Upload folder using huggingface_hub

checkpoint-1511/config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "/mnt/parscratch/users/acp23ay/private/models/OLMo-2-1124-7B-szl_Latn-coli-cw2v/",
+  "architectures": [
+    "Olmo2ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "eos_token_id": 100257,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 4096,
+  "model_type": "olmo2",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "pad_token_id": 100277,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 500000,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.48.0.dev0",
+  "use_cache": true,
+  "vocab_size": 100384
+}

checkpoint-1511/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": 100257,
+  "pad_token_id": 100277,
+  "transformers_version": "4.48.0.dev0"
+}

checkpoint-1511/model-00001-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6957b6d00919f6eb4a191a182bc0fba773efaf07107d7ec2807b88e690038118
+size 4950070472

checkpoint-1511/model-00002-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de0b6a357d03676c4986588aaf4dc96afc3f44a0498514961d2b2cfa1940655e
+size 4991622816

checkpoint-1511/model-00003-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1038a7d346a681a8fb6949e054c2246f3f3c80e826855c1cec6af1155cc607d
+size 4924546896

checkpoint-1511/model-00004-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33c0a21146ace02d36928787ae7f34be868f7c81b894dea98ec7b0e5a3ea01b2
+size 4857404928

checkpoint-1511/model-00005-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7f82a3014ccfc0f67208f63785ae4bc71f09aba86db48bf6139d18178abbfed
+size 4857404928

checkpoint-1511/model-00006-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f686d465a4fa39eff6ee97210806cdd312bb3f289d489d5e58f10d687276812
+size 4614509080

checkpoint-1511/model.safetensors.index.json

@@ -0,0 +1,362 @@
+{
+  "metadata": {
+    "total_size": 29195517952
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00006-of-00006.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.norm.weight": "model-00006-of-00006.safetensors"
+  }
+}

checkpoint-1511/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82ca7f1972e0dd7c222267480cea15f6a52cebd131cf972e6cbe8764cea24620
+size 13055369358

checkpoint-1511/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:386fcc8cc1089aade9450d86fb239ea3483f455fd2d78d8378645feecfec9d69
+size 14244

checkpoint-1511/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0365bbbd991e8527ab663603b79aa7e119fcda5870530787b4c16138808638f0
+size 1064

