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-mean/",
+  "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 sha256:797a6f3258a078a1f3d61de3a5d83a1999b6fdb94b0d16b53f60704ee94e0713
+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 sha256:98f60e1221de97c1a89f0b67c0d81d09b91ce1151d4b08b963c789267ff60882
+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 sha256:7c92a346705ab16e08a3e71db62e1fbf9b66245269da27e77ad79bd67681bb2b
+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": 24.50491714477539,
+      "learning_rate": 2.631578947368421e-06,
+      "loss": 6.5473,
+      "step": 4
+    },
+    {
+      "epoch": 0.005294506949040371,
+      "grad_norm": 17.426511764526367,
+      "learning_rate": 5.263157894736842e-06,
+      "loss": 6.2116,
+      "step": 8
+    },
+    {
+      "epoch": 0.007941760423560556,
+      "grad_norm": 6.35976505279541,
+      "learning_rate": 7.894736842105263e-06,
+      "loss": 5.7967,
+      "step": 12
+    },
+    {
+      "epoch": 0.010589013898080741,
+      "grad_norm": 5.454939842224121,
+      "learning_rate": 1.0526315789473684e-05,
+      "loss": 5.3365,
+      "step": 16
+    },
+    {
+      "epoch": 0.013236267372600927,
+      "grad_norm": 4.607099533081055,
+      "learning_rate": 1.3157894736842106e-05,
+      "loss": 4.7105,
+      "step": 20
+    },
+    {
+      "epoch": 0.01588352084712111,
+      "grad_norm": 3.6498019695281982,
+      "learning_rate": 1.5789473684210526e-05,
+      "loss": 4.286,
+      "step": 24
+    },
+    {
+      "epoch": 0.018530774321641297,
+      "grad_norm": 4.196900844573975,
+      "learning_rate": 1.8421052631578947e-05,
+      "loss": 4.3134,
+      "step": 28
+    },
+    {
+      "epoch": 0.021178027796161483,
+      "grad_norm": 3.617469072341919,
+      "learning_rate": 2.105263157894737e-05,
+      "loss": 3.7494,
+      "step": 32
+    },
+    {
+      "epoch": 0.02382528127068167,
+      "grad_norm": 3.05267333984375,
+      "learning_rate": 2.368421052631579e-05,
+      "loss": 3.8046,
+      "step": 36
+    },
+    {
+      "epoch": 0.026472534745201854,
+      "grad_norm": 2.607614517211914,
+      "learning_rate": 2.6315789473684212e-05,
+      "loss": 3.385,
+      "step": 40
+    },
+    {
+      "epoch": 0.02911978821972204,
+      "grad_norm": 2.536888837814331,
+      "learning_rate": 2.8947368421052634e-05,
+      "loss": 3.3516,
+      "step": 44
+    },
+    {
+      "epoch": 0.03176704169424222,
+      "grad_norm": 2.315871000289917,
+      "learning_rate": 3.157894736842105e-05,
+      "loss": 3.0795,
+      "step": 48
+    },
+    {
+      "epoch": 0.03441429516876241,
+      "grad_norm": 2.3058571815490723,
+      "learning_rate": 3.421052631578947e-05,
+      "loss": 3.0708,
+      "step": 52
+    },
+    {
+      "epoch": 0.037061548643282594,
+      "grad_norm": 2.067796230316162,
+      "learning_rate": 3.6842105263157895e-05,
+      "loss": 2.8311,
+      "step": 56
+    },
+    {
+      "epoch": 0.03970880211780278,
+      "grad_norm": 1.9578440189361572,
+      "learning_rate": 3.9473684210526316e-05,
+      "loss": 2.696,
+      "step": 60
+    },
+    {
+      "epoch": 0.042356055592322965,
+      "grad_norm": 2.043933629989624,
+      "learning_rate": 4.210526315789474e-05,
+      "loss": 2.7501,
+      "step": 64
+    },
+    {
+      "epoch": 0.04500330906684315,
+      "grad_norm": 1.82830810546875,
+      "learning_rate": 4.473684210526316e-05,
+      "loss": 2.5058,
+      "step": 68
+    },
+    {
+      "epoch": 0.04765056254136334,
+      "grad_norm": 1.8841806650161743,
+      "learning_rate": 4.736842105263158e-05,
+      "loss": 2.5708,
+      "step": 72
+    },
+    {
+      "epoch": 0.05029781601588352,
+      "grad_norm": 1.9775539636611938,
+      "learning_rate": 5e-05,
+      "loss": 2.6332,
+      "step": 76
+    },
+    {
+      "epoch": 0.05294506949040371,
+      "grad_norm": 1.7908610105514526,
+      "learning_rate": 5.2631578947368424e-05,
+      "loss": 2.5441,
+      "step": 80
+    },
+    {
+      "epoch": 0.05559232296492389,
+      "grad_norm": 1.977647066116333,
+      "learning_rate": 5.526315789473685e-05,
+      "loss": 2.3617,
+      "step": 84
+    },
+    {
+      "epoch": 0.05823957643944408,
+      "grad_norm": 2.008470296859741,
+      "learning_rate": 5.789473684210527e-05,
+      "loss": 2.3994,
+      "step": 88
+    },
+    {
+      "epoch": 0.06088682991396426,
+      "grad_norm": 2.070720911026001,
+      "learning_rate": 6.052631578947369e-05,
+      "loss": 2.3509,
+      "step": 92
+    },
+    {
+      "epoch": 0.06353408338848444,
+      "grad_norm": 2.0442869663238525,
+      "learning_rate": 6.31578947368421e-05,
+      "loss": 2.35,
+      "step": 96
+    },
+    {
+      "epoch": 0.06618133686300463,
+      "grad_norm": 1.8274725675582886,
+      "learning_rate": 6.578947368421054e-05,
+      "loss": 2.2802,
+      "step": 100
+    },
+    {
+      "epoch": 0.06882859033752482,
+      "grad_norm": 1.9744892120361328,
+      "learning_rate": 6.842105263157895e-05,
+      "loss": 2.4711,
+      "step": 104
+    },
+    {
+      "epoch": 0.071475843812045,
+      "grad_norm": 1.881946086883545,
+      "learning_rate": 7.105263157894737e-05,
+      "loss": 2.3495,
+      "step": 108
+    },
+    {
+      "epoch": 0.07412309728656519,
+      "grad_norm": 1.7632906436920166,
+      "learning_rate": 7.368421052631579e-05,
+      "loss": 2.1906,
+      "step": 112
+    },
+    {
+      "epoch": 0.07677035076108538,
+      "grad_norm": 1.8465447425842285,
+      "learning_rate": 7.631578947368422e-05,
+      "loss": 2.4193,
+      "step": 116
+    },
+    {
+      "epoch": 0.07941760423560557,
+      "grad_norm": 1.978273868560791,
+      "learning_rate": 7.894736842105263e-05,
+      "loss": 2.3229,
+      "step": 120
+    },
+    {
+      "epoch": 0.08206485771012574,
+      "grad_norm": 1.9878270626068115,
+      "learning_rate": 8.157894736842105e-05,
+      "loss": 2.3028,
+      "step": 124
+    },
+    {
+      "epoch": 0.08471211118464593,
+      "grad_norm": 1.7065322399139404,
+      "learning_rate": 8.421052631578948e-05,
+      "loss": 2.244,
+      "step": 128
+    },
+    {
+      "epoch": 0.08735936465916612,
+      "grad_norm": 1.8170701265335083,
+      "learning_rate": 8.68421052631579e-05,
+      "loss": 2.1112,
+      "step": 132
+    },
+    {
+      "epoch": 0.0900066181336863,
+      "grad_norm": 1.9288476705551147,
+      "learning_rate": 8.947368421052632e-05,
+      "loss": 2.3551,
+      "step": 136
+    },
+    {
+      "epoch": 0.09265387160820648,
+      "grad_norm": 1.8695253133773804,
+      "learning_rate": 9.210526315789474e-05,
+      "loss": 2.2814,
+      "step": 140
+    },
+    {
+      "epoch": 0.09530112508272667,
+      "grad_norm": 1.7066093683242798,
+      "learning_rate": 9.473684210526316e-05,
+      "loss": 1.9036,
+      "step": 144
+    },
+    {
+      "epoch": 0.09794837855724686,
+      "grad_norm": 1.8588757514953613,
+      "learning_rate": 9.736842105263158e-05,
+      "loss": 2.0139,
+      "step": 148
+    },
+    {
+      "epoch": 0.10059563203176704,
+      "grad_norm": 1.789518117904663,
+      "learning_rate": 0.0001,
+      "loss": 2.1809,
+      "step": 152
+    },
+    {
+      "epoch": 0.10324288550628723,
+      "grad_norm": 1.9242740869522095,
+      "learning_rate": 9.999952071344157e-05,
+      "loss": 2.301,
+      "step": 156
+    },
+    {
+      "epoch": 0.10589013898080742,
+      "grad_norm": 1.7974549531936646,
+      "learning_rate": 9.999808286295485e-05,
+      "loss": 2.2312,
+      "step": 160
+    },
+    {
+      "epoch": 0.10853739245532759,
+      "grad_norm": 1.7276233434677124,
+      "learning_rate": 9.999568647610555e-05,
+      "loss": 2.1109,
+      "step": 164
+    },
+    {
+      "epoch": 0.11118464592984778,
+      "grad_norm": 1.8286519050598145,
+      "learning_rate": 9.999233159883593e-05,
+      "loss": 2.0782,
+      "step": 168
+    },
+    {
+      "epoch": 0.11383189940436797,
+      "grad_norm": 1.919313907623291,
+      "learning_rate": 9.998801829546386e-05,
+      "loss": 2.0693,
+      "step": 172
+    },
+    {
+      "epoch": 0.11647915287888816,
+      "grad_norm": 1.6544960737228394,
+      "learning_rate": 9.998274664868173e-05,
+      "loss": 2.0982,
+      "step": 176
+    },
+    {
+      "epoch": 0.11912640635340833,
+      "grad_norm": 1.8223872184753418,
+      "learning_rate": 9.997651675955466e-05,
+      "loss": 2.1379,
+      "step": 180
+    },
+    {
+      "epoch": 0.12177365982792852,
+      "grad_norm": 1.7743052244186401,
+      "learning_rate": 9.996932874751877e-05,
+      "loss": 2.0637,
+      "step": 184
+    },
