Training in progress, step 100

README.md

@@ -4,8 +4,8 @@ library_name: transformers
 model_name: nmt
 tags:
 - generated_from_trainer
-- grpo
 - trl
+- grpo
 licence: license
 ---
 
@@ -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/oleg-dats/nmt/runs/y1qtriux) 
+[<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/oleg-dats/nmt/runs/2ixkud5t) 
 
 
 This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).

adapter_config.json

@@ -28,12 +28,12 @@
   "revision": null,
   "target_modules": [
     "up_proj",
-    "o_proj",
     "k_proj",
-    "v_proj",
+    "o_proj",
     "q_proj",
     "down_proj",
-    "gate_proj"
+    "gate_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:b774632159f6df91102d9b2ceaca15303e626645941ed28bb54bc81ed9fed22a
+oid sha256:6fca9cbc348d328e521ee81a005473145f04f1aef651f2b485932a148c8fb649
 size 1260191096

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:efb916066b6a1349830dcb25d4a7360bff0fc8d892b200a06e9bf137e6f90e40
+oid sha256:56bf876f6bb3cc4317493c7e348dd329e19163ec277906faf60680b3bd646db5
 size 6392