checkpoint-1511/trainer_state.json

@@ -0,0 +1,2672 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 1511,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0026472534745201853,
+      "grad_norm": 10.610686302185059,
+      "learning_rate": 2.631578947368421e-06,
+      "loss": 6.7026,
+      "step": 4
+    },
+    {
+      "epoch": 0.005294506949040371,
+      "grad_norm": 7.664066314697266,
+      "learning_rate": 5.263157894736842e-06,
+      "loss": 6.1495,
+      "step": 8
+    },
+    {
+      "epoch": 0.007941760423560556,
+      "grad_norm": 9.318928718566895,
+      "learning_rate": 7.894736842105263e-06,
+      "loss": 5.5933,
+      "step": 12
+    },
+    {
+      "epoch": 0.010589013898080741,
+      "grad_norm": 4.640157699584961,
+      "learning_rate": 1.0526315789473684e-05,
+      "loss": 5.2675,
+      "step": 16
+    },
+    {
+      "epoch": 0.013236267372600927,
+      "grad_norm": 4.8390631675720215,
+      "learning_rate": 1.3157894736842106e-05,
+      "loss": 4.7426,
+      "step": 20
+    },
+    {
+      "epoch": 0.01588352084712111,
+      "grad_norm": 3.759246587753296,
+      "learning_rate": 1.5789473684210526e-05,
+      "loss": 4.393,
+      "step": 24
+    },
+    {
+      "epoch": 0.018530774321641297,
+      "grad_norm": 4.17201042175293,
+      "learning_rate": 1.8421052631578947e-05,
+      "loss": 4.53,
+      "step": 28
+    },
+    {
+      "epoch": 0.021178027796161483,
+      "grad_norm": 3.6570184230804443,
+      "learning_rate": 2.105263157894737e-05,
+      "loss": 4.0309,
+      "step": 32
+    },
+    {
+      "epoch": 0.02382528127068167,
+      "grad_norm": 3.2023730278015137,
+      "learning_rate": 2.368421052631579e-05,
+      "loss": 4.0887,
+      "step": 36
+    },
+    {
+      "epoch": 0.026472534745201854,
+      "grad_norm": 3.0748138427734375,
+      "learning_rate": 2.6315789473684212e-05,
+      "loss": 3.7956,
+      "step": 40
+    },
+    {
+      "epoch": 0.02911978821972204,
+      "grad_norm": 4.828571796417236,
+      "learning_rate": 2.8947368421052634e-05,
+      "loss": 3.7004,
+      "step": 44
+    },
+    {
+      "epoch": 0.03176704169424222,
+      "grad_norm": 3.378938674926758,
+      "learning_rate": 3.157894736842105e-05,
+      "loss": 3.4547,
+      "step": 48
+    },
+    {
+      "epoch": 0.03441429516876241,
+      "grad_norm": 2.6294522285461426,
+      "learning_rate": 3.421052631578947e-05,
+      "loss": 3.4377,
+      "step": 52
+    },
+    {
+      "epoch": 0.037061548643282594,
+      "grad_norm": 3.0871987342834473,
+      "learning_rate": 3.6842105263157895e-05,
+      "loss": 3.1353,
+      "step": 56
+    },
+    {
+      "epoch": 0.03970880211780278,
+      "grad_norm": 2.4088680744171143,
+      "learning_rate": 3.9473684210526316e-05,
+      "loss": 2.9782,
+      "step": 60
+    },
+    {
+      "epoch": 0.042356055592322965,
+      "grad_norm": 2.408381700515747,
+      "learning_rate": 4.210526315789474e-05,
+      "loss": 2.9992,
+      "step": 64
+    },
+    {
+      "epoch": 0.04500330906684315,
+      "grad_norm": 2.2480554580688477,
+      "learning_rate": 4.473684210526316e-05,
+      "loss": 2.715,
+      "step": 68
+    },
+    {
+      "epoch": 0.04765056254136334,
+      "grad_norm": 2.2693371772766113,
+      "learning_rate": 4.736842105263158e-05,
+      "loss": 2.7635,
+      "step": 72
+    },
+    {
+      "epoch": 0.05029781601588352,
+      "grad_norm": 2.228167772293091,
+      "learning_rate": 5e-05,
+      "loss": 2.7817,
+      "step": 76
+    },
+    {
+      "epoch": 0.05294506949040371,
+      "grad_norm": 2.0800728797912598,
+      "learning_rate": 5.2631578947368424e-05,
+      "loss": 2.6518,
+      "step": 80
+    },
+    {
+      "epoch": 0.05559232296492389,
+      "grad_norm": 2.3304054737091064,
+      "learning_rate": 5.526315789473685e-05,
+      "loss": 2.5043,
+      "step": 84
+    },
+    {
+      "epoch": 0.05823957643944408,
+      "grad_norm": 3.0883028507232666,
+      "learning_rate": 5.789473684210527e-05,
+      "loss": 2.5016,
+      "step": 88
+    },
+    {
+      "epoch": 0.06088682991396426,
+      "grad_norm": 2.159608840942383,
+      "learning_rate": 6.052631578947369e-05,
+      "loss": 2.4707,
+      "step": 92
+    },
+    {
+      "epoch": 0.06353408338848444,
+      "grad_norm": 2.28015398979187,
+      "learning_rate": 6.31578947368421e-05,
+      "loss": 2.4652,
+      "step": 96
+    },
+    {
+      "epoch": 0.06618133686300463,
+      "grad_norm": 1.9228318929672241,
+      "learning_rate": 6.578947368421054e-05,
+      "loss": 2.3957,
+      "step": 100
+    },
+    {
+      "epoch": 0.06882859033752482,
+      "grad_norm": 11.928667068481445,
+      "learning_rate": 6.842105263157895e-05,
+      "loss": 2.555,
+      "step": 104
+    },
+    {
+      "epoch": 0.071475843812045,
+      "grad_norm": 2.0567405223846436,
+      "learning_rate": 7.105263157894737e-05,
+      "loss": 2.4527,
+      "step": 108
+    },
+    {
+      "epoch": 0.07412309728656519,
+      "grad_norm": 1.8643274307250977,
+      "learning_rate": 7.368421052631579e-05,
+      "loss": 2.2892,
+      "step": 112
+    },
+    {
+      "epoch": 0.07677035076108538,
+      "grad_norm": 2.0170955657958984,
+      "learning_rate": 7.631578947368422e-05,
+      "loss": 2.4977,
+      "step": 116
+    },
+    {
+      "epoch": 0.07941760423560557,
+      "grad_norm": 2.071643114089966,
+      "learning_rate": 7.894736842105263e-05,
+      "loss": 2.417,
+      "step": 120
+    },
+    {
+      "epoch": 0.08206485771012574,
+      "grad_norm": 1.9377845525741577,
+      "learning_rate": 8.157894736842105e-05,
+      "loss": 2.3889,
+      "step": 124
+    },
+    {
+      "epoch": 0.08471211118464593,
+      "grad_norm": 1.820202112197876,
+      "learning_rate": 8.421052631578948e-05,
+      "loss": 2.3357,
+      "step": 128
+    },
+    {
+      "epoch": 0.08735936465916612,
+      "grad_norm": 1.9323867559432983,
+      "learning_rate": 8.68421052631579e-05,
+      "loss": 2.1983,
+      "step": 132
+    },
+    {
+      "epoch": 0.0900066181336863,
+      "grad_norm": 1.919612169265747,
+      "learning_rate": 8.947368421052632e-05,
+      "loss": 2.4496,
+      "step": 136
+    },
+    {
+      "epoch": 0.09265387160820648,
+      "grad_norm": 1.991562008857727,
+      "learning_rate": 9.210526315789474e-05,
+      "loss": 2.3531,
+      "step": 140
+    },
+    {
+      "epoch": 0.09530112508272667,
+      "grad_norm": 1.8413500785827637,
+      "learning_rate": 9.473684210526316e-05,
+      "loss": 1.9909,
+      "step": 144
+    },
+    {
+      "epoch": 0.09794837855724686,
+      "grad_norm": 1.91356360912323,
+      "learning_rate": 9.736842105263158e-05,
+      "loss": 2.0991,
+      "step": 148
+    },
+    {
+      "epoch": 0.10059563203176704,
+      "grad_norm": 1.8988940715789795,
+      "learning_rate": 0.0001,
+      "loss": 2.2675,
+      "step": 152
+    },
+    {
+      "epoch": 0.10324288550628723,
+      "grad_norm": 2.7324349880218506,
+      "learning_rate": 9.999952071344157e-05,
+      "loss": 2.3948,
+      "step": 156
+    },
+    {
+      "epoch": 0.10589013898080742,
+      "grad_norm": 1.8660908937454224,
+      "learning_rate": 9.999808286295485e-05,
+      "loss": 2.3205,
+      "step": 160
+    },
+    {
+      "epoch": 0.10853739245532759,
+      "grad_norm": 1.8560384511947632,
+      "learning_rate": 9.999568647610555e-05,
+      "loss": 2.1836,
+      "step": 164
+    },
+    {
+      "epoch": 0.11118464592984778,
+      "grad_norm": 1.8613684177398682,
+      "learning_rate": 9.999233159883593e-05,
+      "loss": 2.1611,
+      "step": 168
+    },
+    {
+      "epoch": 0.11383189940436797,
+      "grad_norm": 1.9662210941314697,
+      "learning_rate": 9.998801829546386e-05,
+      "loss": 2.1433,
+      "step": 172
+    },
+    {
+      "epoch": 0.11647915287888816,
+      "grad_norm": 1.722792387008667,
+      "learning_rate": 9.998274664868173e-05,
+      "loss": 2.1694,
+      "step": 176
+    },
+    {
+      "epoch": 0.11912640635340833,
+      "grad_norm": 1.8309135437011719,
+      "learning_rate": 9.997651675955466e-05,
+      "loss": 2.198,
+      "step": 180
+    },
+    {
+      "epoch": 0.12177365982792852,
+      "grad_norm": 5.9480204582214355,
+      "learning_rate": 9.996932874751877e-05,
+      "loss": 2.1246,
+      "step": 184
+    },
+    {
+      "epoch": 0.12442091330244871,
+      "grad_norm": 1.776167392730713,
+      "learning_rate": 9.996118275037873e-05,
+      "loss": 2.2526,
+      "step": 188
+    },
+    {
+      "epoch": 0.1270681667769689,
+      "grad_norm": 1.7050362825393677,
+      "learning_rate": 9.995207892430524e-05,
+      "loss": 2.1991,
+      "step": 192
+    },
+    {
+      "epoch": 0.12971542025148908,
+      "grad_norm": 1.8881738185882568,
+      "learning_rate": 9.994201744383196e-05,
+      "loss": 2.2478,
+      "step": 196
+    },
+    {
+      "epoch": 0.13236267372600927,
+      "grad_norm": 1.6576926708221436,
+      "learning_rate": 9.993099850185216e-05,
+      "loss": 1.9793,
+      "step": 200
+    },
+    {
+      "epoch": 0.13500992720052946,
+      "grad_norm": 1.8202764987945557,
+      "learning_rate": 9.991902230961511e-05,
+      "loss": 2.3154,
+      "step": 204
+    },
+    {
+      "epoch": 0.13765718067504965,
+      "grad_norm": 1.8540215492248535,
+      "learning_rate": 9.99060890967219e-05,
+      "loss": 2.3001,
+      "step": 208
+    },
+    {
+      "epoch": 0.14030443414956983,
+      "grad_norm": 2.091411590576172,
+      "learning_rate": 9.989219911112113e-05,
+      "loss": 2.3242,
+      "step": 212
+    },
+    {
+      "epoch": 0.14295168762409,
+      "grad_norm": 2.0059611797332764,
+      "learning_rate": 9.987735261910417e-05,
+      "loss": 2.0196,
+      "step": 216
+    },
+    {
+      "epoch": 0.14559894109861019,
+      "grad_norm": 1.6925115585327148,
+      "learning_rate": 9.986154990529995e-05,
+      "loss": 2.2002,
+      "step": 220
+    },
+    {
+      "epoch": 0.14824619457313037,
+      "grad_norm": 1.5420494079589844,
+      "learning_rate": 9.984479127266961e-05,
+      "loss": 1.9154,
+      "step": 224
+    },
+    {
+      "epoch": 0.15089344804765056,
+      "grad_norm": 1.5900177955627441,
+      "learning_rate": 9.982707704250065e-05,
+      "loss": 1.8659,
+      "step": 228
+    },
+    {
+      "epoch": 0.15354070152217075,
+      "grad_norm": 1.7766467332839966,
+      "learning_rate": 9.980840755440075e-05,
+      "loss": 2.287,
+      "step": 232
+    },
+    {
+      "epoch": 0.15618795499669094,
+      "grad_norm": 1.6030980348587036,
+      "learning_rate": 9.978878316629133e-05,
+      "loss": 1.978,
+      "step": 236
+    },
+    {
+      "epoch": 0.15883520847121113,
+      "grad_norm": 1.698317289352417,
+      "learning_rate": 9.976820425440058e-05,
+      "loss": 2.0607,
+      "step": 240
+    },
+    {
+      "epoch": 0.1614824619457313,
+      "grad_norm": 1.7810238599777222,
+      "learning_rate": 9.974667121325634e-05,
+      "loss": 2.1114,
+      "step": 244
+    },
+    {
+      "epoch": 0.16412971542025148,
+      "grad_norm": 1.692539930343628,
+      "learning_rate": 9.972418445567851e-05,
+      "loss": 1.9014,
+      "step": 248
+    },
+    {
+      "epoch": 0.16677696889477167,
+      "grad_norm": 1.5846136808395386,
+      "learning_rate": 9.97007444127711e-05,
