Training in progress, step 50

README.md

@@ -3,9 +3,9 @@ library_name: transformers
 model_name: Qwen3-VL-2B-1
 tags:
 - generated_from_trainer
-- hf_jobs
-- trl
 - grpo
+- trl
+- hf_jobs
 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/flofiz-universit-de-bourgogne/GRPO_HTML/runs/evr1zsdn) 
+[<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/flofiz-universit-de-bourgogne/GRPO_HTML/runs/w3kma2u3) 
 
 
 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

@@ -25,13 +25,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "down_proj",
-    "q_proj",
     "v_proj",
-    "o_proj",
     "up_proj",
     "gate_proj",
-    "k_proj"
+    "k_proj",
+    "down_proj",
+    "o_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:c54a0d4c1dac18234cc22f7ef3ed33b2ded02955909eebc120291c7f29836599
+oid sha256:d67722f6f6bbcbfde71c09f37abee841a094a2aba3de42c8f8e33a472af4f57e
 size 69788656

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:93379f4ede47fc93c2740939916b05f9ff1375a401a00578b74755dcdf957303
+oid sha256:ced3232eeeb418f48257f0916e00a2da85101d1c07d731e6566d86ee64673fb8
 size 8273