Downloads · 30 days
0
alirezaaminzadeh/or-reasoning-7b-grpo
or-reasoning-7b-grpo is a text generation model from alirezaaminzadeh. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.
Second-stage GRPO (Group Relative Policy Optimization) fine-tuning with solver-verifiable rewards.
Downloads · 30 days
0
Access
Public
Updated Aug 9, 2026
Repo size
—
Likes
0
Public
Click a slice to open those files.
Other1.5 KB · 51%
From the Hugging Face model README
Second-stage GRPO (Group Relative Policy Optimization) fine-tuning with solver-verifiable rewards.
Rewards are computed from actual solver/compiler feedback on generated Pyomo code:
| Signal | Reward |
|---|---|
| Code compiles | +1 |
| Feasible solution | +2 |
| Correct objective (≤15% error) | +4 |
| Optimal objective (≤2% error) | +8 |
| Parameter | Value |
|---|---|
| Base | alirezaaminzadeh/or-reasoning-7b-sft |
| Method | GRPO (TRL) |
| Reward | Custom solver-verifiable function |
| Prompts | code_generation tasks from SFT corpus |
| Hardware | a10g-large (HF Jobs) |
Use for code generation and repair tasks where solver feedback quality matters most.
from transformers import pipeline
pipe = pipeline("text-generation", model="alirezaaminzadeh/or-reasoning-7b-grpo")