+    {
+      "epoch": 0.12442091330244871,
+      "grad_norm": 1.7228261232376099,
+      "learning_rate": 9.996118275037873e-05,
+      "loss": 2.1696,
+      "step": 188
+    },
+    {
+      "epoch": 0.1270681667769689,
+      "grad_norm": 1.6266913414001465,
+      "learning_rate": 9.995207892430524e-05,
+      "loss": 2.1247,
+      "step": 192
+    },
+    {
+      "epoch": 0.12971542025148908,
+      "grad_norm": 1.8206615447998047,
+      "learning_rate": 9.994201744383196e-05,
+      "loss": 2.1831,
+      "step": 196
+    },
+    {
+      "epoch": 0.13236267372600927,
+      "grad_norm": 1.943579912185669,
+      "learning_rate": 9.993099850185216e-05,
+      "loss": 1.9262,
+      "step": 200
+    },
+    {
+      "epoch": 0.13500992720052946,
+      "grad_norm": 1.89098060131073,
+      "learning_rate": 9.991902230961511e-05,
+      "loss": 2.2636,
+      "step": 204
+    },
+    {
+      "epoch": 0.13765718067504965,
+      "grad_norm": 1.8418017625808716,
+      "learning_rate": 9.99060890967219e-05,
+      "loss": 2.2454,
+      "step": 208
+    },
+    {
+      "epoch": 0.14030443414956983,
+      "grad_norm": 1.7433375120162964,
+      "learning_rate": 9.989219911112113e-05,
+      "loss": 2.2591,
+      "step": 212
+    },
+    {
+      "epoch": 0.14295168762409,
+      "grad_norm": 1.885964035987854,
+      "learning_rate": 9.987735261910417e-05,
+      "loss": 1.9402,
+      "step": 216
+    },
+    {
+      "epoch": 0.14559894109861019,
+      "grad_norm": 1.626397728919983,
+      "learning_rate": 9.986154990529995e-05,
+      "loss": 2.119,
+      "step": 220
+    },
+    {
+      "epoch": 0.14824619457313037,
+      "grad_norm": 1.5490047931671143,
+      "learning_rate": 9.984479127266961e-05,
+      "loss": 1.8635,
+      "step": 224
+    },
+    {
+      "epoch": 0.15089344804765056,
+      "grad_norm": 1.5588316917419434,
+      "learning_rate": 9.982707704250065e-05,
+      "loss": 1.8135,
+      "step": 228
+    },
+    {
+      "epoch": 0.15354070152217075,
+      "grad_norm": 1.9416462182998657,
+      "learning_rate": 9.980840755440075e-05,
+      "loss": 2.2288,
+      "step": 232
+    },
+    {
+      "epoch": 0.15618795499669094,
+      "grad_norm": 1.5774728059768677,
+      "learning_rate": 9.978878316629133e-05,
+      "loss": 1.9254,
+      "step": 236
+    },
+    {
+      "epoch": 0.15883520847121113,
+      "grad_norm": 1.6661707162857056,
+      "learning_rate": 9.976820425440058e-05,
+      "loss": 2.0111,
+      "step": 240
+    },
+    {
+      "epoch": 0.1614824619457313,
+      "grad_norm": 1.5805509090423584,
+      "learning_rate": 9.974667121325634e-05,
+      "loss": 2.0657,
+      "step": 244
+    },
+    {
+      "epoch": 0.16412971542025148,
+      "grad_norm": 1.7854478359222412,
+      "learning_rate": 9.972418445567851e-05,
+      "loss": 1.8586,
+      "step": 248
+    },
+    {
+      "epoch": 0.16677696889477167,
+      "grad_norm": 1.61441171169281,
+      "learning_rate": 9.97007444127711e-05,
+      "loss": 1.9234,
+      "step": 252
+    },
+    {
+      "epoch": 0.16942422236929186,
+      "grad_norm": 2.154454469680786,
+      "learning_rate": 9.967635153391401e-05,
+      "loss": 1.949,
+      "step": 256
+    },
+    {
+      "epoch": 0.17207147584381205,
+      "grad_norm": 1.5182636976242065,
+      "learning_rate": 9.965100628675441e-05,
+      "loss": 2.013,
+      "step": 260
+    },
+    {
+      "epoch": 0.17471872931833224,
+      "grad_norm": 1.751714825630188,
+      "learning_rate": 9.962470915719775e-05,
+      "loss": 1.9629,
+      "step": 264
+    },
+    {
+      "epoch": 0.17736598279285243,
+      "grad_norm": 1.5807703733444214,
+      "learning_rate": 9.959746064939846e-05,
+      "loss": 1.8705,
+      "step": 268
+    },
+    {
+      "epoch": 0.1800132362673726,
+      "grad_norm": 1.7142225503921509,
+      "learning_rate": 9.956926128575026e-05,
+      "loss": 2.0033,
+      "step": 272
+    },
+    {
+      "epoch": 0.18266048974189278,
+      "grad_norm": 1.555530309677124,
+      "learning_rate": 9.954011160687622e-05,
+      "loss": 1.8995,
+      "step": 276
+    },
+    {
+      "epoch": 0.18530774321641297,
+      "grad_norm": 1.5679693222045898,
+      "learning_rate": 9.951001217161829e-05,
+      "loss": 2.042,
+      "step": 280
+    },
+    {
+      "epoch": 0.18795499669093316,
+      "grad_norm": 1.6399418115615845,
+      "learning_rate": 9.947896355702666e-05,
+      "loss": 2.0388,
+      "step": 284
+    },
+    {
+      "epoch": 0.19060225016545335,
+      "grad_norm": 1.7505602836608887,
+      "learning_rate": 9.944696635834867e-05,
+      "loss": 1.9648,
+      "step": 288
+    },
+    {
+      "epoch": 0.19324950363997354,
+      "grad_norm": 1.4888848066329956,
+      "learning_rate": 9.941402118901744e-05,
+      "loss": 1.8595,
+      "step": 292
+    },
+    {
+      "epoch": 0.19589675711449372,
+      "grad_norm": 1.4739277362823486,
+      "learning_rate": 9.938012868064e-05,
+      "loss": 1.7959,
+      "step": 296
+    },
+    {
+      "epoch": 0.1985440105890139,
+      "grad_norm": 1.5393471717834473,
+      "learning_rate": 9.934528948298533e-05,
+      "loss": 1.9469,
+      "step": 300
+    },
+    {
+      "epoch": 0.20119126406353408,
+      "grad_norm": 1.5673627853393555,
+      "learning_rate": 9.930950426397179e-05,
+      "loss": 1.9332,
+      "step": 304
+    },
+    {
+      "epoch": 0.20383851753805426,
+      "grad_norm": 1.6461111307144165,
+      "learning_rate": 9.927277370965435e-05,
+      "loss": 1.8055,
+      "step": 308
+    },
+    {
+      "epoch": 0.20648577101257445,
+      "grad_norm": 1.5950462818145752,
+      "learning_rate": 9.923509852421145e-05,
+      "loss": 1.8414,
+      "step": 312
+    },
+    {
+      "epoch": 0.20913302448709464,
+      "grad_norm": 1.433727741241455,
+      "learning_rate": 9.919647942993148e-05,
+      "loss": 1.9514,
+      "step": 316
+    },
+    {
+      "epoch": 0.21178027796161483,
+      "grad_norm": 1.445776343345642,
+      "learning_rate": 9.915691716719898e-05,
+      "loss": 1.7297,
+      "step": 320
+    },
+    {
+      "epoch": 0.21442753143613502,
+      "grad_norm": 1.9325745105743408,
+      "learning_rate": 9.911641249448036e-05,
+      "loss": 1.9855,
+      "step": 324
+    },
+    {
+      "epoch": 0.21707478491065518,
+      "grad_norm": 1.494813323020935,
+      "learning_rate": 9.907496618830942e-05,
+      "loss": 1.7916,
+      "step": 328
+    },
+    {
+      "epoch": 0.21972203838517537,
+      "grad_norm": 1.4863932132720947,
+      "learning_rate": 9.903257904327249e-05,
+      "loss": 1.8029,
+      "step": 332
+    },
+    {
+      "epoch": 0.22236929185969556,
+      "grad_norm": 1.594827651977539,
+      "learning_rate": 9.898925187199308e-05,
+      "loss": 1.9516,
+      "step": 336
+    },
+    {
+      "epoch": 0.22501654533421575,
+      "grad_norm": 1.5738781690597534,
+      "learning_rate": 9.894498550511646e-05,
+      "loss": 1.8997,
+      "step": 340
+    },
+    {
+      "epoch": 0.22766379880873594,
+      "grad_norm": 1.5598024129867554,
+      "learning_rate": 9.88997807912936e-05,
+      "loss": 1.9249,
+      "step": 344
+    },
+    {
+      "epoch": 0.23031105228325613,
+      "grad_norm": 1.4761321544647217,
+      "learning_rate": 9.885363859716497e-05,
+      "loss": 1.7571,
+      "step": 348
+    },
+    {
+      "epoch": 0.23295830575777632,
+      "grad_norm": 1.4266904592514038,
+      "learning_rate": 9.88065598073439e-05,
+      "loss": 1.9811,
+      "step": 352
+    },
+    {
+      "epoch": 0.23560555923229648,
+      "grad_norm": 1.5371057987213135,
+      "learning_rate": 9.875854532439964e-05,
+      "loss": 1.8021,
+      "step": 356
+    },
+    {
+      "epoch": 0.23825281270681667,
+      "grad_norm": 1.380096673965454,
+      "learning_rate": 9.870959606884004e-05,
+      "loss": 1.8223,
+      "step": 360
+    },
+    {
+      "epoch": 0.24090006618133686,
+      "grad_norm": 1.632664680480957,
+      "learning_rate": 9.865971297909393e-05,
+      "loss": 2.006,
+      "step": 364
+    },
+    {
+      "epoch": 0.24354731965585705,
+      "grad_norm": 1.3765276670455933,
+      "learning_rate": 9.860889701149307e-05,
+      "loss": 1.7893,
+      "step": 368
+    },
+    {
+      "epoch": 0.24619457313037724,
+      "grad_norm": 1.5789958238601685,
+      "learning_rate": 9.855714914025384e-05,
+      "loss": 1.9381,
+      "step": 372
+    },
+    {
+      "epoch": 0.24884182660489743,
+      "grad_norm": 1.8294042348861694,
+      "learning_rate": 9.850447035745866e-05,
+      "loss": 1.8584,
+      "step": 376
+    },
+    {
+      "epoch": 0.2514890800794176,