+      "loss": 1.9669,
+      "step": 252
+    },
+    {
+      "epoch": 0.16942422236929186,
+      "grad_norm": 1.6079951524734497,
+      "learning_rate": 9.967635153391401e-05,
+      "loss": 2.0006,
+      "step": 256
+    },
+    {
+      "epoch": 0.17207147584381205,
+      "grad_norm": 1.5769267082214355,
+      "learning_rate": 9.965100628675441e-05,
+      "loss": 2.0626,
+      "step": 260
+    },
+    {
+      "epoch": 0.17471872931833224,
+      "grad_norm": 1.8569773435592651,
+      "learning_rate": 9.962470915719775e-05,
+      "loss": 2.0044,
+      "step": 264
+    },
+    {
+      "epoch": 0.17736598279285243,
+      "grad_norm": 1.6288201808929443,
+      "learning_rate": 9.959746064939846e-05,
+      "loss": 1.916,
+      "step": 268
+    },
+    {
+      "epoch": 0.1800132362673726,
+      "grad_norm": 1.810739278793335,
+      "learning_rate": 9.956926128575026e-05,
+      "loss": 2.0607,
+      "step": 272
+    },
+    {
+      "epoch": 0.18266048974189278,
+      "grad_norm": 1.709789514541626,
+      "learning_rate": 9.954011160687622e-05,
+      "loss": 1.9518,
+      "step": 276
+    },
+    {
+      "epoch": 0.18530774321641297,
+      "grad_norm": 1.5789042711257935,
+      "learning_rate": 9.951001217161829e-05,
+      "loss": 2.0973,
+      "step": 280
+    },
+    {
+      "epoch": 0.18795499669093316,
+      "grad_norm": 1.6972695589065552,
+      "learning_rate": 9.947896355702666e-05,
+      "loss": 2.0861,
+      "step": 284
+    },
+    {
+      "epoch": 0.19060225016545335,
+      "grad_norm": 1.84464693069458,
+      "learning_rate": 9.944696635834867e-05,
+      "loss": 2.0149,
+      "step": 288
+    },
+    {
+      "epoch": 0.19324950363997354,
+      "grad_norm": 1.5586496591567993,
+      "learning_rate": 9.941402118901744e-05,
+      "loss": 1.9106,
+      "step": 292
+    },
+    {
+      "epoch": 0.19589675711449372,
+      "grad_norm": 1.498591423034668,
+      "learning_rate": 9.938012868064e-05,
+      "loss": 1.841,
+      "step": 296
+    },
+    {
+      "epoch": 0.1985440105890139,
+      "grad_norm": 1.595574140548706,
+      "learning_rate": 9.934528948298533e-05,
+      "loss": 1.9874,
+      "step": 300
+    },
+    {
+      "epoch": 0.20119126406353408,
+      "grad_norm": 1.6719728708267212,
+      "learning_rate": 9.930950426397179e-05,
+      "loss": 1.9742,
+      "step": 304
+    },
+    {
+      "epoch": 0.20383851753805426,
+      "grad_norm": 1.7218730449676514,
+      "learning_rate": 9.927277370965435e-05,
+      "loss": 1.8508,
+      "step": 308
+    },
+    {
+      "epoch": 0.20648577101257445,
+      "grad_norm": 1.6651699542999268,
+      "learning_rate": 9.923509852421145e-05,
+      "loss": 1.8824,
+      "step": 312
+    },
+    {
+      "epoch": 0.20913302448709464,
+      "grad_norm": 1.4921704530715942,
+      "learning_rate": 9.919647942993148e-05,
+      "loss": 1.9895,
+      "step": 316
+    },
+    {
+      "epoch": 0.21178027796161483,
+      "grad_norm": 1.5061111450195312,
+      "learning_rate": 9.915691716719898e-05,
+      "loss": 1.7748,
+      "step": 320
+    },
+    {
+      "epoch": 0.21442753143613502,
+      "grad_norm": 1.7094080448150635,
+      "learning_rate": 9.911641249448036e-05,
+      "loss": 2.0328,
+      "step": 324
+    },
+    {
+      "epoch": 0.21707478491065518,
+      "grad_norm": 1.55011785030365,
+      "learning_rate": 9.907496618830942e-05,
+      "loss": 1.8329,
+      "step": 328
+    },
+    {
+      "epoch": 0.21972203838517537,
+      "grad_norm": 3.727109432220459,
+      "learning_rate": 9.903257904327249e-05,
+      "loss": 1.8444,
+      "step": 332
+    },
+    {
+      "epoch": 0.22236929185969556,
+      "grad_norm": 1.7217804193496704,
+      "learning_rate": 9.898925187199308e-05,
+      "loss": 1.9953,
+      "step": 336
+    },
+    {
+      "epoch": 0.22501654533421575,
+      "grad_norm": 1.707441806793213,
+      "learning_rate": 9.894498550511646e-05,
+      "loss": 1.9449,
+      "step": 340
+    },
+    {
+      "epoch": 0.22766379880873594,
+      "grad_norm": 1.6487103700637817,
+      "learning_rate": 9.88997807912936e-05,
+      "loss": 1.9678,
+      "step": 344
+    },
+    {
+      "epoch": 0.23031105228325613,
+      "grad_norm": 1.5246243476867676,
+      "learning_rate": 9.885363859716497e-05,
+      "loss": 1.7917,
+      "step": 348
+    },
+    {
+      "epoch": 0.23295830575777632,
+      "grad_norm": 1.4904006719589233,
+      "learning_rate": 9.88065598073439e-05,
+      "loss": 2.0179,
+      "step": 352
+    },
+    {
+      "epoch": 0.23560555923229648,
+      "grad_norm": 1.547134280204773,
+      "learning_rate": 9.875854532439964e-05,
+      "loss": 1.8441,
+      "step": 356
+    },
+    {
+      "epoch": 0.23825281270681667,
+      "grad_norm": 1.443660855293274,
+      "learning_rate": 9.870959606884004e-05,
+      "loss": 1.8638,
+      "step": 360
+    },
+    {
+      "epoch": 0.24090006618133686,
+      "grad_norm": 1.6207844018936157,
+      "learning_rate": 9.865971297909393e-05,
+      "loss": 2.0387,
+      "step": 364
+    },
+    {
+      "epoch": 0.24354731965585705,
+      "grad_norm": 1.408623218536377,
+      "learning_rate": 9.860889701149307e-05,
+      "loss": 1.8255,
+      "step": 368
+    },
+    {
+      "epoch": 0.24619457313037724,
+      "grad_norm": 1.5555808544158936,
+      "learning_rate": 9.855714914025384e-05,
+      "loss": 1.9796,
+      "step": 372
+    },
+    {
+      "epoch": 0.24884182660489743,
+      "grad_norm": 1.7005336284637451,
+      "learning_rate": 9.850447035745866e-05,
+      "loss": 1.8875,
+      "step": 376
+    },
+    {
+      "epoch": 0.2514890800794176,
+      "grad_norm": 1.5231842994689941,
+      "learning_rate": 9.845086167303679e-05,
+      "loss": 1.9094,
+      "step": 380
+    },
+    {
+      "epoch": 0.2541363335539378,
+      "grad_norm": 1.6607714891433716,
+      "learning_rate": 9.839632411474513e-05,
+      "loss": 2.0968,
+      "step": 384
+    },
+    {
+      "epoch": 0.256783587028458,
+      "grad_norm": 1.6053401231765747,
+      "learning_rate": 9.83408587281484e-05,
+      "loss": 1.9487,
+      "step": 388
+    },
+    {
+      "epoch": 0.25943084050297816,
+      "grad_norm": 1.6648757457733154,
+      "learning_rate": 9.828446657659918e-05,
+      "loss": 2.0489,
+      "step": 392
+    },
+    {
+      "epoch": 0.26207809397749837,
+      "grad_norm": 1.3566395044326782,
+      "learning_rate": 9.82271487412175e-05,
+      "loss": 1.7258,
+      "step": 396
+    },
+    {
+      "epoch": 0.26472534745201853,
+      "grad_norm": 1.6526752710342407,
+      "learning_rate": 9.816890632087006e-05,
+      "loss": 1.8442,
+      "step": 400
+    },
+    {
+      "epoch": 0.2673726009265387,
+      "grad_norm": 1.5632907152175903,
+      "learning_rate": 9.810974043214922e-05,
+      "loss": 1.9051,
+      "step": 404
+    },
+    {
+      "epoch": 0.2700198544010589,
+      "grad_norm": 1.646107792854309,
+      "learning_rate": 9.804965220935161e-05,
+      "loss": 1.9339,
+      "step": 408
+    },
+    {
+      "epoch": 0.2726671078755791,
+      "grad_norm": 1.6232599020004272,
+      "learning_rate": 9.798864280445632e-05,
+      "loss": 1.8981,
+      "step": 412
+    },
+    {
+      "epoch": 0.2753143613500993,
+      "grad_norm": 1.5264742374420166,
+      "learning_rate": 9.792671338710285e-05,
+      "loss": 1.8395,
+      "step": 416
+    },
+    {
+      "epoch": 0.27796161482461945,
+      "grad_norm": 2.12361741065979,
+      "learning_rate": 9.786386514456872e-05,
+      "loss": 1.9673,
+      "step": 420
+    },
+    {
+      "epoch": 0.28060886829913967,
+      "grad_norm": 1.5487679243087769,
+      "learning_rate": 9.780009928174661e-05,
+      "loss": 1.9475,
+      "step": 424
+    },
+    {
+      "epoch": 0.28325612177365983,
+      "grad_norm": 1.5165311098098755,
+      "learning_rate": 9.773541702112137e-05,
+      "loss": 1.9555,
+      "step": 428
+    },
+    {
+      "epoch": 0.28590337524818,
+      "grad_norm": 1.5391660928726196,
+      "learning_rate": 9.766981960274653e-05,
+      "loss": 1.8691,
+      "step": 432
+    },
+    {
+      "epoch": 0.2885506287227002,
+      "grad_norm": 1.5682833194732666,
+      "learning_rate": 9.760330828422053e-05,
+      "loss": 1.7862,
+      "step": 436
+    },
+    {
+      "epoch": 0.29119788219722037,
+      "grad_norm": 1.403124451637268,
+      "learning_rate": 9.753588434066258e-05,
+      "loss": 1.9392,
+      "step": 440
+    },
+    {
+      "epoch": 0.2938451356717406,
+      "grad_norm": 1.4275953769683838,
+      "learning_rate": 9.746754906468832e-05,
+      "loss": 1.9314,
+      "step": 444
+    },
+    {
+      "epoch": 0.29649238914626075,
+      "grad_norm": 1.4967845678329468,
+      "learning_rate": 9.73983037663849e-05,
+      "loss": 1.854,
+      "step": 448
+    },
+    {
+      "epoch": 0.29913964262078097,
+      "grad_norm": 1.601547122001648,
+      "learning_rate": 9.732814977328592e-05,
+      "loss": 1.9411,
+      "step": 452
+    },
+    {
+      "epoch": 0.3017868960953011,
+      "grad_norm": 1.412458896636963,
+      "learning_rate": 9.725708843034605e-05,
+      "loss": 1.8616,
+      "step": 456
+    },
+    {
+      "epoch": 0.3044341495698213,
+      "grad_norm": 1.4348307847976685,
+      "learning_rate": 9.718512109991514e-05,
+      "loss": 1.8257,
+      "step": 460
+    },
+    {
+      "epoch": 0.3070814030443415,
+      "grad_norm": 1.4989802837371826,
+      "learning_rate": 9.711224916171215e-05,
+      "loss": 1.9711,
+      "step": 464
+    },
+    {
+      "epoch": 0.30972865651886167,
+      "grad_norm": 1.5016189813613892,
+      "learning_rate": 9.703847401279871e-05,
+      "loss": 1.8066,
+      "step": 468
+    },
+    {
+      "epoch": 0.3123759099933819,
+      "grad_norm": 1.5456262826919556,
+      "learning_rate": 9.69637970675523e-05,
+      "loss": 1.7682,
+      "step": 472
+    },
+    {
+      "epoch": 0.31502316346790205,
+      "grad_norm": 1.649239420890808,
+      "learning_rate": 9.688821975763918e-05,
+      "loss": 1.9982,
+      "step": 476
+    },
+    {
+      "epoch": 0.31767041694242226,
+      "grad_norm": 1.5132896900177002,
+      "learning_rate": 9.681174353198687e-05,
+      "loss": 1.6853,
+      "step": 480
+    },
+    {
+      "epoch": 0.3203176704169424,
+      "grad_norm": 1.5706440210342407,
+      "learning_rate": 9.673436985675645e-05,
+      "loss": 1.8175,
+      "step": 484
+    },
+    {
+      "epoch": 0.3229649238914626,
+      "grad_norm": 1.4862254858016968,
+      "learning_rate": 9.665610021531447e-05,
+      "loss": 1.9173,
+      "step": 488
+    },
+    {
+      "epoch": 0.3256121773659828,
+      "grad_norm": 1.6407462358474731,
+      "learning_rate": 9.657693610820437e-05,
+      "loss": 2.0596,
+      "step": 492
+    },
+    {
+      "epoch": 0.32825943084050296,
+      "grad_norm": 1.6441835165023804,
+      "learning_rate": 9.649687905311785e-05,
+      "loss": 2.0295,
+      "step": 496
+    },
+    {
+      "epoch": 0.3309066843150232,
+      "grad_norm": 1.378661870956421,
+      "learning_rate": 9.641593058486574e-05,
+      "loss": 1.7642,
+      "step": 500
+    },
+    {
+      "epoch": 0.33355393778954334,
+      "grad_norm": 1.4671963453292847,
+      "learning_rate": 9.633409225534855e-05,
+      "loss": 1.8116,
+      "step": 504
