50 lines
1.5 KiB
Markdown
50 lines
1.5 KiB
Markdown
|
---
|
||
|
|
license: apache-2.0
|
||
|
|
base_model: laion/r2egym-nl2bash-stack-bugsseq-fixthink-again
|
||
|
|
tags:
|
||
|
|
- reinforcement-learning
|
||
|
|
- code
|
||
|
|
- pymethods2test
|
||
|
|
- r2egym
|
||
|
|
- rl
|
||
|
|
- rloo-n
|
||
|
|
- terminus-structured
|
||
|
|
language:
|
||
|
|
- en
|
||
|
|
pipeline_tag: text-generation
|
||
|
|
library_name: transformers
|
||
|
|
---
|
||
|
|
|
||
|
|
# rl_pymethods2test-r2egym_terminus-structured
|
||
|
|
|
||
|
|
RL-trained Qwen3-8B with structured tool calls (terminus-structured agent).
|
||
|
|
|
||
|
|
## Training Pipeline
|
||
|
|
|
||
|
|
SFT (r2egym+nl2bash+swesmith) → RL mixed dataset (37 steps) → RL full r2egym (55 steps) → RL pymethods2test (156 steps, full epoch)
|
||
|
|
|
||
|
|
## Key Results
|
||
|
|
|
||
|
|
- **SWEBench-100 pass@3**: 37-42% (depending on eval run)
|
||
|
|
- **Pymethods2test pass@8**: 91-97%
|
||
|
|
- **SWEBench in-train eval**: up to 14 fully solved at various checkpoints
|
||
|
|
- Training on test-writing (pymethods) maintains code-editing ability (SWEBench)
|
||
|
|
|
||
|
|
## Training Details
|
||
|
|
|
||
|
|
- **Base model**: [laion/r2egym-nl2bash-stack-bugsseq-fixthink-again](https://huggingface.co/laion/r2egym-nl2bash-stack-bugsseq-fixthink-again)
|
||
|
|
- **Agent**: terminus-structured (bash, view, edit, create, search tools)
|
||
|
|
- **Algorithm**: RLOO-N
|
||
|
|
- **Learning rate**: 1e-5
|
||
|
|
- **Context**: 32k (24k input + 8k output)
|
||
|
|
- **Framework**: BenSkyRL + Harbor (JSC HPC)
|
||
|
|
|
||
|
|
## Usage
|
||
|
|
|
||
|
|
```python
|
||
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
||
|
|
|
||
|
|
model = AutoModelForCausalLM.from_pretrained("laion/rl_pymethods2test-r2egym_terminus-structured")
|
||
|
|
tokenizer = AutoTokenizer.from_pretrained("laion/rl_pymethods2test-r2egym_terminus-structured")
|
||
|
|
```
|