pierre-livetrend/llama381binstruct_summarize_short

README.md

@@ -27,7 +27,7 @@ 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/livetrend/huggingface/runs/2vsckc48) 
+[<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/livetrend/huggingface/runs/uphd67vi) 
 
 
 This model was trained with SFT.

adapter_config.json

@@ -24,13 +24,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
-    "o_proj",
+    "up_proj",
     "gate_proj",
-    "k_proj",
     "q_proj",
+    "k_proj",
+    "o_proj",
     "down_proj",
-    "up_proj"
+    "v_proj"
   ],
   "task_type": "CAUSAL_LM",
   "trainable_token_indices": null,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:df4e7f3e7369533e1bf3cbb550f2f983f55b6504c36c52ef3579741a29fc9145
+oid sha256:03b93f877444858e3c839e3aa2c9ae2420359088f0240c5fbd4f2bc343c364da
 size 167832240

runs/Apr01_20-20-37_7abccc83e92d/events.out.tfevents.1743539003.7abccc83e92d.1313.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b492532559b25b035610172228f439296e537364cf5f96373ff6f88e85f3478
+size 29690

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:813e90b8bfeb26e2e3dc4a21f5f3a5dfa4d6885612b68aad0aedb5d4990da383
+oid sha256:a95be7ce851690c3eb0a52a28b6ab466ef4b4903232130d399184709d631dce6
 size 5688