110 lines
4.5 KiB
Markdown
110 lines
4.5 KiB
Markdown
---
|
|
license: apache-2.0
|
|
library_name: transformers
|
|
pipeline_tag: text-generation
|
|
base_model:
|
|
- Qwen/Qwen2.5-Coder-7B-Instruct
|
|
datasets:
|
|
- TIGER-Lab/FIM-Midtraining-400K
|
|
- R2E-Gym/R2EGym-SFT-Trajectories
|
|
tags:
|
|
- code
|
|
- software-engineering
|
|
- agent
|
|
---
|
|
|
|
# FIM-7B
|
|
|
|
[📄 Paper](https://arxiv.org/abs/2607.12463) · [💻 GitHub](https://github.com/TIGER-AI-Lab/FIM-Midtraining) · [🤗 Dataset](https://huggingface.co/datasets/TIGER-Lab/FIM-Midtraining-400K) · [🤗 Collection](https://huggingface.co/collections/TIGER-Lab/fim-midtraining)
|
|
|
|
**FIM-7B** is the 7B coding-agent model of *"Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models"*: `Qwen2.5-Coder-7B-Instruct`, mid-trained on function-aware FIM data, then post-trained on R2E-Gym agent trajectories with the upstream recipe unmodified. The mid-training stage is the only difference from a standard R2E-Gym reproduction — and it is worth **+2.8 points on SWE-Bench-Verified and +3.7 on SWE-Bench-Lite**.
|
|
|
|
## Training pipeline
|
|
|
|
- **Base model**: [`Qwen/Qwen2.5-Coder-7B-Instruct`](https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct)
|
|
- **FIM mid-training**: [`midtraining/configs/fim_midtrain.yaml`](https://github.com/TIGER-AI-Lab/FIM-Midtraining/blob/main/midtraining/configs/fim_midtrain.yaml) on [TIGER-Lab/FIM-Midtraining-400K](https://huggingface.co/datasets/TIGER-Lab/FIM-Midtraining-400K) (as-run copy: [`FIM_Midtrain_7B.yaml`](https://github.com/TIGER-AI-Lab/FIM-Midtraining/blob/main/midtraining/configs/FIM_Midtrain_7B.yaml)) → intermediate checkpoint released as [TIGER-Lab/FIM-Mid-7B](https://huggingface.co/TIGER-Lab/FIM-Mid-7B)
|
|
- **Post-training**: SFT on R2E-Gym agent trajectories — [`posttraining/r2egym/`](https://github.com/TIGER-AI-Lab/FIM-Midtraining/tree/main/posttraining/r2egym) (as-run copy: [`FIM_Posttrain_7B.yaml`](https://github.com/TIGER-AI-Lab/FIM-Midtraining/blob/main/posttraining/r2egym/FIM_Posttrain_7B.yaml))
|
|
|
|
## Results
|
|
|
|
Means over three evaluation seeds, identical harness for both arms (paper Table 1):
|
|
|
|
| Setting | SWE-Bench-Verified | SWE-Bench-Lite |
|
|
|---|---|---|
|
|
| Qwen2.5-Coder-7B-Instruct + R2E-Gym (reproduced) | 15.00 | 11.33 |
|
|
| **FIM-7B (+ FIM mid-training)** | **17.80** | **15.00** |
|
|
| Δ | +2.80 | +3.67 |
|
|
|
|
## Evaluate on SWE-Bench Verified
|
|
|
|
FIM-7B is evaluated with the **R2E-Gym agent scaffold** (fixed by its post-training pipeline). The complete pinned walkthrough lives at [`evaluation/swebench/released_checkpoints.md`](https://github.com/TIGER-AI-Lab/FIM-Midtraining/blob/main/evaluation/swebench/released_checkpoints.md).
|
|
|
|
### 1. Serve the model with vLLM
|
|
|
|
```bash
|
|
CUDA_VISIBLE_DEVICES=0 \
|
|
VLLM_ALLOW_LONG_MAX_MODEL_LEN=1 \
|
|
python -m vllm.entrypoints.openai.api_server \
|
|
--model TIGER-Lab/FIM-7B \
|
|
--served-model-name FIM-7B \
|
|
--host 127.0.0.1 \
|
|
--port 8400 \
|
|
--tensor-parallel-size 1 \
|
|
--max-model-len 65536 \
|
|
--hf-overrides '{"max_position_embeddings": 65536}' \
|
|
--enable-prefix-caching \
|
|
--gpu-memory-utilization 0.9 \
|
|
> vllm_fim7b.log 2>&1 &
|
|
```
|
|
|
|
Wait until the server is up (model load takes ~1 minute):
|
|
|
|
```bash
|
|
curl -s http://127.0.0.1:8400/v1/models
|
|
```
|
|
|
|
### 2. Run the agent on SWE-Bench Verified
|
|
|
|
From an upstream, unmodified [R2E-Gym](https://github.com/R2E-Gym/R2E-Gym) checkout (Docker required):
|
|
|
|
```bash
|
|
export OPENAI_API_KEY=EMPTY
|
|
export LLM_BASE_URL="http://127.0.0.1:8400/v1"
|
|
|
|
uv run python src/r2egym/agenthub/run/edit.py runagent_multiple \
|
|
--dataset "R2E-Gym/SWE-Bench-Verified" \
|
|
--split "test" \
|
|
--start_idx 0 \
|
|
--k 500 \
|
|
--traj_dir "./traj" \
|
|
--exp_name "FIM-7B_swebench_verified_r1" \
|
|
--llm_name "openai/FIM-7B" \
|
|
--scaffold "r2egym" \
|
|
--backend "docker" \
|
|
--use_fn_calling False \
|
|
--temperature 0 \
|
|
--max_steps 40 \
|
|
--max_steps_absolute 100 \
|
|
--max_workers 6 \
|
|
--max_reward_calc_time 1200 \
|
|
--max_tokens 65536 \
|
|
--use_existing True
|
|
```
|
|
|
|
For SWE-Bench Lite, use `--dataset "R2E-Gym/SWE-Bench-Lite" --k 300`.
|
|
|
|
### 3. Score with the official SWE-bench harness
|
|
|
|
Convert the trajectories to a submission and score with the official harness — [`evaluation/swebench/score.sh`](https://github.com/TIGER-AI-Lab/FIM-Midtraining/blob/main/evaluation/swebench/score.sh). The reported number is `resolved_instances / total_instances`.
|
|
|
|
## Citation
|
|
|
|
```bibtex
|
|
@article{wang2026fim,
|
|
title={Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models},
|
|
author={Wang, Yubo and Liang, Jiarong and Zhang, Yuxuan and Liu, Xuye and Wei, Cong and Zhang, Yuyu and Nie, Ping and Chen, Wenhu},
|
|
journal={arXiv preprint arXiv:2607.12463},
|
|
year={2026}
|
|
}
|
|
```
|