+      "grad_norm": 1.5388972759246826,
+      "learning_rate": 9.845086167303679e-05,
+      "loss": 1.8763,
+      "step": 380
+    },
+    {
+      "epoch": 0.2541363335539378,
+      "grad_norm": 1.5301390886306763,
+      "learning_rate": 9.839632411474513e-05,
+      "loss": 2.0612,
+      "step": 384
+    },
+    {
+      "epoch": 0.256783587028458,
+      "grad_norm": 1.546277642250061,
+      "learning_rate": 9.83408587281484e-05,
+      "loss": 1.9085,
+      "step": 388
+    },
+    {
+      "epoch": 0.25943084050297816,
+      "grad_norm": 1.5818853378295898,
+      "learning_rate": 9.828446657659918e-05,
+      "loss": 2.0181,
+      "step": 392
+    },
+    {
+      "epoch": 0.26207809397749837,
+      "grad_norm": 1.2648255825042725,
+      "learning_rate": 9.82271487412175e-05,
+      "loss": 1.6947,
+      "step": 396
+    },
+    {
+      "epoch": 0.26472534745201853,
+      "grad_norm": 1.541934847831726,
+      "learning_rate": 9.816890632087006e-05,
+      "loss": 1.8053,
+      "step": 400
+    },
+    {
+      "epoch": 0.2673726009265387,
+      "grad_norm": 1.5966472625732422,
+      "learning_rate": 9.810974043214922e-05,
+      "loss": 1.8733,
+      "step": 404
+    },
+    {
+      "epoch": 0.2700198544010589,
+      "grad_norm": 1.5871154069900513,
+      "learning_rate": 9.804965220935161e-05,
+      "loss": 1.896,
+      "step": 408
+    },
+    {
+      "epoch": 0.2726671078755791,
+      "grad_norm": 1.4850573539733887,
+      "learning_rate": 9.798864280445632e-05,
+      "loss": 1.8494,
+      "step": 412
+    },
+    {
+      "epoch": 0.2753143613500993,
+      "grad_norm": 1.4737725257873535,
+      "learning_rate": 9.792671338710285e-05,
+      "loss": 1.8145,
+      "step": 416
+    },
+    {
+      "epoch": 0.27796161482461945,
+      "grad_norm": 1.5895408391952515,
+      "learning_rate": 9.786386514456872e-05,
+      "loss": 1.9279,
+      "step": 420
+    },
+    {
+      "epoch": 0.28060886829913967,
+      "grad_norm": 1.522838830947876,
+      "learning_rate": 9.780009928174661e-05,
+      "loss": 1.9103,
+      "step": 424
+    },
+    {
+      "epoch": 0.28325612177365983,
+      "grad_norm": 1.4890238046646118,
+      "learning_rate": 9.773541702112137e-05,
+      "loss": 1.9306,
+      "step": 428
+    },
+    {
+      "epoch": 0.28590337524818,
+      "grad_norm": 1.5047945976257324,
+      "learning_rate": 9.766981960274653e-05,
+      "loss": 1.8459,
+      "step": 432
+    },
+    {
+      "epoch": 0.2885506287227002,
+      "grad_norm": 1.4997539520263672,
+      "learning_rate": 9.760330828422053e-05,
+      "loss": 1.7442,
+      "step": 436
+    },
+    {
+      "epoch": 0.29119788219722037,
+      "grad_norm": 1.389294981956482,
+      "learning_rate": 9.753588434066258e-05,
+      "loss": 1.9077,
+      "step": 440
+    },
+    {
+      "epoch": 0.2938451356717406,
+      "grad_norm": 1.3641945123672485,
+      "learning_rate": 9.746754906468832e-05,
+      "loss": 1.8979,
+      "step": 444
+    },
+    {
+      "epoch": 0.29649238914626075,
+      "grad_norm": 1.5315138101577759,
+      "learning_rate": 9.73983037663849e-05,
+      "loss": 1.8207,
+      "step": 448
+    },
+    {
+      "epoch": 0.29913964262078097,
+      "grad_norm": 1.5057647228240967,
+      "learning_rate": 9.732814977328592e-05,
+      "loss": 1.911,
+      "step": 452
+    },
+    {
+      "epoch": 0.3017868960953011,
+      "grad_norm": 1.368912696838379,
+      "learning_rate": 9.725708843034605e-05,
+      "loss": 1.8377,
+      "step": 456
+    },
+    {
+      "epoch": 0.3044341495698213,
+      "grad_norm": 1.389817714691162,
+      "learning_rate": 9.718512109991514e-05,
+      "loss": 1.7907,
+      "step": 460
+    },
+    {
+      "epoch": 0.3070814030443415,
+      "grad_norm": 1.7318735122680664,
+      "learning_rate": 9.711224916171215e-05,
+      "loss": 1.9412,
+      "step": 464
+    },
+    {
+      "epoch": 0.30972865651886167,
+      "grad_norm": 1.4791710376739502,
+      "learning_rate": 9.703847401279871e-05,
+      "loss": 1.7754,
+      "step": 468
+    },
+    {
+      "epoch": 0.3123759099933819,
+      "grad_norm": 1.3618526458740234,
+      "learning_rate": 9.69637970675523e-05,
+      "loss": 1.73,
+      "step": 472
+    },
+    {
+      "epoch": 0.31502316346790205,
+      "grad_norm": 1.5649083852767944,
+      "learning_rate": 9.688821975763918e-05,
+      "loss": 1.9635,
+      "step": 476
+    },
+    {
+      "epoch": 0.31767041694242226,
+      "grad_norm": 1.3701534271240234,
+      "learning_rate": 9.681174353198687e-05,
+      "loss": 1.6581,
+      "step": 480
+    },
+    {
+      "epoch": 0.3203176704169424,
+      "grad_norm": 1.4764872789382935,
+      "learning_rate": 9.673436985675645e-05,
+      "loss": 1.794,
+      "step": 484
+    },
+    {
+      "epoch": 0.3229649238914626,
+      "grad_norm": 1.4432624578475952,
+      "learning_rate": 9.665610021531447e-05,
+      "loss": 1.9016,
+      "step": 488
+    },
+    {
+      "epoch": 0.3256121773659828,
+      "grad_norm": 1.572975993156433,
+      "learning_rate": 9.657693610820437e-05,
+      "loss": 2.035,
+      "step": 492
+    },
+    {
+      "epoch": 0.32825943084050296,
+      "grad_norm": 1.5382163524627686,
+      "learning_rate": 9.649687905311785e-05,
+      "loss": 2.0041,
+      "step": 496
+    },
+    {
+      "epoch": 0.3309066843150232,
+      "grad_norm": 1.3413423299789429,
+      "learning_rate": 9.641593058486574e-05,
+      "loss": 1.7448,
+      "step": 500
+    },
+    {
+      "epoch": 0.33355393778954334,
+      "grad_norm": 1.4374409914016724,
+      "learning_rate": 9.633409225534855e-05,
+      "loss": 1.7816,
+      "step": 504
+    },
+    {
+      "epoch": 0.33620119126406356,
+      "grad_norm": 1.4096835851669312,
+      "learning_rate": 9.625136563352671e-05,
+      "loss": 1.772,
+      "step": 508
+    },
+    {
+      "epoch": 0.3388484447385837,
+      "grad_norm": 2.1890769004821777,
+      "learning_rate": 9.616775230539057e-05,
+      "loss": 1.8641,
+      "step": 512
+    },
+    {
+      "epoch": 0.3414956982131039,
+      "grad_norm": 1.4621169567108154,
+      "learning_rate": 9.608325387392986e-05,
+      "loss": 1.7406,
+      "step": 516
+    },
+    {
+      "epoch": 0.3441429516876241,
+      "grad_norm": 1.4140963554382324,
+      "learning_rate": 9.599787195910313e-05,
+      "loss": 1.6127,
+      "step": 520
+    },
+    {
+      "epoch": 0.34679020516214426,
+      "grad_norm": 1.459409236907959,
+      "learning_rate": 9.591160819780649e-05,
+      "loss": 1.7579,
+      "step": 524
+    },
+    {
+      "epoch": 0.3494374586366645,
+      "grad_norm": 1.7444220781326294,
+      "learning_rate": 9.582446424384242e-05,
+      "loss": 1.8177,
+      "step": 528
+    },
+    {
+      "epoch": 0.35208471211118464,
+      "grad_norm": 1.4114232063293457,
+      "learning_rate": 9.573644176788794e-05,
+      "loss": 1.7955,
+      "step": 532
+    },
+    {
+      "epoch": 0.35473196558570486,
+      "grad_norm": 1.4076716899871826,
+      "learning_rate": 9.564754245746264e-05,
+      "loss": 1.9122,
+      "step": 536
+    },
+    {
+      "epoch": 0.357379219060225,
+      "grad_norm": 1.4209445714950562,
+      "learning_rate": 9.555776801689632e-05,
+      "loss": 1.8108,
+      "step": 540
+    },
+    {
+      "epoch": 0.3600264725347452,
+      "grad_norm": 1.5626829862594604,
+      "learning_rate": 9.546712016729624e-05,
+      "loss": 1.9285,
+      "step": 544
+    },
+    {
+      "epoch": 0.3626737260092654,
+      "grad_norm": 1.4253438711166382,
+      "learning_rate": 9.537560064651427e-05,
+      "loss": 1.6505,
+      "step": 548
+    },
+    {
+      "epoch": 0.36532097948378556,
+      "grad_norm": 1.447141170501709,
+      "learning_rate": 9.528321120911346e-05,
+      "loss": 1.8303,
+      "step": 552
+    },
+    {
+      "epoch": 0.3679682329583058,
+      "grad_norm": 1.4913408756256104,
+      "learning_rate": 9.51899536263344e-05,
+      "loss": 1.8382,
+      "step": 556
+    },
+    {
+      "epoch": 0.37061548643282594,
+      "grad_norm": 1.5191394090652466,
+      "learning_rate": 9.509582968606136e-05,
+      "loss": 1.7477,
+      "step": 560
+    },
+    {
+      "epoch": 0.37326273990734615,
+      "grad_norm": 1.3612414598464966,
+      "learning_rate": 9.500084119278788e-05,
+      "loss": 1.7101,
+      "step": 564
+    },
+    {
+      "epoch": 0.3759099933818663,
+      "grad_norm": 1.3365185260772705,
+      "learning_rate": 9.49049899675823e-05,
+      "loss": 1.8855,
+      "step": 568
+    },
+    {
+      "epoch": 0.3785572468563865,
+      "grad_norm": 1.4907687902450562,
