Training in progress, step 5

adapter_config.json

@@ -1,7 +1,7 @@
 {
   "alpha_pattern": {},
   "auto_mapping": null,
-  "base_model_name_or_path": "unsloth/Llama-3.2-1B-Instruct",
+  "base_model_name_or_path": "unsloth/Llama-3.2-3B-Instruct",
   "bias": "none",
   "fan_in_fan_out": null,
   "inference_mode": true,
@@ -20,13 +20,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "gate_proj",
-    "k_proj",
     "up_proj",
-    "o_proj",
-    "q_proj",
     "v_proj",
-    "down_proj"
+    "k_proj",
+    "q_proj",
+    "gate_proj",
+    "down_proj",
+    "o_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:55355fd5952cf80856902cdedbaa5578c5033d8b5a54622199ec88d09ed34369
-size 45118424
+oid sha256:5cdcca758e4f51b86ca61d6dd0e1f205a4136be14bada11320e201462d69597d
+size 97307544

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "unsloth/Llama-3.2-1B-Instruct",
+  "_name_or_path": "unsloth/Llama-3.2-3B-Instruct",
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -7,16 +7,16 @@
   "attention_dropout": 0.0,
   "bos_token_id": 128000,
   "eos_token_id": 128009,
-  "head_dim": 64,
+  "head_dim": 128,
   "hidden_act": "silu",
-  "hidden_size": 2048,
+  "hidden_size": 3072,
   "initializer_range": 0.02,
   "intermediate_size": 8192,
   "max_position_embeddings": 131072,
   "mlp_bias": false,
   "model_type": "llama",
-  "num_attention_heads": 32,
-  "num_hidden_layers": 16,
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
   "num_key_value_heads": 8,
   "pretraining_tp": 1,
   "quantization_config": {

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4f3fba08eeeb586587427cc2d5f17808d9e4f299ce6f2dc52ed4d0f7fb28b25a
+oid sha256:0796ff0b1a86f9f71fa065e52907f37a5901a78d76587d76fddb76319a6b4ce3
 size 6648