Hugging Face's logo

Yukang
/
Qwen2.5-3B-Open-R1-Code-GRPO

Text Generation
Transformers
Safetensors
qwen2
Generated from Trainer
open-r1
trl
grpo
conversational
text-generation-inference
Model card Files
xet
Community
Yukang commited on
Commit
58b82f6
·
verified ·
1 Parent(s): 2ebcfd8

Training in progress, step 509

Browse files
Files changed (7) hide show
  1. README.md +4 -2
  2. all_results.json +4 -4
  3. model-00001-of-00002.safetensors +1 -1
  4. model-00002-of-00002.safetensors +1 -1
  5. train_results.json +4 -4
  6. trainer_state.json +0 -0
  7. training_args.bin +1 -1
README.md CHANGED
@@ -1,9 +1,11 @@
1
  ---
2
  base_model: Qwen/Qwen2.5-3B-Instruct
 
3
  library_name: transformers
4
  model_name: Qwen2.5-3B-Open-R1-Code-GRPO
5
  tags:
6
  - generated_from_trainer
 
7
  - trl
8
  - grpo
9
  licence: license
@@ -11,7 +13,7 @@ licence: license
11
 
12
  # Model Card for Qwen2.5-3B-Open-R1-Code-GRPO
13
 
14
- This model is a fine-tuned version of [Qwen/Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct).
15
  It has been trained using [TRL](https://github.com/huggingface/trl).
16
 
17
  ## Quick start
@@ -27,7 +29,7 @@ print(output["generated_text"])
27
 
28
  ## Training procedure
29
 
30
- [<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/chenyukang2020-nvidia/huggingface/runs/3g7r9wqj)
31
 
32
 
33
  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).
 
1
  ---
2
  base_model: Qwen/Qwen2.5-3B-Instruct
3
+ datasets: open-r1/verifiable-coding-problems-python
4
  library_name: transformers
5
  model_name: Qwen2.5-3B-Open-R1-Code-GRPO
6
  tags:
7
  - generated_from_trainer
8
+ - open-r1
9
  - trl
10
  - grpo
11
  licence: license
 
13
 
14
  # Model Card for Qwen2.5-3B-Open-R1-Code-GRPO
15
 
16
+ This model is a fine-tuned version of [Qwen/Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct) on the [open-r1/verifiable-coding-problems-python](https://huggingface.co/datasets/open-r1/verifiable-coding-problems-python) dataset.
17
  It has been trained using [TRL](https://github.com/huggingface/trl).
18
 
19
  ## Quick start
 
29
 
30
  ## Training procedure
31
 
32
+ [<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/chenyukang2020-nvidia/huggingface/runs/oi2qmn1n)
33
 
34
 
35
  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).
all_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "total_flos": 0.0,
3
- "train_loss": 1.0038919752364893e-06,
4
- "train_runtime": 66.6209,
5
  "train_samples": 35735,
6
- "train_samples_per_second": 3842.636,
7
- "train_steps_per_second": 7.505
8
  }
 
1
  {
2
  "total_flos": 0.0,
3
+ "train_loss": 1.338229644823614e-07,
4
+ "train_runtime": 85.7242,
5
  "train_samples": 35735,
6
+ "train_samples_per_second": 2986.32,
7
+ "train_steps_per_second": 5.833
8
  }
model-00001-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:68f6c005f51e6799b89abb9e83cd85053f8f05a54a4f194f6cfa64fc76e42de3
3
  size 4957560304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b1d502d4f82db4644242a755b9b94d4ffc8f3d1b72853af66ecd31a9576f466
3
  size 4957560304
model-00002-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:46f21cf01ad1a1a895444093101daa61bffe73440097137650e811c072a6d2b3
3
  size 1214366696
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6206f3f3cc2b9baa25ae4aa104637ab3541a459fab63b307ba4d9214702c398e
3
  size 1214366696
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "total_flos": 0.0,
3
- "train_loss": 1.0038919752364893e-06,
4
- "train_runtime": 66.6209,
5
  "train_samples": 35735,
6
- "train_samples_per_second": 3842.636,
7
- "train_steps_per_second": 7.505
8
  }
 
1
  {
2
  "total_flos": 0.0,
3
+ "train_loss": 1.338229644823614e-07,
4
+ "train_runtime": 85.7242,
5
  "train_samples": 35735,
6
+ "train_samples_per_second": 2986.32,
7
+ "train_steps_per_second": 5.833
8
  }
trainer_state.json CHANGED
The diff for this file is too large to render. See raw diff
 
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:de80784f5e91d7fd3b630ab022d6b5968dd6d5ce36dbd966c41599d90a607619
3
  size 8504
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:23bee333f99f395c381089c2b3b9115dcc7dea96675d14499a0f8209ea7f8cbe
3
  size 8504