+      "learning_rate": 9.480827784805278e-05,
+      "loss": 1.8158,
+      "step": 572
+    },
+    {
+      "epoch": 0.3812045003309067,
+      "grad_norm": 1.2549834251403809,
+      "learning_rate": 9.471070668831208e-05,
+      "loss": 1.6304,
+      "step": 576
+    },
+    {
+      "epoch": 0.38385175380542685,
+      "grad_norm": 1.6914743185043335,
+      "learning_rate": 9.4612278358942e-05,
+      "loss": 1.6976,
+      "step": 580
+    },
+    {
+      "epoch": 0.38649900727994707,
+      "grad_norm": 1.5349342823028564,
+      "learning_rate": 9.451299474695754e-05,
+      "loss": 1.7323,
+      "step": 584
+    },
+    {
+      "epoch": 0.38914626075446723,
+      "grad_norm": 1.4379171133041382,
+      "learning_rate": 9.441285775577075e-05,
+      "loss": 1.7762,
+      "step": 588
+    },
+    {
+      "epoch": 0.39179351422898745,
+      "grad_norm": 1.360475778579712,
+      "learning_rate": 9.431186930515419e-05,
+      "loss": 1.7328,
+      "step": 592
+    },
+    {
+      "epoch": 0.3944407677035076,
+      "grad_norm": 1.4364429712295532,
+      "learning_rate": 9.421003133120412e-05,
+      "loss": 1.7363,
+      "step": 596
+    },
+    {
+      "epoch": 0.3970880211780278,
+      "grad_norm": 1.4598385095596313,
+      "learning_rate": 9.410734578630343e-05,
+      "loss": 1.6917,
+      "step": 600
+    },
+    {
+      "epoch": 0.399735274652548,
+      "grad_norm": 1.3313078880310059,
+      "learning_rate": 9.400381463908416e-05,
+      "loss": 1.8008,
+      "step": 604
+    },
+    {
+      "epoch": 0.40238252812706815,
+      "grad_norm": 1.5070075988769531,
+      "learning_rate": 9.389943987438983e-05,
+      "loss": 1.669,
+      "step": 608
+    },
+    {
+      "epoch": 0.40502978160158837,
+      "grad_norm": 1.3858133554458618,
+      "learning_rate": 9.379422349323728e-05,
+      "loss": 1.6599,
+      "step": 612
+    },
+    {
+      "epoch": 0.40767703507610853,
+      "grad_norm": 1.3775012493133545,
+      "learning_rate": 9.368816751277843e-05,
+      "loss": 1.628,
+      "step": 616
+    },
+    {
+      "epoch": 0.41032428855062875,
+      "grad_norm": 1.3733761310577393,
+      "learning_rate": 9.358127396626147e-05,
+      "loss": 1.6797,
+      "step": 620
+    },
+    {
+      "epoch": 0.4129715420251489,
+      "grad_norm": 1.760237455368042,
+      "learning_rate": 9.347354490299205e-05,
+      "loss": 1.7479,
+      "step": 624
+    },
+    {
+      "epoch": 0.41561879549966907,
+      "grad_norm": 1.2483643293380737,
+      "learning_rate": 9.336498238829384e-05,
+      "loss": 1.6595,
+      "step": 628
+    },
+    {
+      "epoch": 0.4182660489741893,
+      "grad_norm": 2.099116563796997,
+      "learning_rate": 9.325558850346897e-05,
+      "loss": 1.6933,
+      "step": 632
+    },
+    {
+      "epoch": 0.42091330244870945,
+      "grad_norm": 1.3913215398788452,
+      "learning_rate": 9.31453653457582e-05,
+      "loss": 1.6433,
+      "step": 636
+    },
+    {
+      "epoch": 0.42356055592322966,
+      "grad_norm": 1.3813973665237427,
+      "learning_rate": 9.303431502830065e-05,
+      "loss": 1.6652,
+      "step": 640
+    },
+    {
+      "epoch": 0.4262078093977498,
+      "grad_norm": 1.496819019317627,
+      "learning_rate": 9.292243968009331e-05,
+      "loss": 1.747,
+      "step": 644
+    },
+    {
+      "epoch": 0.42885506287227004,
+      "grad_norm": 1.37201988697052,
+      "learning_rate": 9.280974144595018e-05,
+      "loss": 1.6331,
+      "step": 648
+    },
+    {
+      "epoch": 0.4315023163467902,
+      "grad_norm": 1.505353331565857,
+      "learning_rate": 9.269622248646124e-05,
+      "loss": 1.7717,
+      "step": 652
+    },
+    {
+      "epoch": 0.43414956982131037,
+      "grad_norm": 1.8498897552490234,
+      "learning_rate": 9.258188497795093e-05,
+      "loss": 1.6643,
+      "step": 656
+    },
+    {
+      "epoch": 0.4367968232958306,
+      "grad_norm": 1.2886799573898315,
+      "learning_rate": 9.24667311124365e-05,
+      "loss": 1.777,
+      "step": 660
+    },
+    {
+      "epoch": 0.43944407677035074,
+      "grad_norm": 1.283218502998352,
+      "learning_rate": 9.23507630975859e-05,
+      "loss": 1.6958,
+      "step": 664
+    },
+    {
+      "epoch": 0.44209133024487096,
+      "grad_norm": 1.3919546604156494,
+      "learning_rate": 9.223398315667561e-05,
+      "loss": 1.6515,
+      "step": 668
+    },
+    {
+      "epoch": 0.4447385837193911,
+      "grad_norm": 1.4083247184753418,
+      "learning_rate": 9.211639352854787e-05,
+      "loss": 1.7531,
+      "step": 672
+    },
+    {
+      "epoch": 0.44738583719391134,
+      "grad_norm": 1.2739989757537842,
+      "learning_rate": 9.199799646756777e-05,
+      "loss": 1.7694,
+      "step": 676
+    },
+    {
+      "epoch": 0.4500330906684315,
+      "grad_norm": 1.4435306787490845,
+      "learning_rate": 9.187879424358014e-05,
+      "loss": 1.8044,
+      "step": 680
+    },
+    {
+      "epoch": 0.45268034414295166,
+      "grad_norm": 1.4848833084106445,
+      "learning_rate": 9.17587891418659e-05,
+      "loss": 1.6531,
+      "step": 684
+    },
+    {
+      "epoch": 0.4553275976174719,
+      "grad_norm": 1.527485966682434,
+      "learning_rate": 9.163798346309837e-05,
+      "loss": 1.8783,
+      "step": 688
+    },
+    {
+      "epoch": 0.45797485109199204,
+      "grad_norm": 1.2369976043701172,
+      "learning_rate": 9.151637952329903e-05,
+      "loss": 1.5479,
+      "step": 692
+    },
+    {
+      "epoch": 0.46062210456651226,
+      "grad_norm": 1.4693775177001953,
+      "learning_rate": 9.139397965379327e-05,
+      "loss": 1.7891,
+      "step": 696
+    },
+    {
+      "epoch": 0.4632693580410324,
+      "grad_norm": 1.6788188219070435,
+      "learning_rate": 9.127078620116556e-05,
+      "loss": 1.7637,
+      "step": 700
+    },
+    {
+      "epoch": 0.46591661151555264,
+      "grad_norm": 1.3309741020202637,
+      "learning_rate": 9.114680152721453e-05,
+      "loss": 1.6053,
+      "step": 704
+    },
+    {
+      "epoch": 0.4685638649900728,
+      "grad_norm": 1.509023904800415,
+      "learning_rate": 9.102202800890772e-05,
+      "loss": 1.8784,
+      "step": 708
+    },
+    {
+      "epoch": 0.47121111846459296,
+      "grad_norm": 1.3232872486114502,
+      "learning_rate": 9.089646803833589e-05,
+      "loss": 1.6745,
+      "step": 712
+    },
+    {
+      "epoch": 0.4738583719391132,
+      "grad_norm": 1.3540325164794922,
+      "learning_rate": 9.077012402266731e-05,
+      "loss": 1.6668,
+      "step": 716
+    },
+    {
+      "epoch": 0.47650562541363334,
+      "grad_norm": 1.3100489377975464,
+      "learning_rate": 9.064299838410152e-05,
+      "loss": 1.6188,
+      "step": 720
+    },
+    {
+      "epoch": 0.47915287888815355,
+      "grad_norm": 1.3783172369003296,
+      "learning_rate": 9.051509355982293e-05,
+      "loss": 1.6491,
+      "step": 724
+    },
+    {
+      "epoch": 0.4818001323626737,
+      "grad_norm": 1.27851402759552,
+      "learning_rate": 9.038641200195404e-05,
+      "loss": 1.8925,
+      "step": 728
+    },
+    {
+      "epoch": 0.48444738583719393,
+      "grad_norm": 1.4370380640029907,
+      "learning_rate": 9.025695617750848e-05,
+      "loss": 1.7996,
+      "step": 732
+    },
+    {
+      "epoch": 0.4870946393117141,
+      "grad_norm": 1.4078205823898315,
+      "learning_rate": 9.012672856834373e-05,
+      "loss": 1.8554,
+      "step": 736
+    },
+    {
+      "epoch": 0.48974189278623426,
+      "grad_norm": 1.3553669452667236,
+      "learning_rate": 8.999573167111348e-05,
+      "loss": 1.5417,
+      "step": 740
+    },
+    {
+      "epoch": 0.4923891462607545,
+      "grad_norm": 1.4759166240692139,
+      "learning_rate": 8.986396799721983e-05,
+      "loss": 1.6143,
+      "step": 744
+    },
+    {
+      "epoch": 0.49503639973527463,
+      "grad_norm": 1.3601372241973877,
+      "learning_rate": 8.973144007276508e-05,
+      "loss": 1.7011,
+      "step": 748
+    },
+    {
+      "epoch": 0.49768365320979485,
+      "grad_norm": 1.425181269645691,
+      "learning_rate": 8.959815043850336e-05,
+      "loss": 1.672,
+      "step": 752
+    },
+    {
+      "epoch": 0.500330906684315,
+      "grad_norm": 1.440303921699524,
+      "learning_rate": 8.946410164979184e-05,
+      "loss": 1.8008,
+      "step": 756
+    },
+    {
+      "epoch": 0.5029781601588352,
+      "grad_norm": 1.4576961994171143,
+      "learning_rate": 8.932929627654185e-05,
+      "loss": 1.5234,
+      "step": 760
+    },
+    {
+      "epoch": 0.5056254136333554,
+      "grad_norm": 1.3088816404342651,
+      "learning_rate": 8.919373690316952e-05,
+      "loss": 1.701,