+    },
+    {
+      "epoch": 0.33620119126406356,
+      "grad_norm": 1.5696290731430054,
+      "learning_rate": 9.625136563352671e-05,
+      "loss": 1.8052,
+      "step": 508
+    },
+    {
+      "epoch": 0.3388484447385837,
+      "grad_norm": 1.4909025430679321,
+      "learning_rate": 9.616775230539057e-05,
+      "loss": 1.8951,
+      "step": 512
+    },
+    {
+      "epoch": 0.3414956982131039,
+      "grad_norm": 1.5281769037246704,
+      "learning_rate": 9.608325387392986e-05,
+      "loss": 1.7721,
+      "step": 516
+    },
+    {
+      "epoch": 0.3441429516876241,
+      "grad_norm": 1.4955800771713257,
+      "learning_rate": 9.599787195910313e-05,
+      "loss": 1.6392,
+      "step": 520
+    },
+    {
+      "epoch": 0.34679020516214426,
+      "grad_norm": 1.4858269691467285,
+      "learning_rate": 9.591160819780649e-05,
+      "loss": 1.7931,
+      "step": 524
+    },
+    {
+      "epoch": 0.3494374586366645,
+      "grad_norm": 1.5280017852783203,
+      "learning_rate": 9.582446424384242e-05,
+      "loss": 1.8464,
+      "step": 528
+    },
+    {
+      "epoch": 0.35208471211118464,
+      "grad_norm": 1.5444755554199219,
+      "learning_rate": 9.573644176788794e-05,
+      "loss": 1.8197,
+      "step": 532
+    },
+    {
+      "epoch": 0.35473196558570486,
+      "grad_norm": 1.5050253868103027,
+      "learning_rate": 9.564754245746264e-05,
+      "loss": 1.9478,
+      "step": 536
+    },
+    {
+      "epoch": 0.357379219060225,
+      "grad_norm": 1.461586356163025,
+      "learning_rate": 9.555776801689632e-05,
+      "loss": 1.8323,
+      "step": 540
+    },
+    {
+      "epoch": 0.3600264725347452,
+      "grad_norm": 1.51008141040802,
+      "learning_rate": 9.546712016729624e-05,
+      "loss": 1.9505,
+      "step": 544
+    },
+    {
+      "epoch": 0.3626737260092654,
+      "grad_norm": 1.442744493484497,
+      "learning_rate": 9.537560064651427e-05,
+      "loss": 1.6729,
+      "step": 548
+    },
+    {
+      "epoch": 0.36532097948378556,
+      "grad_norm": 1.5391943454742432,
+      "learning_rate": 9.528321120911346e-05,
+      "loss": 1.8549,
+      "step": 552
+    },
+    {
+      "epoch": 0.3679682329583058,
+      "grad_norm": 1.5444978475570679,
+      "learning_rate": 9.51899536263344e-05,
+      "loss": 1.8613,
+      "step": 556
+    },
+    {
+      "epoch": 0.37061548643282594,
+      "grad_norm": 3.030534267425537,
+      "learning_rate": 9.509582968606136e-05,
+      "loss": 1.7753,
+      "step": 560
+    },
+    {
+      "epoch": 0.37326273990734615,
+      "grad_norm": 1.452682614326477,
+      "learning_rate": 9.500084119278788e-05,
+      "loss": 1.7413,
+      "step": 564
+    },
+    {
+      "epoch": 0.3759099933818663,
+      "grad_norm": 1.351912260055542,
+      "learning_rate": 9.49049899675823e-05,
+      "loss": 1.9188,
+      "step": 568
+    },
+    {
+      "epoch": 0.3785572468563865,
+      "grad_norm": 1.6070955991744995,
+      "learning_rate": 9.480827784805278e-05,
+      "loss": 1.8472,
+      "step": 572
+    },
+    {
+      "epoch": 0.3812045003309067,
+      "grad_norm": 1.619681715965271,
+      "learning_rate": 9.471070668831208e-05,
+      "loss": 1.6604,
+      "step": 576
+    },
+    {
+      "epoch": 0.38385175380542685,
+      "grad_norm": 1.5847423076629639,
+      "learning_rate": 9.4612278358942e-05,
+      "loss": 1.7116,
+      "step": 580
+    },
+    {
+      "epoch": 0.38649900727994707,
+      "grad_norm": 1.4484244585037231,
+      "learning_rate": 9.451299474695754e-05,
+      "loss": 1.7518,
+      "step": 584
+    },
+    {
+      "epoch": 0.38914626075446723,
+      "grad_norm": 1.4682494401931763,
+      "learning_rate": 9.441285775577075e-05,
+      "loss": 1.7971,
+      "step": 588
+    },
+    {
+      "epoch": 0.39179351422898745,
+      "grad_norm": 1.4084190130233765,
+      "learning_rate": 9.431186930515419e-05,
+      "loss": 1.7613,
+      "step": 592
+    },
+    {
+      "epoch": 0.3944407677035076,
+      "grad_norm": 1.4905308485031128,
+      "learning_rate": 9.421003133120412e-05,
+      "loss": 1.7662,
+      "step": 596
+    },
+    {
+      "epoch": 0.3970880211780278,
+      "grad_norm": 1.5779659748077393,
+      "learning_rate": 9.410734578630343e-05,
+      "loss": 1.7133,
+      "step": 600
+    },
+    {
+      "epoch": 0.399735274652548,
+      "grad_norm": 1.415911316871643,
+      "learning_rate": 9.400381463908416e-05,
+      "loss": 1.8177,
+      "step": 604
+    },
+    {
+      "epoch": 0.40238252812706815,
+      "grad_norm": 1.4674650430679321,
+      "learning_rate": 9.389943987438983e-05,
+      "loss": 1.6904,
+      "step": 608
+    },
+    {
+      "epoch": 0.40502978160158837,
+      "grad_norm": 1.442590355873108,
+      "learning_rate": 9.379422349323728e-05,
+      "loss": 1.686,
+      "step": 612
+    },
+    {
+      "epoch": 0.40767703507610853,
+      "grad_norm": 1.4115177392959595,
+      "learning_rate": 9.368816751277843e-05,
+      "loss": 1.6588,
+      "step": 616
+    },
+    {
+      "epoch": 0.41032428855062875,
+      "grad_norm": 1.407148003578186,
+      "learning_rate": 9.358127396626147e-05,
+      "loss": 1.6998,
+      "step": 620
+    },
+    {
+      "epoch": 0.4129715420251489,
+      "grad_norm": 1.3762400150299072,
+      "learning_rate": 9.347354490299205e-05,
+      "loss": 1.7738,
+      "step": 624
+    },
+    {
+      "epoch": 0.41561879549966907,
+      "grad_norm": 1.3269281387329102,
+      "learning_rate": 9.336498238829384e-05,
+      "loss": 1.6914,
+      "step": 628
+    },
+    {
+      "epoch": 0.4182660489741893,
+      "grad_norm": 1.2647446393966675,
+      "learning_rate": 9.325558850346897e-05,
+      "loss": 1.7098,
+      "step": 632
+    },
+    {
+      "epoch": 0.42091330244870945,
+      "grad_norm": 1.5389339923858643,
+      "learning_rate": 9.31453653457582e-05,
+      "loss": 1.6726,
+      "step": 636
+    },
+    {
+      "epoch": 0.42356055592322966,
+      "grad_norm": 1.525876760482788,
+      "learning_rate": 9.303431502830065e-05,
+      "loss": 1.6839,
+      "step": 640
+    },
+    {
+      "epoch": 0.4262078093977498,
+      "grad_norm": 1.5353223085403442,
+      "learning_rate": 9.292243968009331e-05,
+      "loss": 1.7669,
+      "step": 644
+    },
+    {
+      "epoch": 0.42885506287227004,
+      "grad_norm": 1.3794018030166626,
+      "learning_rate": 9.280974144595018e-05,
+      "loss": 1.6615,
+      "step": 648
+    },
+    {
+      "epoch": 0.4315023163467902,
+      "grad_norm": 1.5229333639144897,
+      "learning_rate": 9.269622248646124e-05,
+      "loss": 1.7976,
+      "step": 652
+    },
+    {
+      "epoch": 0.43414956982131037,
+      "grad_norm": 1.474219799041748,
+      "learning_rate": 9.258188497795093e-05,
+      "loss": 1.6914,
+      "step": 656
+    },
+    {
+      "epoch": 0.4367968232958306,
+      "grad_norm": 1.3414136171340942,
+      "learning_rate": 9.24667311124365e-05,
+      "loss": 1.8003,
+      "step": 660
+    },
+    {
+      "epoch": 0.43944407677035074,
+      "grad_norm": 1.3020344972610474,
+      "learning_rate": 9.23507630975859e-05,
+      "loss": 1.7176,
+      "step": 664
+    },
+    {
+      "epoch": 0.44209133024487096,
+      "grad_norm": 1.4163818359375,
+      "learning_rate": 9.223398315667561e-05,
+      "loss": 1.674,
+      "step": 668
+    },
+    {
+      "epoch": 0.4447385837193911,
+      "grad_norm": 1.3935587406158447,
+      "learning_rate": 9.211639352854787e-05,
+      "loss": 1.7689,
+      "step": 672
+    },
+    {
+      "epoch": 0.44738583719391134,
+      "grad_norm": 1.3712141513824463,
+      "learning_rate": 9.199799646756777e-05,
+      "loss": 1.7988,
+      "step": 676
+    },
+    {
+      "epoch": 0.4500330906684315,
+      "grad_norm": 1.5320765972137451,
+      "learning_rate": 9.187879424358014e-05,
+      "loss": 1.831,
+      "step": 680
+    },
+    {
+      "epoch": 0.45268034414295166,
+      "grad_norm": 1.5243709087371826,
+      "learning_rate": 9.17587891418659e-05,
+      "loss": 1.6793,
+      "step": 684
+    },
+    {
+      "epoch": 0.4553275976174719,
+      "grad_norm": 1.727277159690857,
+      "learning_rate": 9.163798346309837e-05,
+      "loss": 1.9142,
+      "step": 688
+    },
+    {
+      "epoch": 0.45797485109199204,
+      "grad_norm": 1.2457393407821655,
+      "learning_rate": 9.151637952329903e-05,
+      "loss": 1.577,
+      "step": 692
+    },
+    {
+      "epoch": 0.46062210456651226,
+      "grad_norm": 1.939177393913269,
+      "learning_rate": 9.139397965379327e-05,
+      "loss": 1.8163,
+      "step": 696
+    },
+    {
+      "epoch": 0.4632693580410324,
+      "grad_norm": 1.477421760559082,
+      "learning_rate": 9.127078620116556e-05,
+      "loss": 1.7861,
+      "step": 700
+    },
+    {
+      "epoch": 0.46591661151555264,
+      "grad_norm": 1.3966877460479736,
+      "learning_rate": 9.114680152721453e-05,
+      "loss": 1.6275,
+      "step": 704
+    },
+    {
+      "epoch": 0.4685638649900728,
+      "grad_norm": 1.5782712697982788,
+      "learning_rate": 9.102202800890772e-05,
+      "loss": 1.9066,
+      "step": 708
+    },
+    {
+      "epoch": 0.47121111846459296,
+      "grad_norm": 1.3626751899719238,
+      "learning_rate": 9.089646803833589e-05,
+      "loss": 1.7002,
+      "step": 712
+    },
+    {
+      "epoch": 0.4738583719391132,
+      "grad_norm": 1.4438906908035278,
+      "learning_rate": 9.077012402266731e-05,
+      "loss": 1.6942,
+      "step": 716
+    },
+    {
+      "epoch": 0.47650562541363334,
+      "grad_norm": 1.39316725730896,
+      "learning_rate": 9.064299838410152e-05,
+      "loss": 1.6446,
+      "step": 720
+    },
+    {
+      "epoch": 0.47915287888815355,
+      "grad_norm": 1.4561007022857666,
+      "learning_rate": 9.051509355982293e-05,
+      "loss": 1.6668,
+      "step": 724
+    },
+    {
+      "epoch": 0.4818001323626737,
+      "grad_norm": 1.3590714931488037,
+      "learning_rate": 9.038641200195404e-05,
+      "loss": 1.9216,
+      "step": 728
+    },
+    {
+      "epoch": 0.48444738583719393,
+      "grad_norm": 1.575279712677002,
+      "learning_rate": 9.025695617750848e-05,
+      "loss": 1.8161,
+      "step": 732
+    },
+    {
+      "epoch": 0.4870946393117141,
+      "grad_norm": 1.4522175788879395,
+      "learning_rate": 9.012672856834373e-05,
+      "loss": 1.8648,
+      "step": 736
+    },
+    {
+      "epoch": 0.48974189278623426,
+      "grad_norm": 1.3954745531082153,
+      "learning_rate": 8.999573167111348e-05,
+      "loss": 1.5615,
+      "step": 740
+    },
+    {
+      "epoch": 0.4923891462607545,
+      "grad_norm": 1.4717799425125122,
+      "learning_rate": 8.986396799721983e-05,
+      "loss": 1.6382,
+      "step": 744
+    },
+    {
+      "epoch": 0.49503639973527463,
+      "grad_norm": 1.450149416923523,
+      "learning_rate": 8.973144007276508e-05,
+      "loss": 1.7229,
+      "step": 748
+    },
+    {
+      "epoch": 0.49768365320979485,
+      "grad_norm": 1.4680171012878418,
+      "learning_rate": 8.959815043850336e-05,
+      "loss": 1.6906,
+      "step": 752
+    },
+    {
+      "epoch": 0.500330906684315,
+      "grad_norm": 1.4839046001434326,
+      "learning_rate": 8.946410164979184e-05,
+      "loss": 1.8205,
+      "step": 756
+    },
+    {
+      "epoch": 0.5029781601588352,
