Downloads · 30 days
4
11% of all-time downloads
logan7000/cogrpo-n3-selfpeers-qwen25-3b-x-llama32-3b-x-qwen3-1p7b-math345-groupB-llama32-end
cogrpo-n3-selfpeers-qwen25-3b-x-llama32-3b-x-qwen3-1p7b-math345-groupB-llama32-end is a reinforcement learning model from logan7000. Use it for the reinforcement learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
Fine-tuned from meta-llama/Llama-3.2-3B-Instruct with co-GRPO-DP N=3, selfpluspeers: my own vote joins the two peers', so a 2-vs-1 split resolves instead of being discarded. - Co-trained with: Qwen2.5-3B x Llama-3.2-3…
Downloads · 30 days
4
11% of all-time downloads
All-time downloads
38
Public
Parameters
175K
6.4 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors6.4 GB · 100%
From the Hugging Face model README
Fine-tuned from meta-llama/Llama-3.2-3B-Instruct with co-GRPO-DP N=3, self_plus_peers: my own vote joins the two peers', so a 2-vs-1 split resolves instead of being discarded.
Co-trained with: Qwen2.5-3B x Llama-3.2-3B x Qwen3-1.7B-Base (this repo is group B)
Training data: q1716523669/MATH-Level345 (8,740 problems, MATH levels 3-5)
Eval: MATH-500 (500 prompts, pass@1, temperature 0.6)
This checkpoint: step 136 of 136 (2 epochs) — the end one
Its score: 0.5280
best is the step with the highest eval; end is step 136. Both are published
because they answer different questions and the gap between them is a result in
itself. The sibling repo is cogrpo-n3-selfpeers-qwen25-3b-x-llama32-3b-x-qwen3-1p7b-math345-groupB-llama32-best.
Effective batch 128 prompts per optimizer step, lr 3e-6, 2 epochs, 12
generations per prompt, 3072-token completions, temperature 1.0 for rollouts and
0.6 for eval, beta 0, bnpo loss, scale_rewards group, seed 42.
| step | MATH-500 pass@1 |
|---|---|
| 0 | 0.4140 |
| 10 | 0.4900 |
| 20 | 0.4960 |
| 30 | 0.5120 |
| 40 | 0.5060 |
| 50 | 0.5240 |
| 60 | 0.5180 |
| 70 | 0.5080 |
| 80 | 0.5280 |
| 90 | 0.5220 |
| 100 | 0.5340 |
| 110 | 0.5100 |
| 120 | 0.5140 |
| 130 | 0.5280 |
train.log — the complete training log this checkpoint came fromeval_curve.csv — the table above, machine-readablerun_config.json — resolved config as the trainer saw itCode: williamium3000/trl-projects,
branch n3-interaction-modes, projects/co-grpo-dp/.