+      "step": 764
+    },
+    {
+      "epoch": 0.5082726671078756,
+      "grad_norm": 3.7521555423736572,
+      "learning_rate": 8.905742612854628e-05,
+      "loss": 1.6714,
+      "step": 768
+    },
+    {
+      "epoch": 0.5109199205823958,
+      "grad_norm": 1.4540220499038696,
+      "learning_rate": 8.892036656594898e-05,
+      "loss": 1.6276,
+      "step": 772
+    },
+    {
+      "epoch": 0.513567174056916,
+      "grad_norm": 1.3043605089187622,
+      "learning_rate": 8.87825608430099e-05,
+      "loss": 1.635,
+      "step": 776
+    },
+    {
+      "epoch": 0.5162144275314361,
+      "grad_norm": 1.3931020498275757,
+      "learning_rate": 8.864401160166624e-05,
+      "loss": 1.5822,
+      "step": 780
+    },
+    {
+      "epoch": 0.5188616810059563,
+      "grad_norm": 1.3738582134246826,
+      "learning_rate": 8.85047214981096e-05,
+      "loss": 1.694,
+      "step": 784
+    },
+    {
+      "epoch": 0.5215089344804765,
+      "grad_norm": 1.3968422412872314,
+      "learning_rate": 8.83646932027349e-05,
+      "loss": 1.6673,
+      "step": 788
+    },
+    {
+      "epoch": 0.5241561879549967,
+      "grad_norm": 1.4195423126220703,
+      "learning_rate": 8.822392940008937e-05,
+      "loss": 1.5422,
+      "step": 792
+    },
+    {
+      "epoch": 0.5268034414295168,
+      "grad_norm": 1.2660058736801147,
+      "learning_rate": 8.808243278882094e-05,
+      "loss": 1.4875,
+      "step": 796
+    },
+    {
+      "epoch": 0.5294506949040371,
+      "grad_norm": 1.3500608205795288,
+      "learning_rate": 8.794020608162656e-05,
+      "loss": 1.6946,
+      "step": 800
+    },
+    {
+      "epoch": 0.5320979483785573,
+      "grad_norm": 1.6274265050888062,
+      "learning_rate": 8.779725200520021e-05,
+      "loss": 1.6943,
+      "step": 804
+    },
+    {
+      "epoch": 0.5347452018530774,
+      "grad_norm": 1.2186963558197021,
+      "learning_rate": 8.765357330018056e-05,
+      "loss": 1.4563,
+      "step": 808
+    },
+    {
+      "epoch": 0.5373924553275976,
+      "grad_norm": 1.501142978668213,
+      "learning_rate": 8.750917272109848e-05,
+      "loss": 1.6729,
+      "step": 812
+    },
+    {
+      "epoch": 0.5400397088021178,
+      "grad_norm": 1.372517466545105,
+      "learning_rate": 8.736405303632427e-05,
+      "loss": 1.636,
+      "step": 816
+    },
+    {
+      "epoch": 0.542686962276638,
+      "grad_norm": 1.4448741674423218,
+      "learning_rate": 8.721821702801449e-05,
+      "loss": 1.6977,
+      "step": 820
+    },
+    {
+      "epoch": 0.5453342157511581,
+      "grad_norm": 1.4774208068847656,
+      "learning_rate": 8.707166749205866e-05,
+      "loss": 1.7892,
+      "step": 824
+    },
+    {
+      "epoch": 0.5479814692256784,
+      "grad_norm": 1.3137487173080444,
+      "learning_rate": 8.692440723802571e-05,
+      "loss": 1.5086,
+      "step": 828
+    },
+    {
+      "epoch": 0.5506287227001986,
+      "grad_norm": 1.4480420351028442,
+      "learning_rate": 8.677643908911007e-05,
+      "loss": 1.6694,
+      "step": 832
+    },
+    {
+      "epoch": 0.5532759761747187,
+      "grad_norm": 1.4660981893539429,
+      "learning_rate": 8.662776588207747e-05,
+      "loss": 1.632,
+      "step": 836
+    },
+    {
+      "epoch": 0.5559232296492389,
+      "grad_norm": 1.2639222145080566,
+      "learning_rate": 8.647839046721076e-05,
+      "loss": 1.5101,
+      "step": 840
+    },
+    {
+      "epoch": 0.5585704831237591,
+      "grad_norm": 1.3556458950042725,
+      "learning_rate": 8.632831570825508e-05,
+      "loss": 1.7912,
+      "step": 844
+    },
+    {
+      "epoch": 0.5612177365982793,
+      "grad_norm": 1.2261251211166382,
+      "learning_rate": 8.617754448236298e-05,
+      "loss": 1.6547,
+      "step": 848
+    },
+    {
+      "epoch": 0.5638649900727994,
+      "grad_norm": 1.2850754261016846,
+      "learning_rate": 8.602607968003935e-05,
+      "loss": 1.5365,
+      "step": 852
+    },
+    {
+      "epoch": 0.5665122435473197,
+      "grad_norm": 1.3346043825149536,
+      "learning_rate": 8.587392420508598e-05,
+      "loss": 1.6175,
+      "step": 856
+    },
+    {
+      "epoch": 0.5691594970218399,
+      "grad_norm": 1.5381152629852295,
+      "learning_rate": 8.572108097454578e-05,
+      "loss": 1.7967,
+      "step": 860
+    },
+    {
+      "epoch": 0.57180675049636,
+      "grad_norm": 1.2237263917922974,
+      "learning_rate": 8.556755291864701e-05,
+      "loss": 1.6057,
+      "step": 864
+    },
+    {
+      "epoch": 0.5744540039708802,
+      "grad_norm": 1.233619213104248,
+      "learning_rate": 8.541334298074701e-05,
+      "loss": 1.7107,
+      "step": 868
+    },
+    {
+      "epoch": 0.5771012574454004,
+      "grad_norm": 1.2423778772354126,
+      "learning_rate": 8.525845411727581e-05,
+      "loss": 1.4729,
+      "step": 872
+    },
+    {
+      "epoch": 0.5797485109199206,
+      "grad_norm": 7.3384480476379395,
+      "learning_rate": 8.51028892976794e-05,
+      "loss": 1.6363,
+      "step": 876
+    },
+    {
+      "epoch": 0.5823957643944407,
+      "grad_norm": 1.3198407888412476,
+      "learning_rate": 8.494665150436288e-05,
+      "loss": 1.646,
+      "step": 880
+    },
+    {
+      "epoch": 0.585043017868961,
+      "grad_norm": 1.172568678855896,
+      "learning_rate": 8.478974373263318e-05,
+      "loss": 1.4356,
+      "step": 884
+    },
+    {
+      "epoch": 0.5876902713434812,
+      "grad_norm": 1.4879450798034668,
+      "learning_rate": 8.463216899064179e-05,
+      "loss": 1.7847,
+      "step": 888
+    },
+    {
+      "epoch": 0.5903375248180013,
+      "grad_norm": 1.3998438119888306,
+      "learning_rate": 8.447393029932692e-05,
+      "loss": 1.7818,
+      "step": 892
+    },
+    {
+      "epoch": 0.5929847782925215,
+      "grad_norm": 1.3567726612091064,
+      "learning_rate": 8.431503069235565e-05,
+      "loss": 1.5539,
+      "step": 896
+    },
+    {
+      "epoch": 0.5956320317670417,
+      "grad_norm": 1.4983903169631958,
+      "learning_rate": 8.415547321606584e-05,
+      "loss": 1.6477,
+      "step": 900
+    },
+    {
+      "epoch": 0.5982792852415619,
+      "grad_norm": 1.2646454572677612,
+      "learning_rate": 8.399526092940768e-05,
+      "loss": 1.6087,
+      "step": 904
+    },
+    {
+      "epoch": 0.600926538716082,
+      "grad_norm": 1.4137752056121826,
+      "learning_rate": 8.38343969038849e-05,
+      "loss": 1.7626,
+      "step": 908
+    },
+    {
+      "epoch": 0.6035737921906023,
+      "grad_norm": 1.4016697406768799,
+      "learning_rate": 8.367288422349617e-05,
+      "loss": 1.6947,
+      "step": 912
+    },
+    {
+      "epoch": 0.6062210456651225,
+      "grad_norm": 1.331425666809082,
+      "learning_rate": 8.351072598467576e-05,
+      "loss": 1.6358,
+      "step": 916
+    },
+    {
+      "epoch": 0.6088682991396426,
+      "grad_norm": 1.2292309999465942,
+      "learning_rate": 8.334792529623419e-05,
+      "loss": 1.4613,
+      "step": 920
+    },
+    {
+      "epoch": 0.6115155526141628,
+      "grad_norm": 1.3756728172302246,
+      "learning_rate": 8.318448527929877e-05,
+      "loss": 1.5771,
+      "step": 924
+    },
+    {
+      "epoch": 0.614162806088683,
+      "grad_norm": 1.4124281406402588,
+      "learning_rate": 8.302040906725361e-05,
+      "loss": 1.7364,
+      "step": 928
+    },
+    {
+      "epoch": 0.6168100595632032,
+      "grad_norm": 1.298540472984314,
+      "learning_rate": 8.285569980567964e-05,
+      "loss": 1.6394,
+      "step": 932
+    },
+    {
+      "epoch": 0.6194573130377233,
+      "grad_norm": 1.3905584812164307,
+      "learning_rate": 8.269036065229427e-05,
+      "loss": 1.7034,
+      "step": 936
+    },
+    {
+      "epoch": 0.6221045665122436,
+      "grad_norm": 1.4072821140289307,
+      "learning_rate": 8.252439477689082e-05,
+      "loss": 1.6315,
+      "step": 940
+    },
+    {
+      "epoch": 0.6247518199867638,
+      "grad_norm": 1.239159345626831,
+      "learning_rate": 8.235780536127787e-05,
+      "loss": 1.5178,
+      "step": 944
+    },
+    {
+      "epoch": 0.6273990734612839,
+      "grad_norm": 1.3636091947555542,
+      "learning_rate": 8.21905955992181e-05,
+      "loss": 1.6564,
+      "step": 948
+    },
+    {
+      "epoch": 0.6300463269358041,
+      "grad_norm": 1.3506637811660767,
+      "learning_rate": 8.202276869636713e-05,
+      "loss": 1.646,
+      "step": 952
+    },
+    {
+      "epoch": 0.6326935804103243,
+      "grad_norm": 1.4368304014205933,
+      "learning_rate": 8.185432787021216e-05,
+      "loss": 1.5073,
+      "step": 956
+    },
+    {
+      "epoch": 0.6353408338848445,
