Training in progress, step 50

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,
@@ -20,12 +20,12 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
-    "v_proj",
-    "gate_proj",
     "k_proj",
+    "gate_proj",
     "up_proj",
     "o_proj",
+    "v_proj",
+    "q_proj",
     "down_proj"
   ],
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c3f769ff1e4ba4c98c03a624ed01a33e854004d0f07165933ad20c79838391d
-size 45118424
+oid sha256:ef850e00f6ddccf7a92baced38593af32bacbc7deb92d8d8407aa4aa0408de1a
+size 97307544

config.json

@@ -1,6 +1,6 @@
 {
   "_attn_implementation_autoset": true,
-  "_name_or_path": "unsloth/Llama-3.2-1B-Instruct",
+  "_name_or_path": "unsloth/Llama-3.2-3B-Instruct",
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -8,16 +8,16 @@
   "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,
   "rms_norm_eps": 1e-05,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:149f14fb729cab896091dbb95460bb98759373752ba272a6ae1f5cde47ccec9e
+oid sha256:2c10884f4be4e91add9795afd80f7b25e35f322adac12d6b37b6d44aa702d50c
 size 6776