Training in progress, step 362

README.md

@@ -5,8 +5,8 @@ model_name: test-gemini
 tags:
 - generated_from_trainer
 - trl
-- sft
 - unsloth
+- sft
 licence: license
 ---
 
@@ -28,7 +28,7 @@ print(output["generated_text"])
 
 ## Training procedure
 
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/batuhan409/huggingface/runs/43fs3x2p) 
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/batuhan409/huggingface/runs/k89bvfrt) 
 
 
 This model was trained with SFT.

adapter_config.json

@@ -20,7 +20,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 256,
+  "lora_alpha": 128,
   "lora_bias": false,
   "lora_dropout": 0,
   "megatron_config": null,
@@ -35,11 +35,11 @@
   "target_modules": [
     "k_proj",
     "down_proj",
-    "q_proj",
+    "gate_proj",
+    "v_proj",
     "o_proj",
     "up_proj",
-    "gate_proj",
-    "v_proj"
+    "q_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b3ce31ba0e668c10d07dfa8a7e2fd348b76585474273a196e29570d05549cf56
+oid sha256:52aa2324035c1292bf2fa39f648f91ac5671d3d542c9e4e2b5783d5972371b74
 size 1347775120

tokenizer_config.json

@@ -11,7 +11,7 @@
   ],
   "model_max_length": 131072,
   "model_specific_special_tokens": {},
-  "pad_token": "<|reserved_special_token_250|>",
+  "pad_token": "<|eot_id|>",
   "padding_side": "right",
   "tokenizer_class": "TokenizersBackend"
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c30167f40d7b7a967ef44dde73679b16467d012fe5bd5a7b809f4da605df0cb
+oid sha256:35428c9ecf443b790ca66a1a03447a6eeb723685fc36c79a6b97241be28d4f9c
 size 5713