+      "grad_norm": 1.3278450965881348,
+      "learning_rate": 8.168527635001015e-05,
+      "loss": 1.5203,
+      "step": 960
+    },
+    {
+      "epoch": 0.6379880873593646,
+      "grad_norm": 1.2450168132781982,
+      "learning_rate": 8.151561737672591e-05,
+      "loss": 1.7171,
+      "step": 964
+    },
+    {
+      "epoch": 0.6406353408338848,
+      "grad_norm": 1.2755018472671509,
+      "learning_rate": 8.134535420297008e-05,
+      "loss": 1.5675,
+      "step": 968
+    },
+    {
+      "epoch": 0.6432825943084051,
+      "grad_norm": 1.3066191673278809,
+      "learning_rate": 8.117449009293668e-05,
+      "loss": 1.6525,
+      "step": 972
+    },
+    {
+      "epoch": 0.6459298477829252,
+      "grad_norm": 1.2875075340270996,
+      "learning_rate": 8.100302832234056e-05,
+      "loss": 1.6484,
+      "step": 976
+    },
+    {
+      "epoch": 0.6485771012574454,
+      "grad_norm": 1.5069595575332642,
+      "learning_rate": 8.083097217835461e-05,
+      "loss": 1.6251,
+      "step": 980
+    },
+    {
+      "epoch": 0.6512243547319656,
+      "grad_norm": 1.334075927734375,
+      "learning_rate": 8.065832495954668e-05,
+      "loss": 1.743,
+      "step": 984
+    },
+    {
+      "epoch": 0.6538716082064858,
+      "grad_norm": 1.3219469785690308,
+      "learning_rate": 8.048508997581647e-05,
+      "loss": 1.6345,
+      "step": 988
+    },
+    {
+      "epoch": 0.6565188616810059,
+      "grad_norm": 1.4275529384613037,
+      "learning_rate": 8.03112705483319e-05,
+      "loss": 1.7515,
+      "step": 992
+    },
+    {
+      "epoch": 0.6591661151555261,
+      "grad_norm": 1.349526286125183,
+      "learning_rate": 8.013687000946561e-05,
+      "loss": 1.5209,
+      "step": 996
+    },
+    {
+      "epoch": 0.6618133686300464,
+      "grad_norm": 1.3620506525039673,
+      "learning_rate": 7.996189170273096e-05,
+      "loss": 1.6789,
+      "step": 1000
+    },
+    {
+      "epoch": 0.6644606221045665,
+      "grad_norm": 1.2079874277114868,
+      "learning_rate": 7.978633898271795e-05,
+      "loss": 1.3453,
+      "step": 1004
+    },
+    {
+      "epoch": 0.6671078755790867,
+      "grad_norm": 1.3527398109436035,
+      "learning_rate": 7.961021521502895e-05,
+      "loss": 1.5927,
+      "step": 1008
+    },
+    {
+      "epoch": 0.6697551290536069,
+      "grad_norm": 1.3048250675201416,
+      "learning_rate": 7.943352377621414e-05,
+      "loss": 1.643,
+      "step": 1012
+    },
+    {
+      "epoch": 0.6724023825281271,
+      "grad_norm": 1.2111921310424805,
+      "learning_rate": 7.925626805370678e-05,
+      "loss": 1.4432,
+      "step": 1016
+    },
+    {
+      "epoch": 0.6750496360026472,
+      "grad_norm": 1.3531336784362793,
+      "learning_rate": 7.907845144575829e-05,
+      "loss": 1.6235,
+      "step": 1020
+    },
+    {
+      "epoch": 0.6776968894771674,
+      "grad_norm": 1.204720139503479,
+      "learning_rate": 7.890007736137307e-05,
+      "loss": 1.5377,
+      "step": 1024
+    },
+    {
+      "epoch": 0.6803441429516877,
+      "grad_norm": 1.3632683753967285,
+      "learning_rate": 7.872114922024313e-05,
+      "loss": 1.5758,
+      "step": 1028
+    },
+    {
+      "epoch": 0.6829913964262078,
+      "grad_norm": 1.4058332443237305,
+      "learning_rate": 7.854167045268264e-05,
+      "loss": 1.4645,
+      "step": 1032
+    },
+    {
+      "epoch": 0.685638649900728,
+      "grad_norm": 1.2490967512130737,
+      "learning_rate": 7.836164449956199e-05,
+      "loss": 1.5723,
+      "step": 1036
+    },
+    {
+      "epoch": 0.6882859033752482,
+      "grad_norm": 1.3228312730789185,
+      "learning_rate": 7.818107481224198e-05,
+      "loss": 1.466,
+      "step": 1040
+    },
+    {
+      "epoch": 0.6909331568497684,
+      "grad_norm": 1.3664582967758179,
+      "learning_rate": 7.799996485250755e-05,
+      "loss": 1.4823,
+      "step": 1044
+    },
+    {
+      "epoch": 0.6935804103242885,
+      "grad_norm": 1.1946579217910767,
+      "learning_rate": 7.781831809250151e-05,
+      "loss": 1.6093,
+      "step": 1048
+    },
+    {
+      "epoch": 0.6962276637988087,
+      "grad_norm": 1.3534433841705322,
+      "learning_rate": 7.763613801465786e-05,
+      "loss": 1.5823,
+      "step": 1052
+    },
+    {
+      "epoch": 0.698874917273329,
+      "grad_norm": 1.275877833366394,
+      "learning_rate": 7.745342811163507e-05,
+      "loss": 1.508,
+      "step": 1056
+    },
+    {
+      "epoch": 0.7015221707478491,
+      "grad_norm": 1.2870965003967285,
+      "learning_rate": 7.727019188624922e-05,
+      "loss": 1.6452,
+      "step": 1060
+    },
+    {
+      "epoch": 0.7041694242223693,
+      "grad_norm": 1.2805050611495972,
+      "learning_rate": 7.708643285140667e-05,
+      "loss": 1.7463,
+      "step": 1064
+    },
+    {
+      "epoch": 0.7068166776968895,
+      "grad_norm": 1.331794261932373,
+      "learning_rate": 7.690215453003684e-05,
+      "loss": 1.4428,
+      "step": 1068
+    },
+    {
+      "epoch": 0.7094639311714097,
+      "grad_norm": 1.3701887130737305,
+      "learning_rate": 7.671736045502462e-05,
+      "loss": 1.6868,
+      "step": 1072
+    },
+    {
+      "epoch": 0.7121111846459298,
+      "grad_norm": 1.3474302291870117,
+      "learning_rate": 7.653205416914267e-05,
+      "loss": 1.4919,
+      "step": 1076
+    },
+    {
+      "epoch": 0.71475843812045,
+      "grad_norm": 1.6028352975845337,
+      "learning_rate": 7.634623922498348e-05,
+      "loss": 1.5958,
+      "step": 1080
+    },
+    {
+      "epoch": 0.7174056915949703,
+      "grad_norm": 1.2263597249984741,
+      "learning_rate": 7.615991918489125e-05,
+      "loss": 1.7238,
+      "step": 1084
+    },
+    {
+      "epoch": 0.7200529450694904,
+      "grad_norm": 1.4178084135055542,
+      "learning_rate": 7.597309762089359e-05,
+      "loss": 1.48,
+      "step": 1088
+    },
+    {
+      "epoch": 0.7227001985440106,
+      "grad_norm": 1.3942856788635254,
+      "learning_rate": 7.57857781146331e-05,
+      "loss": 1.5336,
+      "step": 1092
+    },
+    {
+      "epoch": 0.7253474520185308,
+      "grad_norm": 1.2155961990356445,
+      "learning_rate": 7.559796425729863e-05,
+      "loss": 1.4977,
+      "step": 1096
+    },
+    {
+      "epoch": 0.727994705493051,
+      "grad_norm": 1.3590655326843262,
+      "learning_rate": 7.540965964955649e-05,
+      "loss": 1.6736,
+      "step": 1100
+    },
+    {
+      "epoch": 0.7306419589675711,
+      "grad_norm": 1.1585520505905151,
+      "learning_rate": 7.522086790148133e-05,
+      "loss": 1.6883,
+      "step": 1104
+    },
+    {
+      "epoch": 0.7332892124420913,
+      "grad_norm": 1.2694188356399536,
+      "learning_rate": 7.503159263248709e-05,
+      "loss": 1.657,
+      "step": 1108
+    },
+    {
+      "epoch": 0.7359364659166115,
+      "grad_norm": 1.2413800954818726,
+      "learning_rate": 7.484183747125742e-05,
+      "loss": 1.4757,
+      "step": 1112
+    },
+    {
+      "epoch": 0.7385837193911317,
+      "grad_norm": 1.1527191400527954,
+      "learning_rate": 7.46516060556763e-05,
+      "loss": 1.5628,
+      "step": 1116
+    },
+    {
+      "epoch": 0.7412309728656519,
+      "grad_norm": 1.5187007188796997,
+      "learning_rate": 7.446090203275809e-05,
+      "loss": 1.6387,
+      "step": 1120
+    },
+    {
+      "epoch": 0.7438782263401721,
+      "grad_norm": 1.3278498649597168,
+      "learning_rate": 7.426972905857781e-05,
+      "loss": 1.5212,
+      "step": 1124
+    },
+    {
+      "epoch": 0.7465254798146923,
+      "grad_norm": 1.4994242191314697,
+      "learning_rate": 7.407809079820094e-05,
+      "loss": 1.7582,
+      "step": 1128
+    },
+    {
+      "epoch": 0.7491727332892124,
+      "grad_norm": 1.2623709440231323,
+      "learning_rate": 7.388599092561315e-05,
+      "loss": 1.6223,
+      "step": 1132
+    },
+    {
+      "epoch": 0.7518199867637326,
+      "grad_norm": 1.3785511255264282,
+      "learning_rate": 7.369343312364993e-05,
+      "loss": 1.5051,
+      "step": 1136
+    },
+    {
+      "epoch": 0.7544672402382528,
+      "grad_norm": 1.2472020387649536,
+      "learning_rate": 7.350042108392594e-05,
+      "loss": 1.419,
+      "step": 1140
+    },
+    {
+      "epoch": 0.757114493712773,
+      "grad_norm": 1.6892167329788208,
+      "learning_rate": 7.330695850676421e-05,
+      "loss": 1.5718,
+      "step": 1144
+    },
+    {
+      "epoch": 0.7597617471872932,
+      "grad_norm": 1.4521297216415405,
+      "learning_rate": 7.311304910112525e-05,
+      "loss": 1.6383,
+      "step": 1148
+    },
+    {
+      "epoch": 0.7624090006618134,
+      "grad_norm": 1.450149655342102,