+      "grad_norm": 1.503608226776123,
+      "learning_rate": 8.932929627654185e-05,
+      "loss": 1.5467,
+      "step": 760
+    },
+    {
+      "epoch": 0.5056254136333554,
+      "grad_norm": 1.3385919332504272,
+      "learning_rate": 8.919373690316952e-05,
+      "loss": 1.7181,
+      "step": 764
+    },
+    {
+      "epoch": 0.5082726671078756,
+      "grad_norm": 1.4352362155914307,
+      "learning_rate": 8.905742612854628e-05,
+      "loss": 1.702,
+      "step": 768
+    },
+    {
+      "epoch": 0.5109199205823958,
+      "grad_norm": 1.5198529958724976,
+      "learning_rate": 8.892036656594898e-05,
+      "loss": 1.6662,
+      "step": 772
+    },
+    {
+      "epoch": 0.513567174056916,
+      "grad_norm": 1.360083818435669,
+      "learning_rate": 8.87825608430099e-05,
+      "loss": 1.6477,
+      "step": 776
+    },
+    {
+      "epoch": 0.5162144275314361,
+      "grad_norm": 1.4632827043533325,
+      "learning_rate": 8.864401160166624e-05,
+      "loss": 1.6028,
+      "step": 780
+    },
+    {
+      "epoch": 0.5188616810059563,
+      "grad_norm": 1.4365050792694092,
+      "learning_rate": 8.85047214981096e-05,
+      "loss": 1.7236,
+      "step": 784
+    },
+    {
+      "epoch": 0.5215089344804765,
+      "grad_norm": 1.3933697938919067,
+      "learning_rate": 8.83646932027349e-05,
+      "loss": 1.678,
+      "step": 788
+    },
+    {
+      "epoch": 0.5241561879549967,
+      "grad_norm": 1.4825220108032227,
+      "learning_rate": 8.822392940008937e-05,
+      "loss": 1.5583,
+      "step": 792
+    },
+    {
+      "epoch": 0.5268034414295168,
+      "grad_norm": 1.331561803817749,
+      "learning_rate": 8.808243278882094e-05,
+      "loss": 1.5029,
+      "step": 796
+    },
+    {
+      "epoch": 0.5294506949040371,
+      "grad_norm": 1.4000210762023926,
+      "learning_rate": 8.794020608162656e-05,
+      "loss": 1.7178,
+      "step": 800
+    },
+    {
+      "epoch": 0.5320979483785573,
+      "grad_norm": 1.4191969633102417,
+      "learning_rate": 8.779725200520021e-05,
+      "loss": 1.7233,
+      "step": 804
+    },
+    {
+      "epoch": 0.5347452018530774,
+      "grad_norm": 1.319602608680725,
+      "learning_rate": 8.765357330018056e-05,
+      "loss": 1.4836,
+      "step": 808
+    },
+    {
+      "epoch": 0.5373924553275976,
+      "grad_norm": 1.5123964548110962,
+      "learning_rate": 8.750917272109848e-05,
+      "loss": 1.6977,
+      "step": 812
+    },
+    {
+      "epoch": 0.5400397088021178,
+      "grad_norm": 1.4251317977905273,
+      "learning_rate": 8.736405303632427e-05,
+      "loss": 1.6526,
+      "step": 816
+    },
+    {
+      "epoch": 0.542686962276638,
+      "grad_norm": 1.515101671218872,
+      "learning_rate": 8.721821702801449e-05,
+      "loss": 1.7234,
+      "step": 820
+    },
+    {
+      "epoch": 0.5453342157511581,
+      "grad_norm": 1.4469647407531738,
+      "learning_rate": 8.707166749205866e-05,
+      "loss": 1.8084,
+      "step": 824
+    },
+    {
+      "epoch": 0.5479814692256784,
+      "grad_norm": 1.3407204151153564,
+      "learning_rate": 8.692440723802571e-05,
+      "loss": 1.5344,
+      "step": 828
+    },
+    {
+      "epoch": 0.5506287227001986,
+      "grad_norm": 1.460547685623169,
+      "learning_rate": 8.677643908911007e-05,
+      "loss": 1.6787,
+      "step": 832
+    },
+    {
+      "epoch": 0.5532759761747187,
+      "grad_norm": 1.409914493560791,
+      "learning_rate": 8.662776588207747e-05,
+      "loss": 1.648,
+      "step": 836
+    },
+    {
+      "epoch": 0.5559232296492389,
+      "grad_norm": 1.2722299098968506,
+      "learning_rate": 8.647839046721076e-05,
+      "loss": 1.5172,
+      "step": 840
+    },
+    {
+      "epoch": 0.5585704831237591,
+      "grad_norm": 1.4182597398757935,
+      "learning_rate": 8.632831570825508e-05,
+      "loss": 1.8057,
+      "step": 844
+    },
+    {
+      "epoch": 0.5612177365982793,
+      "grad_norm": 1.3057011365890503,
+      "learning_rate": 8.617754448236298e-05,
+      "loss": 1.6716,
+      "step": 848
+    },
+    {
+      "epoch": 0.5638649900727994,
+      "grad_norm": 1.347036361694336,
+      "learning_rate": 8.602607968003935e-05,
+      "loss": 1.558,
+      "step": 852
+    },
+    {
+      "epoch": 0.5665122435473197,
+      "grad_norm": 1.415988802909851,
+      "learning_rate": 8.587392420508598e-05,
+      "loss": 1.6299,
+      "step": 856
+    },
+    {
+      "epoch": 0.5691594970218399,
+      "grad_norm": 1.5263333320617676,
+      "learning_rate": 8.572108097454578e-05,
+      "loss": 1.8237,
+      "step": 860
+    },
+    {
+      "epoch": 0.57180675049636,
+      "grad_norm": 1.2556434869766235,
+      "learning_rate": 8.556755291864701e-05,
+      "loss": 1.6314,
+      "step": 864
+    },
+    {
+      "epoch": 0.5744540039708802,
+      "grad_norm": 1.3698112964630127,
+      "learning_rate": 8.541334298074701e-05,
+      "loss": 1.7362,
+      "step": 868
+    },
+    {
+      "epoch": 0.5771012574454004,
+      "grad_norm": 1.2662156820297241,
+      "learning_rate": 8.525845411727581e-05,
+      "loss": 1.4891,
+      "step": 872
+    },
+    {
+      "epoch": 0.5797485109199206,
+      "grad_norm": 1.3394712209701538,
+      "learning_rate": 8.51028892976794e-05,
+      "loss": 1.6579,
+      "step": 876
+    },
+    {
+      "epoch": 0.5823957643944407,
+      "grad_norm": 1.3826072216033936,
+      "learning_rate": 8.494665150436288e-05,
+      "loss": 1.6722,
+      "step": 880
+    },
+    {
+      "epoch": 0.585043017868961,
+      "grad_norm": 1.2150722742080688,
+      "learning_rate": 8.478974373263318e-05,
+      "loss": 1.4603,
+      "step": 884
+    },
+    {
+      "epoch": 0.5876902713434812,
+      "grad_norm": 1.5128766298294067,
+      "learning_rate": 8.463216899064179e-05,
+      "loss": 1.7955,
+      "step": 888
+    },
+    {
+      "epoch": 0.5903375248180013,
+      "grad_norm": 1.3931621313095093,
+      "learning_rate": 8.447393029932692e-05,
+      "loss": 1.8027,
+      "step": 892
+    },
+    {
+      "epoch": 0.5929847782925215,
+      "grad_norm": 1.4115737676620483,
+      "learning_rate": 8.431503069235565e-05,
+      "loss": 1.5827,
+      "step": 896
+    },
+    {
+      "epoch": 0.5956320317670417,
+      "grad_norm": 1.4222979545593262,
+      "learning_rate": 8.415547321606584e-05,
+      "loss": 1.6744,
+      "step": 900
+    },
+    {
+      "epoch": 0.5982792852415619,
+      "grad_norm": 1.3109382390975952,
+      "learning_rate": 8.399526092940768e-05,
+      "loss": 1.6247,
+      "step": 904
+    },
+    {
+      "epoch": 0.600926538716082,
+      "grad_norm": 1.5184141397476196,
+      "learning_rate": 8.38343969038849e-05,
+      "loss": 1.7907,
+      "step": 908
+    },
+    {
+      "epoch": 0.6035737921906023,
+      "grad_norm": 1.4683995246887207,
+      "learning_rate": 8.367288422349617e-05,
+      "loss": 1.7061,
+      "step": 912
+    },
+    {
+      "epoch": 0.6062210456651225,
+      "grad_norm": 1.4031612873077393,
+      "learning_rate": 8.351072598467576e-05,
+      "loss": 1.6612,
+      "step": 916
+    },
+    {
+      "epoch": 0.6088682991396426,
+      "grad_norm": 1.2582640647888184,
+      "learning_rate": 8.334792529623419e-05,
+      "loss": 1.4902,
+      "step": 920
+    },
+    {
+      "epoch": 0.6115155526141628,
+      "grad_norm": 2.3730320930480957,
+      "learning_rate": 8.318448527929877e-05,
+      "loss": 1.6021,
+      "step": 924
+    },
+    {
+      "epoch": 0.614162806088683,
+      "grad_norm": 1.5273563861846924,
+      "learning_rate": 8.302040906725361e-05,
+      "loss": 1.7583,
+      "step": 928
+    },
+    {
+      "epoch": 0.6168100595632032,
+      "grad_norm": 1.3741856813430786,
+      "learning_rate": 8.285569980567964e-05,
+      "loss": 1.6579,
+      "step": 932
+    },
+    {
+      "epoch": 0.6194573130377233,
+      "grad_norm": 1.4378982782363892,
+      "learning_rate": 8.269036065229427e-05,
+      "loss": 1.7302,
+      "step": 936
+    },
+    {
+      "epoch": 0.6221045665122436,
+      "grad_norm": 1.438852310180664,
+      "learning_rate": 8.252439477689082e-05,
+      "loss": 1.6479,
+      "step": 940
+    },
+    {
+      "epoch": 0.6247518199867638,
+      "grad_norm": 1.4666221141815186,
+      "learning_rate": 8.235780536127787e-05,
+      "loss": 1.5313,
+      "step": 944
+    },
+    {
+      "epoch": 0.6273990734612839,
+      "grad_norm": 1.4131202697753906,
+      "learning_rate": 8.21905955992181e-05,
+      "loss": 1.6741,
+      "step": 948
+    },
+    {
+      "epoch": 0.6300463269358041,
+      "grad_norm": 1.3892124891281128,
+      "learning_rate": 8.202276869636713e-05,
+      "loss": 1.6632,
+      "step": 952
+    },
+    {
+      "epoch": 0.6326935804103243,
+      "grad_norm": 1.499426007270813,
+      "learning_rate": 8.185432787021216e-05,
+      "loss": 1.5264,
+      "step": 956
+    },
+    {
+      "epoch": 0.6353408338848445,
+      "grad_norm": 1.3629894256591797,
+      "learning_rate": 8.168527635001015e-05,
+      "loss": 1.5446,
+      "step": 960
+    },
+    {
+      "epoch": 0.6379880873593646,
+      "grad_norm": 1.2739462852478027,
+      "learning_rate": 8.151561737672591e-05,
+      "loss": 1.7349,
+      "step": 964
+    },
+    {
+      "epoch": 0.6406353408338848,
+      "grad_norm": 1.306991696357727,
+      "learning_rate": 8.134535420297008e-05,
+      "loss": 1.5969,
+      "step": 968
+    },
+    {
+      "epoch": 0.6432825943084051,
+      "grad_norm": 1.3341622352600098,
+      "learning_rate": 8.117449009293668e-05,
+      "loss": 1.6726,
+      "step": 972
+    },
+    {
+      "epoch": 0.6459298477829252,
+      "grad_norm": 1.3210887908935547,
+      "learning_rate": 8.100302832234056e-05,
+      "loss": 1.6547,
+      "step": 976
+    },
+    {
+      "epoch": 0.6485771012574454,
+      "grad_norm": 1.6015820503234863,
+      "learning_rate": 8.083097217835461e-05,
+      "loss": 1.6409,
+      "step": 980
+    },
+    {
+      "epoch": 0.6512243547319656,
+      "grad_norm": 1.3567839860916138,
+      "learning_rate": 8.065832495954668e-05,
+      "loss": 1.7578,
+      "step": 984
+    },
+    {
+      "epoch": 0.6538716082064858,
+      "grad_norm": 1.3692917823791504,
+      "learning_rate": 8.048508997581647e-05,
+      "loss": 1.647,
+      "step": 988
+    },
+    {
+      "epoch": 0.6565188616810059,
+      "grad_norm": 1.4648549556732178,
+      "learning_rate": 8.03112705483319e-05,
+      "loss": 1.7708,
+      "step": 992
+    },
+    {
+      "epoch": 0.6591661151555261,
+      "grad_norm": 1.3697330951690674,
+      "learning_rate": 8.013687000946561e-05,
+      "loss": 1.5405,
+      "step": 996
+    },
+    {
+      "epoch": 0.6618133686300464,
+      "grad_norm": 1.4155359268188477,
+      "learning_rate": 7.996189170273096e-05,
+      "loss": 1.708,
+      "step": 1000
+    },
+    {
+      "epoch": 0.6644606221045665,
+      "grad_norm": 1.0955262184143066,
+      "learning_rate": 7.978633898271795e-05,
+      "loss": 1.3617,
+      "step": 1004
+    },
+    {
+      "epoch": 0.6671078755790867,
+      "grad_norm": 1.480194330215454,
+      "learning_rate": 7.961021521502895e-05,
+      "loss": 1.6045,
+      "step": 1008
+    },
+    {
+      "epoch": 0.6697551290536069,
+      "grad_norm": 1.3911713361740112,
+      "learning_rate": 7.943352377621414e-05,
+      "loss": 1.6599,
