Training in progress, step 25

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": [
-    "up_proj",
+    "o_proj",
+    "q_proj",
     "v_proj",
+    "up_proj",
+    "gate_proj",
     "down_proj",
-    "k_proj",
-    "q_proj",
-    "o_proj",
-    "gate_proj"
+    "k_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1fff4c4c5300b09c202586de1b731e5ef75ce95b001250060dc9d64c8d87d214
-size 45118424
+oid sha256:8e681c7412ef2557f51f437204f7dbbccb418dccefea52569f201f7a05c61feb
+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 sha256:538ddb675c19e11b2b5d087f2549e85b6180656767f11457e34a8f69d9e78e58
+oid sha256:a8524cfa1063fe98d8aea167ded37880cba9a9019662d453eab8ed82b8700126
 size 6648