Instructions to use TIGER-Lab/FIM-Mid-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TIGER-Lab/FIM-Mid-8B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="TIGER-Lab/FIM-Mid-8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("TIGER-Lab/FIM-Mid-8B")
model = AutoModelForCausalLM.from_pretrained("TIGER-Lab/FIM-Mid-8B", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use TIGER-Lab/FIM-Mid-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TIGER-Lab/FIM-Mid-8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TIGER-Lab/FIM-Mid-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/TIGER-Lab/FIM-Mid-8B

SGLang

How to use TIGER-Lab/FIM-Mid-8B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "TIGER-Lab/FIM-Mid-8B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TIGER-Lab/FIM-Mid-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "TIGER-Lab/FIM-Mid-8B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TIGER-Lab/FIM-Mid-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use TIGER-Lab/FIM-Mid-8B with Docker Model Runner:
```
docker model run hf.co/TIGER-Lab/FIM-Mid-8B
```

ubowang commited on 17 days ago

Commit

21de27a

verified ·

1 Parent(s): 872402e

Update model card: official GitHub/dataset links, unified training configs, results, citation

Browse files

Files changed (1) hide show

README.md +23 -8

README.md CHANGED Viewed

@@ -4,6 +4,8 @@ library_name: transformers
 pipeline_tag: text-generation
 base_model:
 - Qwen/Qwen3-8B
 tags:
 - code
 - software-engineering
@@ -12,15 +14,17 @@ tags:
 # FIM-Mid-8B
-**FIM-Mid-8B** is the mid-trained checkpoint of the FIM 8B pipeline: `Qwen3-8B` after FIM mid-training, **before** agent post-training. Post-training this checkpoint on SWE-Lego trajectories produces [TIGER-Lab/FIM-8B](https://huggingface.co/TIGER-Lab/FIM-8B).
-## Model
-Local path: `models/FIM-Mid-8B/` (checkpoints are gitignored; do not commit them).
-- Base model: `Qwen/Qwen3-8B`
-- FIM mid-training: `train/FIM_Midtrain_8B.yaml` (AdamW, lr `1.0e-5`, cosine schedule, warmup ratio `0.1`, weight decay `0.05`, one epoch, sequence length `32768`, bf16)
-- Post-training: none — see [TIGER-Lab/FIM-8B](https://huggingface.co/TIGER-Lab/FIM-8B) for the post-trained agent model
 ## Serve with vLLM
@@ -29,7 +33,7 @@ Ships the native Qwen3 40960 context (the yarn extension to 163840 was applied a
 ```bash
 CUDA_VISIBLE_DEVICES=0 \
 python -m vllm.entrypoints.openai.api_server \
-  --model models/FIM-Mid-8B \
   --served-model-name FIM-Mid-8B \
   --host 127.0.0.1 \
   --port 8400 \
@@ -41,4 +45,15 @@ python -m vllm.entrypoints.openai.api_server \
 ## Post-training
-To reproduce FIM-8B, run SWE-Lego trajectory SFT from this checkpoint (LLaMA-Factory, full fine-tuning, lr `1.0e-4`, 4 epochs, cutoff 131072 with yarn rope scaling, `qwen3_nothink` template, error masking enabled).

 pipeline_tag: text-generation
 base_model:
 - Qwen/Qwen3-8B
+datasets:
+- TIGER-Lab/FIM-Midtraining-400K
 tags:
 - code
 - software-engineering
 # FIM-Mid-8B
+[📄 Paper (PDF)](https://github.com/TIGER-AI-Lab/FIM-Midtraining/blob/main/paper.pdf) · [💻 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-Mid-8B** is the mid-trained checkpoint of the FIM 8B pipeline: `Qwen3-8B` after function-aware FIM mid-training, **before** agent post-training. Post-training this checkpoint on SWE-Lego trajectories produces [TIGER-Lab/FIM-8B](https://huggingface.co/TIGER-Lab/FIM-8B).
+It is released for reproducibility and further post-training. The paper deliberately never scores mid-training-only checkpoints — a FIM-only model has degraded instruction-following and cannot be compared fairly against instruction-tuned baselines; every reported gain is one that *survives* post-training.
+## Training
+- **Base model**: [`Qwen/Qwen3-8B`](https://huggingface.co/Qwen/Qwen3-8B)
+- **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) — AdamW, lr `1.0e-5`, cosine schedule, warmup ratio `0.1`, weight decay `0.05`, one epoch, sequence length `32768`, bf16 (as-run copy: [`FIM_Midtrain_8B.yaml`](https://github.com/TIGER-AI-Lab/FIM-Midtraining/blob/main/midtraining/configs/FIM_Midtrain_8B.yaml))
+- **Post-training**: none — see [TIGER-Lab/FIM-8B](https://huggingface.co/TIGER-Lab/FIM-8B) for the post-trained agent model
 ## Serve with vLLM
 ```bash
 CUDA_VISIBLE_DEVICES=0 \
 python -m vllm.entrypoints.openai.api_server \
+  --model TIGER-Lab/FIM-Mid-8B \
   --served-model-name FIM-Mid-8B \
   --host 127.0.0.1 \
   --port 8400 \
 ## Post-training
+To reproduce FIM-8B, run SWE-Lego trajectory SFT from this checkpoint — the exact config is [`posttraining/swe_lego/FIM_Posttrain_8B.yaml`](https://github.com/TIGER-AI-Lab/FIM-Midtraining/blob/main/posttraining/swe_lego/FIM_Posttrain_8B.yaml) (LLaMA-Factory, full fine-tuning, lr `1.0e-4`, **2 epochs** — the official SWE-Lego recipe's 4 overfits this base — cutoff 131072 with yarn rope scaling, `qwen3_nothink` template, `turn_mask` enabled), which already points at this repo id. See [`posttraining/swe_lego/`](https://github.com/TIGER-AI-Lab/FIM-Midtraining/tree/main/posttraining/swe_lego) for the walkthrough.
+## 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},
+  year={2026}
+}
+```