+      "step": 1012
+    },
+    {
+      "epoch": 0.6724023825281271,
+      "grad_norm": 1.230700969696045,
+      "learning_rate": 7.925626805370678e-05,
+      "loss": 1.4586,
+      "step": 1016
+    },
+    {
+      "epoch": 0.6750496360026472,
+      "grad_norm": 1.3873647451400757,
+      "learning_rate": 7.907845144575829e-05,
+      "loss": 1.6458,
+      "step": 1020
+    },
+    {
+      "epoch": 0.6776968894771674,
+      "grad_norm": 1.2631981372833252,
+      "learning_rate": 7.890007736137307e-05,
+      "loss": 1.5528,
+      "step": 1024
+    },
+    {
+      "epoch": 0.6803441429516877,
+      "grad_norm": 1.4044338464736938,
+      "learning_rate": 7.872114922024313e-05,
+      "loss": 1.5926,
+      "step": 1028
+    },
+    {
+      "epoch": 0.6829913964262078,
+      "grad_norm": 1.4623993635177612,
+      "learning_rate": 7.854167045268264e-05,
+      "loss": 1.4877,
+      "step": 1032
+    },
+    {
+      "epoch": 0.685638649900728,
+      "grad_norm": 1.2731891870498657,
+      "learning_rate": 7.836164449956199e-05,
+      "loss": 1.5832,
+      "step": 1036
+    },
+    {
+      "epoch": 0.6882859033752482,
+      "grad_norm": 1.3587007522583008,
+      "learning_rate": 7.818107481224198e-05,
+      "loss": 1.496,
+      "step": 1040
+    },
+    {
+      "epoch": 0.6909331568497684,
+      "grad_norm": 1.5598070621490479,
+      "learning_rate": 7.799996485250755e-05,
+      "loss": 1.4943,
+      "step": 1044
+    },
+    {
+      "epoch": 0.6935804103242885,
+      "grad_norm": 1.2438719272613525,
+      "learning_rate": 7.781831809250151e-05,
+      "loss": 1.6271,
+      "step": 1048
+    },
+    {
+      "epoch": 0.6962276637988087,
+      "grad_norm": 1.393099069595337,
+      "learning_rate": 7.763613801465786e-05,
+      "loss": 1.6096,
+      "step": 1052
+    },
+    {
+      "epoch": 0.698874917273329,
+      "grad_norm": 1.3444733619689941,
+      "learning_rate": 7.745342811163507e-05,
+      "loss": 1.5174,
+      "step": 1056
+    },
+    {
+      "epoch": 0.7015221707478491,
+      "grad_norm": 1.3353770971298218,
+      "learning_rate": 7.727019188624922e-05,
+      "loss": 1.6596,
+      "step": 1060
+    },
+    {
+      "epoch": 0.7041694242223693,
+      "grad_norm": 1.3728408813476562,
+      "learning_rate": 7.708643285140667e-05,
+      "loss": 1.7569,
+      "step": 1064
+    },
+    {
+      "epoch": 0.7068166776968895,
+      "grad_norm": 1.3551334142684937,
+      "learning_rate": 7.690215453003684e-05,
+      "loss": 1.4542,
+      "step": 1068
+    },
+    {
+      "epoch": 0.7094639311714097,
+      "grad_norm": 1.3880891799926758,
+      "learning_rate": 7.671736045502462e-05,
+      "loss": 1.7059,
+      "step": 1072
+    },
+    {
+      "epoch": 0.7121111846459298,
+      "grad_norm": 1.4413353204727173,
+      "learning_rate": 7.653205416914267e-05,
+      "loss": 1.5083,
+      "step": 1076
+    },
+    {
+      "epoch": 0.71475843812045,
+      "grad_norm": 1.3269354104995728,
+      "learning_rate": 7.634623922498348e-05,
+      "loss": 1.616,
+      "step": 1080
+    },
+    {
+      "epoch": 0.7174056915949703,
+      "grad_norm": 1.290656566619873,
+      "learning_rate": 7.615991918489125e-05,
+      "loss": 1.7394,
+      "step": 1084
+    },
+    {
+      "epoch": 0.7200529450694904,
+      "grad_norm": 1.2843822240829468,
+      "learning_rate": 7.597309762089359e-05,
+      "loss": 1.4916,
+      "step": 1088
+    },
+    {
+      "epoch": 0.7227001985440106,
+      "grad_norm": 1.2776360511779785,
+      "learning_rate": 7.57857781146331e-05,
+      "loss": 1.5602,
+      "step": 1092
+    },
+    {
+      "epoch": 0.7253474520185308,
+      "grad_norm": 1.2273900508880615,
+      "learning_rate": 7.559796425729863e-05,
+      "loss": 1.507,
+      "step": 1096
+    },
+    {
+      "epoch": 0.727994705493051,
+      "grad_norm": 1.425979495048523,
+      "learning_rate": 7.540965964955649e-05,
+      "loss": 1.6868,
+      "step": 1100
+    },
+    {
+      "epoch": 0.7306419589675711,
+      "grad_norm": 1.1589537858963013,
+      "learning_rate": 7.522086790148133e-05,
+      "loss": 1.7104,
+      "step": 1104
+    },
+    {
+      "epoch": 0.7332892124420913,
+      "grad_norm": 1.309058427810669,
+      "learning_rate": 7.503159263248709e-05,
+      "loss": 1.676,
+      "step": 1108
+    },
+    {
+      "epoch": 0.7359364659166115,
+      "grad_norm": 1.2926828861236572,
+      "learning_rate": 7.484183747125742e-05,
+      "loss": 1.4925,
+      "step": 1112
+    },
+    {
+      "epoch": 0.7385837193911317,
+      "grad_norm": 1.1898349523544312,
+      "learning_rate": 7.46516060556763e-05,
+      "loss": 1.5822,
+      "step": 1116
+    },
+    {
+      "epoch": 0.7412309728656519,
+      "grad_norm": 1.5024652481079102,
+      "learning_rate": 7.446090203275809e-05,
+      "loss": 1.6646,
+      "step": 1120
+    },
+    {
+      "epoch": 0.7438782263401721,
+      "grad_norm": 1.3592578172683716,
+      "learning_rate": 7.426972905857781e-05,
+      "loss": 1.5341,
+      "step": 1124
+    },
+    {
+      "epoch": 0.7465254798146923,
+      "grad_norm": 1.3298940658569336,
+      "learning_rate": 7.407809079820094e-05,
+      "loss": 1.7761,
+      "step": 1128
+    },
+    {
+      "epoch": 0.7491727332892124,
+      "grad_norm": 1.2832707166671753,
+      "learning_rate": 7.388599092561315e-05,
+      "loss": 1.6397,
+      "step": 1132
+    },
+    {
+      "epoch": 0.7518199867637326,
+      "grad_norm": 1.4493275880813599,
+      "learning_rate": 7.369343312364993e-05,
+      "loss": 1.5245,
+      "step": 1136
+    },
+    {
+      "epoch": 0.7544672402382528,
+      "grad_norm": 1.239708423614502,
+      "learning_rate": 7.350042108392594e-05,
+      "loss": 1.4402,
+      "step": 1140
+    },
+    {
+      "epoch": 0.757114493712773,
+      "grad_norm": 1.5889708995819092,
+      "learning_rate": 7.330695850676421e-05,
+      "loss": 1.5915,
+      "step": 1144
+    },
+    {
+      "epoch": 0.7597617471872932,
+      "grad_norm": 2.788710832595825,
+      "learning_rate": 7.311304910112525e-05,
+      "loss": 1.6522,
+      "step": 1148
+    },
+    {
+      "epoch": 0.7624090006618134,
+      "grad_norm": 1.4260072708129883,
+      "learning_rate": 7.291869658453594e-05,
+      "loss": 1.792,
+      "step": 1152
+    },
+    {
+      "epoch": 0.7650562541363336,
+      "grad_norm": 1.3381611108779907,
+      "learning_rate": 7.272390468301821e-05,
+      "loss": 1.6558,
+      "step": 1156
+    },
+    {
+      "epoch": 0.7677035076108537,
+      "grad_norm": 1.2346041202545166,
+      "learning_rate": 7.252867713101771e-05,
+      "loss": 1.3681,
+      "step": 1160
+    },
+    {
+      "epoch": 0.7703507610853739,
+      "grad_norm": 1.2784842252731323,
+      "learning_rate": 7.233301767133205e-05,
+      "loss": 1.5347,
+      "step": 1164
+    },
+    {
+      "epoch": 0.7729980145598941,
+      "grad_norm": 1.4147677421569824,
+      "learning_rate": 7.213693005503924e-05,
+      "loss": 1.6521,
+      "step": 1168
+    },
+    {
+      "epoch": 0.7756452680344142,
+      "grad_norm": 1.3260513544082642,
+      "learning_rate": 7.194041804142557e-05,
+      "loss": 1.7057,
+      "step": 1172
+    },
+    {
+      "epoch": 0.7782925215089345,
+      "grad_norm": 1.23674476146698,
+      "learning_rate": 7.174348539791375e-05,
+      "loss": 1.3705,
+      "step": 1176
+    },
+    {
+      "epoch": 0.7809397749834547,
+      "grad_norm": 1.3101705312728882,
+      "learning_rate": 7.154613589999054e-05,
+      "loss": 1.7117,
+      "step": 1180
+    },
+    {
+      "epoch": 0.7835870284579749,
+      "grad_norm": 1.287341594696045,
+      "learning_rate": 7.13483733311344e-05,
+      "loss": 1.4148,
+      "step": 1184
+    },
+    {
+      "epoch": 0.786234281932495,
+      "grad_norm": 1.5907158851623535,
+      "learning_rate": 7.115020148274295e-05,
+      "loss": 1.6679,
+      "step": 1188
+    },
+    {
+      "epoch": 0.7888815354070152,
+      "grad_norm": 1.3115376234054565,
+      "learning_rate": 7.095162415406034e-05,
+      "loss": 1.5572,
+      "step": 1192
+    },
+    {
+      "epoch": 0.7915287888815354,
+      "grad_norm": 1.3476297855377197,
+      "learning_rate": 7.075264515210435e-05,
+      "loss": 1.5828,
+      "step": 1196
+    },
+    {
+      "epoch": 0.7941760423560555,
+      "grad_norm": 1.3659868240356445,
+      "learning_rate": 7.055326829159341e-05,
+      "loss": 1.5468,
+      "step": 1200
+    },
+    {
+      "epoch": 0.7968232958305758,
+      "grad_norm": 1.4729256629943848,
+      "learning_rate": 7.03534973948735e-05,
+      "loss": 1.6008,
+      "step": 1204
+    },
+    {
+      "epoch": 0.799470549305096,
+      "grad_norm": 1.3182849884033203,
+      "learning_rate": 7.015333629184484e-05,
+      "loss": 1.5134,
+      "step": 1208
+    },
+    {
+      "epoch": 0.8021178027796162,
+      "grad_norm": 1.334165096282959,
+      "learning_rate": 6.995278881988847e-05,
+      "loss": 1.5952,
+      "step": 1212
+    },
+    {
+      "epoch": 0.8047650562541363,
+      "grad_norm": 1.3390055894851685,
+      "learning_rate": 6.975185882379271e-05,
+      "loss": 1.4782,
+      "step": 1216
+    },
+    {
+      "epoch": 0.8074123097286565,
+      "grad_norm": 1.3522272109985352,
+      "learning_rate": 6.955055015567942e-05,
+      "loss": 1.5075,
+      "step": 1220
+    },
+    {
+      "epoch": 0.8100595632031767,
+      "grad_norm": 4.2179059982299805,
+      "learning_rate": 6.934886667493012e-05,
+      "loss": 1.4707,
+      "step": 1224
+    },
+    {
+      "epoch": 0.8127068166776968,
+      "grad_norm": 1.2791635990142822,
+      "learning_rate": 6.914681224811208e-05,
+      "loss": 1.5571,
+      "step": 1228
+    },
+    {
+      "epoch": 0.8153540701522171,
+      "grad_norm": 1.2795437574386597,
+      "learning_rate": 6.894439074890414e-05,
+      "loss": 1.5616,
+      "step": 1232
+    },
+    {
+      "epoch": 0.8180013236267373,
+      "grad_norm": 1.3185795545578003,
+      "learning_rate": 6.874160605802244e-05,
+      "loss": 1.6333,
+      "step": 1236
+    },
+    {
+      "epoch": 0.8206485771012575,
+      "grad_norm": 1.3139797449111938,
+      "learning_rate": 6.853846206314605e-05,
+      "loss": 1.5663,
+      "step": 1240
+    },
+    {
+      "epoch": 0.8232958305757776,
+      "grad_norm": 1.2083790302276611,
+      "learning_rate": 6.833496265884241e-05,
+      "loss": 1.5103,
+      "step": 1244
+    },
+    {
+      "epoch": 0.8259430840502978,
+      "grad_norm": 1.5744483470916748,
+      "learning_rate": 6.813111174649269e-05,
+      "loss": 1.7545,
+      "step": 1248
+    },
+    {
+      "epoch": 0.828590337524818,
+      "grad_norm": 1.2797157764434814,
+      "learning_rate": 6.792691323421698e-05,
+      "loss": 1.5789,
+      "step": 1252
+    },
+    {
+      "epoch": 0.8312375909993381,
+      "grad_norm": 1.27366304397583,
+      "learning_rate": 6.772237103679937e-05,
+      "loss": 1.6271,
+      "step": 1256
+    },
+    {
+      "epoch": 0.8338848444738584,
+      "grad_norm": 1.2030060291290283,
+      "learning_rate": 6.751748907561288e-05,
+      "loss": 1.4017,
+      "step": 1260
+    },
+    {
+      "epoch": 0.8365320979483786,
+      "grad_norm": 1.2918180227279663,
+      "learning_rate": 6.731227127854434e-05,
+      "loss": 1.5686,
+      "step": 1264
