Instructions to use alireza7/GrepSeek-Qwen3.5-9B-SFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use alireza7/GrepSeek-Qwen3.5-9B-SFT with Transformers:

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

pipe = pipeline("text-generation", model="alireza7/GrepSeek-Qwen3.5-9B-SFT")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("alireza7/GrepSeek-Qwen3.5-9B-SFT")
model = AutoModelForMultimodalLM.from_pretrained("alireza7/GrepSeek-Qwen3.5-9B-SFT", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use alireza7/GrepSeek-Qwen3.5-9B-SFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "alireza7/GrepSeek-Qwen3.5-9B-SFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "alireza7/GrepSeek-Qwen3.5-9B-SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/alireza7/GrepSeek-Qwen3.5-9B-SFT

SGLang

How to use alireza7/GrepSeek-Qwen3.5-9B-SFT 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 "alireza7/GrepSeek-Qwen3.5-9B-SFT" \
    --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": "alireza7/GrepSeek-Qwen3.5-9B-SFT",
		"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 "alireza7/GrepSeek-Qwen3.5-9B-SFT" \
        --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": "alireza7/GrepSeek-Qwen3.5-9B-SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use alireza7/GrepSeek-Qwen3.5-9B-SFT with Docker Model Runner:
```
docker model run hf.co/alireza7/GrepSeek-Qwen3.5-9B-SFT
```

GrepSeek-Qwen3.5-9B-SFT

The cold-start SFT policy for GrepSeek — a Direct Corpus Interaction (DCI) search agent that answers questions by issuing Unix shell commands (rg, grep, head, …) directly over a 21M-passage Wikipedia corpus, instead of retrieving from a dense/sparse index. This checkpoint is Qwen/Qwen3.5-9B supervised-fine-tuned on the cold-start trajectories; it is the initialization for the RL stage and corresponds to the "w/o GRPO" ablation in the paper.

📄 GrepSeek: Training Search Agents for Direct Corpus Interaction · 💻 https://github.com/alirezasalemi7/grepseek

Base: Qwen/Qwen3.5-9B
Training data: alireza7/GrepSeek-ColdStart-SFT-10k (5k NQ + 5k HotpotQA, Answer-Aware-Tutor / Answer-Blind-Planner trajectories)
Recipe: verl FSDP, 1 epoch, AdamW, peak LR 5e-6 (constant + 5% warmup), 16,384-token sequences, bf16, Ulysses SP=4, on 4×A100-80GB.
RL-optimized successor: alireza7/GrepSeek-Qwen3.5-9B-GRPO

Why SFT first: RL directly from the base model over millions of documents is unstable (overly broad queries, context blow-ups, OOM). Cold-start SFT instills concise, causally-grounded shell-search behavior — establishing the low-level retrieval "primitives" (fixed-string -F matching, | head truncation, cascaded rg ... | rg ... filtering) that RL later refines.

⚠️ A tool-using agent, not a standalone chatbot

The model emits <tool_call> shell commands that must be executed against the Wikipedia corpus and fed back as <tool_response> turns. To use it you need: (1) the corpus PeterJinGo/wiki-18-corpus, (2) a tool-calling vLLM server, and (3) the GrepSeek inference harness (grep tool

agent loop), all in the code repo.

Usage

git clone https://github.com/alirezasalemi7/grepseek && cd grepseek
# env: TRAINING_ENV.md  ·  corpus: cold_start_sft/download_corpus.py

# 1. serve this checkpoint
MODEL_PATH=alireza7/GrepSeek-Qwen3.5-9B-SFT bash rl/serve_rl.sh         # -> http://localhost:10730/v1

# 2. run the agent (paper inference: temperature 0.6, <=6 turns, 16k context)
GREPSEEK_CORPUS_ROOT=/path/to/wiki_18_corpus \
  bash inference/run_inference.sh --base_url http://localhost:10730/v1 \
    --model grepseek --temperature 0.6 --input my_questions.jsonl --out_dir out

Evaluation (token-F1 / EM, micro-average over 7 QA benchmarks)

This SFT-only policy already substantially beats the untuned base model, but RL adds large gains on multi-hop reasoning:

variant	micro-avg F1	micro-avg EM
base (no SFT, no RL)	0.3314	0.2836
this model (SFT only)	0.4249	0.3569
+ GRPO → `GrepSeek-Qwen3.5-9B-GRPO`	0.5691	0.4948

(7 benchmarks: NQ, TriviaQA, PopQA, HotpotQA, 2WikiMultiHopQA, MuSiQue, Bamboogle; trained only on NQ + HotpotQA, the rest are out-of-distribution.)

License

Inherits the license of the base model Qwen/Qwen3.5-9B — confirm and update the license field above if needed.

Citation

@misc{salemi2026grepseektrainingsearchagents,
      title={GrepSeek: Training Search Agents for Direct Corpus Interaction},
      author={Alireza Salemi and Chang Zeng and Atharva Nijasure and Jui-Hui Chung and Razieh Rahimi and Fernando Diaz and Hamed Zamani},
      year={2026},
      eprint={2605.29307},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2605.29307},
}