+      "learning_rate": 7.291869658453594e-05,
+      "loss": 1.771,
+      "step": 1152
+    },
+    {
+      "epoch": 0.7650562541363336,
+      "grad_norm": 1.3068790435791016,
+      "learning_rate": 7.272390468301821e-05,
+      "loss": 1.6414,
+      "step": 1156
+    },
+    {
+      "epoch": 0.7677035076108537,
+      "grad_norm": 1.1887469291687012,
+      "learning_rate": 7.252867713101771e-05,
+      "loss": 1.3455,
+      "step": 1160
+    },
+    {
+      "epoch": 0.7703507610853739,
+      "grad_norm": 1.2392699718475342,
+      "learning_rate": 7.233301767133205e-05,
+      "loss": 1.5139,
+      "step": 1164
+    },
+    {
+      "epoch": 0.7729980145598941,
+      "grad_norm": 1.353925347328186,
+      "learning_rate": 7.213693005503924e-05,
+      "loss": 1.6324,
+      "step": 1168
+    },
+    {
+      "epoch": 0.7756452680344142,
+      "grad_norm": 1.2792888879776,
+      "learning_rate": 7.194041804142557e-05,
+      "loss": 1.69,
+      "step": 1172
+    },
+    {
+      "epoch": 0.7782925215089345,
+      "grad_norm": 1.1825402975082397,
+      "learning_rate": 7.174348539791375e-05,
+      "loss": 1.3613,
+      "step": 1176
+    },
+    {
+      "epoch": 0.7809397749834547,
+      "grad_norm": 1.2615066766738892,
+      "learning_rate": 7.154613589999054e-05,
+      "loss": 1.6972,
+      "step": 1180
+    },
+    {
+      "epoch": 0.7835870284579749,
+      "grad_norm": 1.239867091178894,
+      "learning_rate": 7.13483733311344e-05,
+      "loss": 1.403,
+      "step": 1184
+    },
+    {
+      "epoch": 0.786234281932495,
+      "grad_norm": 1.3656786680221558,
+      "learning_rate": 7.115020148274295e-05,
+      "loss": 1.6528,
+      "step": 1188
+    },
+    {
+      "epoch": 0.7888815354070152,
+      "grad_norm": 1.2590436935424805,
+      "learning_rate": 7.095162415406034e-05,
+      "loss": 1.5411,
+      "step": 1192
+    },
+    {
+      "epoch": 0.7915287888815354,
+      "grad_norm": 1.2784417867660522,
+      "learning_rate": 7.075264515210435e-05,
+      "loss": 1.5618,
+      "step": 1196
+    },
+    {
+      "epoch": 0.7941760423560555,
+      "grad_norm": 1.3260300159454346,
+      "learning_rate": 7.055326829159341e-05,
+      "loss": 1.5295,
+      "step": 1200
+    },
+    {
+      "epoch": 0.7968232958305758,
+      "grad_norm": 5.832207202911377,
+      "learning_rate": 7.03534973948735e-05,
+      "loss": 1.5864,
+      "step": 1204
+    },
+    {
+      "epoch": 0.799470549305096,
+      "grad_norm": 1.2828547954559326,
+      "learning_rate": 7.015333629184484e-05,
+      "loss": 1.5081,
+      "step": 1208
+    },
+    {
+      "epoch": 0.8021178027796162,
+      "grad_norm": 1.2997095584869385,
+      "learning_rate": 6.995278881988847e-05,
+      "loss": 1.5827,
+      "step": 1212
+    },
+    {
+      "epoch": 0.8047650562541363,
+      "grad_norm": 1.2829680442810059,
+      "learning_rate": 6.975185882379271e-05,
+      "loss": 1.4565,
+      "step": 1216
+    },
+    {
+      "epoch": 0.8074123097286565,
+      "grad_norm": 1.3034470081329346,
+      "learning_rate": 6.955055015567942e-05,
+      "loss": 1.4973,
+      "step": 1220
+    },
+    {
+      "epoch": 0.8100595632031767,
+      "grad_norm": 1.170404314994812,
+      "learning_rate": 6.934886667493012e-05,
+      "loss": 1.4518,
+      "step": 1224
+    },
+    {
+      "epoch": 0.8127068166776968,
+      "grad_norm": 1.2815779447555542,
+      "learning_rate": 6.914681224811208e-05,
+      "loss": 1.546,
+      "step": 1228
+    },
+    {
+      "epoch": 0.8153540701522171,
+      "grad_norm": 1.227200984954834,
+      "learning_rate": 6.894439074890414e-05,
+      "loss": 1.5478,
+      "step": 1232
+    },
+    {
+      "epoch": 0.8180013236267373,
+      "grad_norm": 1.2927132844924927,
+      "learning_rate": 6.874160605802244e-05,
+      "loss": 1.6184,
+      "step": 1236
+    },
+    {
+      "epoch": 0.8206485771012575,
+      "grad_norm": 1.2327131032943726,
+      "learning_rate": 6.853846206314605e-05,
+      "loss": 1.5553,
+      "step": 1240
+    },
+    {
+      "epoch": 0.8232958305757776,
+      "grad_norm": 1.1886876821517944,
+      "learning_rate": 6.833496265884241e-05,
+      "loss": 1.4956,
+      "step": 1244
+    },
+    {
+      "epoch": 0.8259430840502978,
+      "grad_norm": 1.4828628301620483,
+      "learning_rate": 6.813111174649269e-05,
+      "loss": 1.7339,
+      "step": 1248
+    },
+    {
+      "epoch": 0.828590337524818,
+      "grad_norm": 1.2269375324249268,
+      "learning_rate": 6.792691323421698e-05,
+      "loss": 1.5712,
+      "step": 1252
+    },
+    {
+      "epoch": 0.8312375909993381,
+      "grad_norm": 1.4898347854614258,
+      "learning_rate": 6.772237103679937e-05,
+      "loss": 1.6172,
+      "step": 1256
+    },
+    {
+      "epoch": 0.8338848444738584,
+      "grad_norm": 1.1373467445373535,
+      "learning_rate": 6.751748907561288e-05,
+      "loss": 1.3869,
+      "step": 1260
+    },
+    {
+      "epoch": 0.8365320979483786,
+      "grad_norm": 1.2607003450393677,
+      "learning_rate": 6.731227127854434e-05,
+      "loss": 1.5501,
+      "step": 1264
+    },
+    {
+      "epoch": 0.8391793514228988,
+      "grad_norm": 1.357080340385437,
+      "learning_rate": 6.710672157991899e-05,
+      "loss": 1.5804,
+      "step": 1268
+    },
+    {
+      "epoch": 0.8418266048974189,
+      "grad_norm": 1.300445318222046,
+      "learning_rate": 6.690084392042513e-05,
+      "loss": 1.4547,
+      "step": 1272
+    },
+    {
+      "epoch": 0.8444738583719391,
+      "grad_norm": 1.281031608581543,
+      "learning_rate": 6.669464224703861e-05,
+      "loss": 1.5843,
+      "step": 1276
+    },
+    {
+      "epoch": 0.8471211118464593,
+      "grad_norm": 1.2201812267303467,
+      "learning_rate": 6.648812051294697e-05,
+      "loss": 1.4422,
+      "step": 1280
+    },
+    {
+      "epoch": 0.8497683653209794,
+      "grad_norm": 1.2445136308670044,
+      "learning_rate": 6.628128267747391e-05,
+      "loss": 1.5826,
+      "step": 1284
+    },
+    {
+      "epoch": 0.8524156187954997,
+      "grad_norm": 1.383170247077942,
+      "learning_rate": 6.607413270600319e-05,
+      "loss": 1.6194,
+      "step": 1288
+    },
+    {
+      "epoch": 0.8550628722700199,
+      "grad_norm": 1.370076060295105,
+      "learning_rate": 6.586667456990267e-05,
+      "loss": 1.6408,
+      "step": 1292
+    },
+    {
+      "epoch": 0.8577101257445401,
+      "grad_norm": 1.293721318244934,
+      "learning_rate": 6.565891224644822e-05,
+      "loss": 1.5066,
+      "step": 1296
+    },
+    {
+      "epoch": 0.8603573792190602,
+      "grad_norm": 1.4381659030914307,
+      "learning_rate": 6.545084971874738e-05,
+      "loss": 1.5161,
+      "step": 1300
+    },
+    {
+      "epoch": 0.8630046326935804,
+      "grad_norm": 1.3525183200836182,
+      "learning_rate": 6.524249097566306e-05,
+      "loss": 1.6022,
+      "step": 1304
+    },
+    {
+      "epoch": 0.8656518861681006,
+      "grad_norm": 1.1742914915084839,
+      "learning_rate": 6.503384001173707e-05,
+      "loss": 1.3307,
+      "step": 1308
+    },
+    {
+      "epoch": 0.8682991396426207,
+      "grad_norm": 1.275770664215088,
+      "learning_rate": 6.48249008271135e-05,
+      "loss": 1.5092,
+      "step": 1312
+    },
+    {
+      "epoch": 0.870946393117141,
+      "grad_norm": 1.3267558813095093,
+      "learning_rate": 6.461567742746206e-05,
+      "loss": 1.6288,
+      "step": 1316
+    },
+    {
+      "epoch": 0.8735936465916612,
+      "grad_norm": 1.1977699995040894,
+      "learning_rate": 6.440617382390128e-05,
+      "loss": 1.5567,
+      "step": 1320
+    },
+    {
+      "epoch": 0.8762409000661814,
+      "grad_norm": 1.1399099826812744,
+      "learning_rate": 6.419639403292161e-05,
+      "loss": 1.5925,
+      "step": 1324
+    },
+    {
+      "epoch": 0.8788881535407015,
+      "grad_norm": 1.3445255756378174,
+      "learning_rate": 6.398634207630841e-05,
+      "loss": 1.5288,
+      "step": 1328
+    },
+    {
+      "epoch": 0.8815354070152217,
+      "grad_norm": 1.2953174114227295,
+      "learning_rate": 6.377602198106483e-05,
+      "loss": 1.5119,
+      "step": 1332
+    },
+    {
+      "epoch": 0.8841826604897419,
+      "grad_norm": 1.2466961145401,
+      "learning_rate": 6.356543777933468e-05,
+      "loss": 1.4559,
+      "step": 1336
+    },
+    {
+      "epoch": 0.886829913964262,
+      "grad_norm": 1.410008430480957,
+      "learning_rate": 6.335459350832504e-05,
+      "loss": 1.6239,
+      "step": 1340
+    },
+    {
+      "epoch": 0.8894771674387822,
+      "grad_norm": 1.2374393939971924,
+      "learning_rate": 6.314349321022893e-05,