+    },
+    {
+      "epoch": 0.8391793514228988,
+      "grad_norm": 1.3792778253555298,
+      "learning_rate": 6.710672157991899e-05,
+      "loss": 1.5969,
+      "step": 1268
+    },
+    {
+      "epoch": 0.8418266048974189,
+      "grad_norm": 1.349704623222351,
+      "learning_rate": 6.690084392042513e-05,
+      "loss": 1.4658,
+      "step": 1272
+    },
+    {
+      "epoch": 0.8444738583719391,
+      "grad_norm": 1.314712405204773,
+      "learning_rate": 6.669464224703861e-05,
+      "loss": 1.6031,
+      "step": 1276
+    },
+    {
+      "epoch": 0.8471211118464593,
+      "grad_norm": 1.2862118482589722,
+      "learning_rate": 6.648812051294697e-05,
+      "loss": 1.4582,
+      "step": 1280
+    },
+    {
+      "epoch": 0.8497683653209794,
+      "grad_norm": 1.249649167060852,
+      "learning_rate": 6.628128267747391e-05,
+      "loss": 1.5963,
+      "step": 1284
+    },
+    {
+      "epoch": 0.8524156187954997,
+      "grad_norm": 1.4206124544143677,
+      "learning_rate": 6.607413270600319e-05,
+      "loss": 1.631,
+      "step": 1288
+    },
+    {
+      "epoch": 0.8550628722700199,
+      "grad_norm": 1.3754596710205078,
+      "learning_rate": 6.586667456990267e-05,
+      "loss": 1.6588,
+      "step": 1292
+    },
+    {
+      "epoch": 0.8577101257445401,
+      "grad_norm": 1.375802993774414,
+      "learning_rate": 6.565891224644822e-05,
+      "loss": 1.528,
+      "step": 1296
+    },
+    {
+      "epoch": 0.8603573792190602,
+      "grad_norm": 1.3465059995651245,
+      "learning_rate": 6.545084971874738e-05,
+      "loss": 1.5359,
+      "step": 1300
+    },
+    {
+      "epoch": 0.8630046326935804,
+      "grad_norm": 1.3681912422180176,
+      "learning_rate": 6.524249097566306e-05,
+      "loss": 1.6129,
+      "step": 1304
+    },
+    {
+      "epoch": 0.8656518861681006,
+      "grad_norm": 1.216261625289917,
+      "learning_rate": 6.503384001173707e-05,
+      "loss": 1.3435,
+      "step": 1308
+    },
+    {
+      "epoch": 0.8682991396426207,
+      "grad_norm": 1.3217042684555054,
+      "learning_rate": 6.48249008271135e-05,
+      "loss": 1.5285,
+      "step": 1312
+    },
+    {
+      "epoch": 0.870946393117141,
+      "grad_norm": 1.3882989883422852,
+      "learning_rate": 6.461567742746206e-05,
+      "loss": 1.637,
+      "step": 1316
+    },
+    {
+      "epoch": 0.8735936465916612,
+      "grad_norm": 1.2400107383728027,
+      "learning_rate": 6.440617382390128e-05,
+      "loss": 1.5658,
+      "step": 1320
+    },
+    {
+      "epoch": 0.8762409000661814,
+      "grad_norm": 1.1711450815200806,
+      "learning_rate": 6.419639403292161e-05,
+      "loss": 1.6047,
+      "step": 1324
+    },
+    {
+      "epoch": 0.8788881535407015,
+      "grad_norm": 1.3696454763412476,
+      "learning_rate": 6.398634207630841e-05,
+      "loss": 1.5399,
+      "step": 1328
+    },
+    {
+      "epoch": 0.8815354070152217,
+      "grad_norm": 1.3397842645645142,
+      "learning_rate": 6.377602198106483e-05,
+      "loss": 1.5335,
+      "step": 1332
+    },
+    {
+      "epoch": 0.8841826604897419,
+      "grad_norm": 1.2820230722427368,
+      "learning_rate": 6.356543777933468e-05,
+      "loss": 1.4757,
+      "step": 1336
+    },
+    {
+      "epoch": 0.886829913964262,
+      "grad_norm": 1.4850640296936035,
+      "learning_rate": 6.335459350832504e-05,
+      "loss": 1.6376,
+      "step": 1340
+    },
+    {
+      "epoch": 0.8894771674387822,
+      "grad_norm": 1.2414631843566895,
+      "learning_rate": 6.314349321022893e-05,
+      "loss": 1.4272,
+      "step": 1344
+    },
+    {
+      "epoch": 0.8921244209133025,
+      "grad_norm": 1.4662775993347168,
+      "learning_rate": 6.293214093214775e-05,
+      "loss": 1.5005,
+      "step": 1348
+    },
+    {
+      "epoch": 0.8947716743878227,
+      "grad_norm": 1.3633265495300293,
+      "learning_rate": 6.272054072601374e-05,
+      "loss": 1.5595,
+      "step": 1352
+    },
+    {
+      "epoch": 0.8974189278623428,
+      "grad_norm": 1.1849807500839233,
+      "learning_rate": 6.250869664851227e-05,
+      "loss": 1.3661,
+      "step": 1356
+    },
+    {
+      "epoch": 0.900066181336863,
+      "grad_norm": 1.2869032621383667,
+      "learning_rate": 6.229661276100412e-05,
+      "loss": 1.4952,
+      "step": 1360
+    },
+    {
+      "epoch": 0.9027134348113832,
+      "grad_norm": 1.2344225645065308,
+      "learning_rate": 6.208429312944754e-05,
+      "loss": 1.4536,
+      "step": 1364
+    },
+    {
+      "epoch": 0.9053606882859033,
+      "grad_norm": 1.4070030450820923,
+      "learning_rate": 6.187174182432033e-05,
+      "loss": 1.5211,
+      "step": 1368
+    },
+    {
+      "epoch": 0.9080079417604235,
+      "grad_norm": 1.2222853899002075,
+      "learning_rate": 6.165896292054187e-05,
+      "loss": 1.5136,
+      "step": 1372
+    },
+    {
+      "epoch": 0.9106551952349438,
+      "grad_norm": 1.3191885948181152,
+      "learning_rate": 6.14459604973949e-05,
+      "loss": 1.4809,
+      "step": 1376
+    },
+    {
+      "epoch": 0.913302448709464,
+      "grad_norm": 1.2976018190383911,
+      "learning_rate": 6.12327386384473e-05,
+      "loss": 1.5621,
+      "step": 1380
+    },
+    {
+      "epoch": 0.9159497021839841,
+      "grad_norm": 1.3803764581680298,
+      "learning_rate": 6.101930143147395e-05,
+      "loss": 1.5469,
+      "step": 1384
+    },
+    {
+      "epoch": 0.9185969556585043,
+      "grad_norm": 1.3461982011795044,
+      "learning_rate": 6.080565296837821e-05,
+      "loss": 1.5434,
+      "step": 1388
+    },
+    {
+      "epoch": 0.9212442091330245,
+      "grad_norm": 1.2731950283050537,
+      "learning_rate": 6.059179734511356e-05,
+      "loss": 1.3728,
+      "step": 1392
+    },
+    {
+      "epoch": 0.9238914626075446,
+      "grad_norm": 1.2360509634017944,
+      "learning_rate": 6.037773866160502e-05,
+      "loss": 1.405,
+      "step": 1396
+    },
+    {
+      "epoch": 0.9265387160820648,
+      "grad_norm": 1.3289110660552979,
+      "learning_rate": 6.0163481021670575e-05,
+      "loss": 1.6851,
+      "step": 1400
+    },
+    {
+      "epoch": 0.9291859695565851,
+      "grad_norm": 1.2140709161758423,
+      "learning_rate": 5.994902853294251e-05,
+      "loss": 1.5109,
+      "step": 1404
+    },
+    {
+      "epoch": 0.9318332230311053,
+      "grad_norm": 1.364023208618164,
+      "learning_rate": 5.973438530678861e-05,
+      "loss": 1.613,
+      "step": 1408
+    },
+    {
+      "epoch": 0.9344804765056254,
+      "grad_norm": 1.3927541971206665,
+      "learning_rate": 5.951955545823342e-05,
+      "loss": 1.6401,
+      "step": 1412
+    },
+    {
+      "epoch": 0.9371277299801456,
+      "grad_norm": 1.2477508783340454,
+      "learning_rate": 5.930454310587929e-05,
+      "loss": 1.4617,
+      "step": 1416
+    },
+    {
+      "epoch": 0.9397749834546658,
+      "grad_norm": 1.455429196357727,
+      "learning_rate": 5.9089352371827446e-05,
+      "loss": 1.7028,
+      "step": 1420
+    },
+    {
+      "epoch": 0.9424222369291859,
+      "grad_norm": 1.379651665687561,
+      "learning_rate": 5.8873987381598924e-05,
+      "loss": 1.6448,
+      "step": 1424
+    },
+    {
+      "epoch": 0.9450694904037061,
+      "grad_norm": 1.2866801023483276,
+      "learning_rate": 5.865845226405553e-05,
+      "loss": 1.4807,
+      "step": 1428
+    },
+    {
+      "epoch": 0.9477167438782264,
+      "grad_norm": 2.283489465713501,
+      "learning_rate": 5.844275115132064e-05,
+      "loss": 1.5081,
+      "step": 1432
+    },
+    {
+      "epoch": 0.9503639973527466,
+      "grad_norm": 1.1935935020446777,
+      "learning_rate": 5.822688817870004e-05,
+      "loss": 1.5479,
+      "step": 1436
+    },
+    {
+      "epoch": 0.9530112508272667,
+      "grad_norm": 1.3389490842819214,
+      "learning_rate": 5.801086748460255e-05,
+      "loss": 1.5592,
+      "step": 1440
+    },
+    {
+      "epoch": 0.9556585043017869,
+      "grad_norm": 1.3386996984481812,
+      "learning_rate": 5.7794693210460804e-05,
+      "loss": 1.5736,
+      "step": 1444
+    },
+    {
+      "epoch": 0.9583057577763071,
+      "grad_norm": 1.3163249492645264,
+      "learning_rate": 5.757836950065172e-05,
+      "loss": 1.4719,
+      "step": 1448
+    },
+    {
+      "epoch": 0.9609530112508272,
+      "grad_norm": 1.3524895906448364,
+      "learning_rate": 5.736190050241719e-05,
+      "loss": 1.7056,
+      "step": 1452
+    },
+    {
+      "epoch": 0.9636002647253474,
+      "grad_norm": 1.2378168106079102,
+      "learning_rate": 5.714529036578443e-05,
+      "loss": 1.4294,
+      "step": 1456
+    },
+    {
+      "epoch": 0.9662475181998676,
+      "grad_norm": 1.353258728981018,
+      "learning_rate": 5.692854324348653e-05,
+      "loss": 1.5733,
+      "step": 1460
+    },
+    {
+      "epoch": 0.9688947716743879,
+      "grad_norm": 1.412914752960205,
+      "learning_rate": 5.6711663290882776e-05,
+      "loss": 1.5015,
+      "step": 1464
+    },
+    {
+      "epoch": 0.971542025148908,
+      "grad_norm": 1.400468349456787,
+      "learning_rate": 5.649465466587902e-05,
+      "loss": 1.6175,
+      "step": 1468
+    },
+    {
+      "epoch": 0.9741892786234282,
+      "grad_norm": 1.3348610401153564,
+      "learning_rate": 5.627752152884794e-05,
+      "loss": 1.5971,
+      "step": 1472
+    },
+    {
+      "epoch": 0.9768365320979484,
+      "grad_norm": 1.3182803392410278,
+      "learning_rate": 5.606026804254931e-05,
+      "loss": 1.4296,
+      "step": 1476
+    },
+    {
+      "epoch": 0.9794837855724685,
+      "grad_norm": 1.1478044986724854,
+      "learning_rate": 5.584289837205012e-05,
+      "loss": 1.3953,
+      "step": 1480
+    },
+    {
+      "epoch": 0.9821310390469887,
+      "grad_norm": 1.3956937789916992,
+      "learning_rate": 5.5625416684644874e-05,
+      "loss": 1.5031,
+      "step": 1484
+    },
+    {
+      "epoch": 0.984778292521509,
+      "grad_norm": 1.275169014930725,
+      "learning_rate": 5.540782714977549e-05,
+      "loss": 1.5208,
+      "step": 1488
+    },
+    {
+      "epoch": 0.9874255459960292,
+      "grad_norm": 1.2562490701675415,
+      "learning_rate": 5.51901339389516e-05,
+      "loss": 1.3706,
+      "step": 1492
+    },
+    {
+      "epoch": 0.9900727994705493,
+      "grad_norm": 1.2528316974639893,
+      "learning_rate": 5.4972341225670354e-05,
+      "loss": 1.5021,
+      "step": 1496
+    },
+    {
+      "epoch": 0.9927200529450695,
+      "grad_norm": 1.196968674659729,
+      "learning_rate": 5.4754453185336586e-05,
+      "loss": 1.5856,
+      "step": 1500
+    },
+    {
+      "epoch": 0.9953673064195897,
+      "grad_norm": 1.2576624155044556,
+      "learning_rate": 5.453647399518262e-05,
+      "loss": 1.3897,
+      "step": 1504
+    },
+    {
+      "epoch": 0.9980145598941098,
+      "grad_norm": 1.2169815301895142,
+      "learning_rate": 5.431840783418832e-05,
+      "loss": 1.3847,
+      "step": 1508
+    }
+  ],
+  "logging_steps": 4,
+  "max_steps": 3022,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 1511,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.5577041204504166e+17,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-1511/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42967f5abbe89175c93a82bf0011a7cbad7fd20693f8a95f4528737ccaf8045d
+size 5432

