Instructions to use gxx27/BioTool-finetuned-Qwen3-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use gxx27/BioTool-finetuned-Qwen3-4B with Transformers:

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

pipe = pipeline("text-generation", model="gxx27/BioTool-finetuned-Qwen3-4B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("gxx27/BioTool-finetuned-Qwen3-4B")
model = AutoModelForCausalLM.from_pretrained("gxx27/BioTool-finetuned-Qwen3-4B")
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

vLLM

How to use gxx27/BioTool-finetuned-Qwen3-4B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "gxx27/BioTool-finetuned-Qwen3-4B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "gxx27/BioTool-finetuned-Qwen3-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/gxx27/BioTool-finetuned-Qwen3-4B

SGLang

How to use gxx27/BioTool-finetuned-Qwen3-4B 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 "gxx27/BioTool-finetuned-Qwen3-4B" \
    --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": "gxx27/BioTool-finetuned-Qwen3-4B",
		"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 "gxx27/BioTool-finetuned-Qwen3-4B" \
        --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": "gxx27/BioTool-finetuned-Qwen3-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use gxx27/BioTool-finetuned-Qwen3-4B with Docker Model Runner:
```
docker model run hf.co/gxx27/BioTool-finetuned-Qwen3-4B
```

gxx27 commited on 19 days ago

Commit

e5338fc

verified ·

1 Parent(s): 268c35e

Update README.md

Browse files

Files changed (1) hide show

README.md +7 -18

README.md CHANGED Viewed

@@ -20,16 +20,11 @@ pipeline_tag: text-generation
 # BioTool-finetuned-Qwen3-4B
-Qwen3-4B-Instruct-2507 fully fine-tuned on the **[BioTool](https://github.com/<user>/BioAPI)**
 training split (5,632 samples). The resulting model produces structured tool calls
 against **127 biomedical APIs** spanning NCBI E-utilities + BLAST, UniProt and
 Ensembl REST.
-This is the headline checkpoint released with:
-> Gao et al., **"BioTool: A Comprehensive Tool-Calling Dataset for Enhancing
-> Biomedical Capabilities of Large Language Models"**, arXiv:2605.05758, 2026.
 ## Quick start
 ```python
@@ -62,20 +57,14 @@ print(tok.decode(out[0][tok(prompt, return_tensors="pt")["input_ids"].shape[1]:]
                  skip_special_tokens=False))
 ```
-The model emits tool calls in the form
-```
-<tool_call>
-{"name": "lookup_by_symbol", "arguments": {"species": "human", "symbol": "BRCA1"}}
-</tool_call>
-```
 To **execute** the resulting tool call, use the Python wrappers in the
-[BioAPI repository](https://github.com/<user>/BioAPI):
 ```python
-import sys; sys.path.insert(0, "/path/to/BioAPI")
 from ensembl.lookup.api import lookup_by_symbol
 print(lookup_by_symbol(species="human", symbol="BRCA1"))
 ```
@@ -100,7 +89,7 @@ print(lookup_by_symbol(species="human", symbol="BRCA1"))
 - **Epochs:** 3
 - **Effective batch size:** 16 (per-device 1 × grad-accum 16)
 - **Precision:** bf16
-- **Hyperparameter file:** `qwen3_4b.yaml` in the BioAPI repo's
   `llamafactory_cfgs/`
 ![training loss](training_loss.png)
@@ -110,7 +99,7 @@ print(lookup_by_symbol(species="human", symbol="BRCA1"))
 On the BioTool test split (1,408 samples), this model achieves the highest
 **BioTool Score** among the open-source models we evaluated, while also being
 the smallest (4B parameters). Per-database breakdowns and head-to-head numbers
-against GPT-5.1, GPT-5.1-Codex, Claude Sonnet 4.5 and Gemini 3.1 Pro are
 reported in the paper.
 ## Citation

 # BioTool-finetuned-Qwen3-4B
+Qwen3-4B-Instruct-2507 fully fine-tuned on the **[BioTool](https://github.com/gxx27/BioTool)**
 training split (5,632 samples). The resulting model produces structured tool calls
 against **127 biomedical APIs** spanning NCBI E-utilities + BLAST, UniProt and
 Ensembl REST.
 ## Quick start
 ```python
                  skip_special_tokens=False))
 ```
 To **execute** the resulting tool call, use the Python wrappers in the
+[BioTool repository](https://github.com/gxx27/BioTool):
 ```python
+import sys
+sys.path.insert(0, "/path/to/BioTool")
 from ensembl.lookup.api import lookup_by_symbol
 print(lookup_by_symbol(species="human", symbol="BRCA1"))
 ```
 - **Epochs:** 3
 - **Effective batch size:** 16 (per-device 1 × grad-accum 16)
 - **Precision:** bf16
+- **Hyperparameter file:** `qwen3_4b.yaml` in the BioTool repo's
   `llamafactory_cfgs/`
 ![training loss](training_loss.png)
 On the BioTool test split (1,408 samples), this model achieves the highest
 **BioTool Score** among the open-source models we evaluated, while also being
 the smallest (4B parameters). Per-database breakdowns and head-to-head numbers
+against GPT-5.1, GPT-5.1-Codex, Claude Sonnet 4.5 and Gemini 3 Pro are
 reported in the paper.
 ## Citation