+      "loss": 1.4162,
+      "step": 1344
+    },
+    {
+      "epoch": 0.8921244209133025,
+      "grad_norm": 1.3700758218765259,
+      "learning_rate": 6.293214093214775e-05,
+      "loss": 1.4784,
+      "step": 1348
+    },
+    {
+      "epoch": 0.8947716743878227,
+      "grad_norm": 1.345596432685852,
+      "learning_rate": 6.272054072601374e-05,
+      "loss": 1.5489,
+      "step": 1352
+    },
+    {
+      "epoch": 0.8974189278623428,
+      "grad_norm": 1.1666315793991089,
+      "learning_rate": 6.250869664851227e-05,
+      "loss": 1.3515,
+      "step": 1356
+    },
+    {
+      "epoch": 0.900066181336863,
+      "grad_norm": 1.2450063228607178,
+      "learning_rate": 6.229661276100412e-05,
+      "loss": 1.4763,
+      "step": 1360
+    },
+    {
+      "epoch": 0.9027134348113832,
+      "grad_norm": 1.1888995170593262,
+      "learning_rate": 6.208429312944754e-05,
+      "loss": 1.4322,
+      "step": 1364
+    },
+    {
+      "epoch": 0.9053606882859033,
+      "grad_norm": 1.3319921493530273,
+      "learning_rate": 6.187174182432033e-05,
+      "loss": 1.5044,
+      "step": 1368
+    },
+    {
+      "epoch": 0.9080079417604235,
+      "grad_norm": 1.2023800611495972,
+      "learning_rate": 6.165896292054187e-05,
+      "loss": 1.5033,
+      "step": 1372
+    },
+    {
+      "epoch": 0.9106551952349438,
+      "grad_norm": 1.3017017841339111,
+      "learning_rate": 6.14459604973949e-05,
+      "loss": 1.4683,
+      "step": 1376
+    },
+    {
+      "epoch": 0.913302448709464,
+      "grad_norm": 1.2657389640808105,
+      "learning_rate": 6.12327386384473e-05,
+      "loss": 1.5533,
+      "step": 1380
+    },
+    {
+      "epoch": 0.9159497021839841,
+      "grad_norm": 1.3227919340133667,
+      "learning_rate": 6.101930143147395e-05,
+      "loss": 1.5239,
+      "step": 1384
+    },
+    {
+      "epoch": 0.9185969556585043,
+      "grad_norm": 1.3174325227737427,
+      "learning_rate": 6.080565296837821e-05,
+      "loss": 1.5259,
+      "step": 1388
+    },
+    {
+      "epoch": 0.9212442091330245,
+      "grad_norm": 1.2424542903900146,
+      "learning_rate": 6.059179734511356e-05,
+      "loss": 1.3573,
+      "step": 1392
+    },
+    {
+      "epoch": 0.9238914626075446,
+      "grad_norm": 1.2109280824661255,
+      "learning_rate": 6.037773866160502e-05,
+      "loss": 1.3831,
+      "step": 1396
+    },
+    {
+      "epoch": 0.9265387160820648,
+      "grad_norm": 1.2729474306106567,
+      "learning_rate": 6.0163481021670575e-05,
+      "loss": 1.674,
+      "step": 1400
+    },
+    {
+      "epoch": 0.9291859695565851,
+      "grad_norm": 1.1736104488372803,
+      "learning_rate": 5.994902853294251e-05,
+      "loss": 1.4935,
+      "step": 1404
+    },
+    {
+      "epoch": 0.9318332230311053,
+      "grad_norm": 1.3021750450134277,
+      "learning_rate": 5.973438530678861e-05,
+      "loss": 1.6066,
+      "step": 1408
+    },
+    {
+      "epoch": 0.9344804765056254,
+      "grad_norm": 1.3625566959381104,
+      "learning_rate": 5.951955545823342e-05,
+      "loss": 1.629,
+      "step": 1412
+    },
+    {
+      "epoch": 0.9371277299801456,
+      "grad_norm": 1.1946239471435547,
+      "learning_rate": 5.930454310587929e-05,
+      "loss": 1.4444,
+      "step": 1416
+    },
+    {
+      "epoch": 0.9397749834546658,
+      "grad_norm": 1.4337393045425415,
+      "learning_rate": 5.9089352371827446e-05,
+      "loss": 1.6888,
+      "step": 1420
+    },
+    {
+      "epoch": 0.9424222369291859,
+      "grad_norm": 1.3422842025756836,
+      "learning_rate": 5.8873987381598924e-05,
+      "loss": 1.6227,
+      "step": 1424
+    },
+    {
+      "epoch": 0.9450694904037061,
+      "grad_norm": 1.2459781169891357,
+      "learning_rate": 5.865845226405553e-05,
+      "loss": 1.4704,
+      "step": 1428
+    },
+    {
+      "epoch": 0.9477167438782264,
+      "grad_norm": 1.5130873918533325,
+      "learning_rate": 5.844275115132064e-05,
+      "loss": 1.5029,
+      "step": 1432
+    },
+    {
+      "epoch": 0.9503639973527466,
+      "grad_norm": 1.127805471420288,
+      "learning_rate": 5.822688817870004e-05,
+      "loss": 1.5289,
+      "step": 1436
+    },
+    {
+      "epoch": 0.9530112508272667,
+      "grad_norm": 1.283653736114502,
+      "learning_rate": 5.801086748460255e-05,
+      "loss": 1.545,
+      "step": 1440
+    },
+    {
+      "epoch": 0.9556585043017869,
+      "grad_norm": 1.290038824081421,
+      "learning_rate": 5.7794693210460804e-05,
+      "loss": 1.5588,
+      "step": 1444
+    },
+    {
+      "epoch": 0.9583057577763071,
+      "grad_norm": 1.2246005535125732,
+      "learning_rate": 5.757836950065172e-05,
+      "loss": 1.4577,
+      "step": 1448
+    },
+    {
+      "epoch": 0.9609530112508272,
+      "grad_norm": 1.3036789894104004,
+      "learning_rate": 5.736190050241719e-05,
+      "loss": 1.6891,
+      "step": 1452
+    },
+    {
+      "epoch": 0.9636002647253474,
+      "grad_norm": 1.2149336338043213,
+      "learning_rate": 5.714529036578443e-05,
+      "loss": 1.4114,
+      "step": 1456
+    },
+    {
+      "epoch": 0.9662475181998676,
+      "grad_norm": 1.1539721488952637,
+      "learning_rate": 5.692854324348653e-05,
+      "loss": 1.5497,
+      "step": 1460
+    },
+    {
+      "epoch": 0.9688947716743879,
+      "grad_norm": 1.3573237657546997,
+      "learning_rate": 5.6711663290882776e-05,
+      "loss": 1.4812,
+      "step": 1464
+    },
+    {
+      "epoch": 0.971542025148908,
+      "grad_norm": 1.338049292564392,
+      "learning_rate": 5.649465466587902e-05,
+      "loss": 1.6043,
+      "step": 1468
+    },
+    {
+      "epoch": 0.9741892786234282,
+      "grad_norm": 1.3066737651824951,
+      "learning_rate": 5.627752152884794e-05,
+      "loss": 1.582,
+      "step": 1472
+    },
+    {
+      "epoch": 0.9768365320979484,
+      "grad_norm": 1.2373597621917725,
+      "learning_rate": 5.606026804254931e-05,
+      "loss": 1.4099,
+      "step": 1476
+    },
+    {
+      "epoch": 0.9794837855724685,
+      "grad_norm": 1.1805121898651123,
+      "learning_rate": 5.584289837205012e-05,
+      "loss": 1.3914,
+      "step": 1480
+    },
+    {
+      "epoch": 0.9821310390469887,
+      "grad_norm": 1.3286755084991455,
+      "learning_rate": 5.5625416684644874e-05,
+      "loss": 1.4803,
+      "step": 1484
+    },
+    {
+      "epoch": 0.984778292521509,
+      "grad_norm": 1.2538626194000244,
+      "learning_rate": 5.540782714977549e-05,
+      "loss": 1.5063,
+      "step": 1488
+    },
+    {
+      "epoch": 0.9874255459960292,
+      "grad_norm": 1.2164523601531982,
+      "learning_rate": 5.51901339389516e-05,
+      "loss": 1.3555,
+      "step": 1492
+    },
+    {
+      "epoch": 0.9900727994705493,
+      "grad_norm": 1.217489242553711,
+      "learning_rate": 5.4972341225670354e-05,
+      "loss": 1.4818,
+      "step": 1496
+    },
+    {
+      "epoch": 0.9927200529450695,
+      "grad_norm": 1.170462727546692,
+      "learning_rate": 5.4754453185336586e-05,
+      "loss": 1.5693,
+      "step": 1500
+    },
+    {
+      "epoch": 0.9953673064195897,
+      "grad_norm": 1.2590230703353882,
+      "learning_rate": 5.453647399518262e-05,
+      "loss": 1.3735,
+      "step": 1504
+    },
+    {
+      "epoch": 0.9980145598941098,
+      "grad_norm": 1.1807870864868164,
+      "learning_rate": 5.431840783418832e-05,
+      "loss": 1.3643,
+      "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 sha256:22807a4d0d98b07d9135feb748772a4c53397f31504eaf19f1d1408e42ec0932
+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-mean/",
+  "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 sha256:62150dfca850474c23b0420dfe091c7a57574f909a74466a615d0dba05a97414
+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:fca1f1c7bd769380a3b6ebf36010ac29d9eb940e34a22d50bd65540b14a99268
+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 sha256:5d6731a327e10326468aa08e2ab0c47be9aa123272f74fe29cb86738fdee5efb
+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 sha256:22807a4d0d98b07d9135feb748772a4c53397f31504eaf19f1d1408e42ec0932
+size 5432

config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "/mnt/parscratch/users/acp23ay/private/models/OLMo-2-1124-7B-szl_Latn-coli-mean/",
+  "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 sha256:62150dfca850474c23b0420dfe091c7a57574f909a74466a615d0dba05a97414
+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:fca1f1c7bd769380a3b6ebf36010ac29d9eb940e34a22d50bd65540b14a99268
+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 sha256:22807a4d0d98b07d9135feb748772a4c53397f31504eaf19f1d1408e42ec0932
+size 5432