checkpoint-3022/config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "/mnt/parscratch/users/acp23ay/private/models/OLMo-2-1124-7B-szl_Latn-coli-cw2v/",
+  "architectures": [
+    "Olmo2ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "eos_token_id": 100257,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 4096,
+  "model_type": "olmo2",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "pad_token_id": 100277,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 500000,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.48.0.dev0",
+  "use_cache": true,
+  "vocab_size": 100384
+}

checkpoint-3022/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": 100257,
+  "pad_token_id": 100277,
+  "transformers_version": "4.48.0.dev0"
+}

checkpoint-3022/model-00001-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e6657750276ebbedd168b965b409b5f5f816f40d9e61319b7201312b91e3ea4
+size 4950070472

checkpoint-3022/model-00002-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de0b6a357d03676c4986588aaf4dc96afc3f44a0498514961d2b2cfa1940655e
+size 4991622816

checkpoint-3022/model-00003-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1038a7d346a681a8fb6949e054c2246f3f3c80e826855c1cec6af1155cc607d
+size 4924546896

checkpoint-3022/model-00004-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33c0a21146ace02d36928787ae7f34be868f7c81b894dea98ec7b0e5a3ea01b2
+size 4857404928

checkpoint-3022/model-00005-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7f82a3014ccfc0f67208f63785ae4bc71f09aba86db48bf6139d18178abbfed
+size 4857404928

checkpoint-3022/model-00006-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d71f8a8d07738eaf013a697ae5e979a146de61dbfb9a222f222a5740647d2406
+size 4614509080

checkpoint-3022/model.safetensors.index.json

@@ -0,0 +1,362 @@
+{
+  "metadata": {
+    "total_size": 29195517952
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00006-of-00006.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.norm.weight": "model-00006-of-00006.safetensors"
+  }
+}

checkpoint-3022/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:602649e7a0c50d583a6b5e0f741160349d17f906ce616b3bc852a0d9392f4406
+size 13055369358

checkpoint-3022/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ff264f99d31b522cc7e2a4eac9d38606d0c58a34c0adc74d71e0ca8b371dc36
+size 14244

checkpoint-3022/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03d0d2d27d290ea394fe2967fe1ee75ca9e753a7e55c04097547cd7579eb33ea
+size 1064

checkpoint-3022/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42967f5abbe89175c93a82bf0011a7cbad7fd20693f8a95f4528737ccaf8045d
+size 5432

config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "/mnt/parscratch/users/acp23ay/private/models/OLMo-2-1124-7B-szl_Latn-coli-cw2v/",
+  "architectures": [
+    "Olmo2ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "eos_token_id": 100257,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 4096,
+  "model_type": "olmo2",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "pad_token_id": 100277,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 500000,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.48.0.dev0",
+  "use_cache": true,
+  "vocab_size": 100384
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": 100257,
+  "pad_token_id": 100277,
+  "transformers_version": "4.48.0.dev0"
+}

model-00001-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e6657750276ebbedd168b965b409b5f5f816f40d9e61319b7201312b91e3ea4
+size 4950070472

model-00002-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de0b6a357d03676c4986588aaf4dc96afc3f44a0498514961d2b2cfa1940655e
+size 4991622816

model-00003-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1038a7d346a681a8fb6949e054c2246f3f3c80e826855c1cec6af1155cc607d
+size 4924546896

model-00004-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33c0a21146ace02d36928787ae7f34be868f7c81b894dea98ec7b0e5a3ea01b2
+size 4857404928

model-00005-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7f82a3014ccfc0f67208f63785ae4bc71f09aba86db48bf6139d18178abbfed
+size 4857404928

model-00006-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d71f8a8d07738eaf013a697ae5e979a146de61dbfb9a222f222a5740647d2406
+size 4614509080

model.safetensors.index.json

@@ -0,0 +1,362 @@
+{
+  "metadata": {
+    "total_size": 29195517952
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00006-of-00006.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.post_feedforward_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.16.self_attn.k_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.q_norm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.post_feedforward_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.22.self_attn.k_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.q_norm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.post_feedforward_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.28.self_attn.k_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.q_norm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.post_feedforward_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.k_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.q_norm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.post_feedforward_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.k_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.q_norm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.post_feedforward_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.k_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.q_norm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.norm.weight": "model-00006-of-00006.safetensors"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|pad|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,189 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "100256": {
+      "content": "<|extra_id_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100257": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100258": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100259": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100260": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100261": {
+      "content": "|||PHONE_NUMBER|||",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100262": {
+      "content": "|||EMAIL_ADDRESS|||",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100263": {
+      "content": "|||IP_ADDRESS|||",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100264": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100265": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100266": {
+      "content": "<|extra_id_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100267": {
+      "content": "<|extra_id_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100268": {
+      "content": "<|extra_id_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100269": {
+      "content": "<|extra_id_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100270": {
+      "content": "<|extra_id_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100271": {
+      "content": "<|extra_id_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100272": {
+      "content": "<|extra_id_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100273": {
+      "content": "<|extra_id_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100274": {
+      "content": "<|extra_id_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100275": {
+      "content": "<|extra_id_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100276": {
+      "content": "<|endofprompt|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100277": {
+      "content": "<|pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<|pad|>",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42967f5abbe89175c93a82bf0011a7cbad7fd20693f8a95f4528737ccaf8045d
+size 5432