Instructions to use LARK-Lab/EnvFactory-1.7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use LARK-Lab/EnvFactory-1.7B with Transformers:

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

pipe = pipeline("text-generation", model="LARK-Lab/EnvFactory-1.7B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("LARK-Lab/EnvFactory-1.7B")
model = AutoModelForCausalLM.from_pretrained("LARK-Lab/EnvFactory-1.7B")
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 LARK-Lab/EnvFactory-1.7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "LARK-Lab/EnvFactory-1.7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "LARK-Lab/EnvFactory-1.7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/LARK-Lab/EnvFactory-1.7B

SGLang

How to use LARK-Lab/EnvFactory-1.7B 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 "LARK-Lab/EnvFactory-1.7B" \
    --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": "LARK-Lab/EnvFactory-1.7B",
		"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 "LARK-Lab/EnvFactory-1.7B" \
        --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": "LARK-Lab/EnvFactory-1.7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use LARK-Lab/EnvFactory-1.7B with Docker Model Runner:
```
docker model run hf.co/LARK-Lab/EnvFactory-1.7B
```

RolandXMR commited on 1 day ago

Commit

6e578be

verified ·

1 Parent(s): 7097e55

Upload README.md with huggingface_hub

Browse files

Files changed (1) hide show

README.md +6 -8

README.md CHANGED Viewed

@@ -56,7 +56,7 @@ base_model:
 We propose **EnvFactory**, a fully automated framework that addresses the challenges of equipping LLMs with tool-use capabilities via Agentic Reinforcement Learning (Agentic RL). EnvFactory autonomously explores and verifies stateful, executable tool environments from authentic resources, and synthesizes natural multi-turn trajectories through topology-aware sampling and calibrated refinement, producing grounded queries with implicit intents.
-This model is the official **EnvFactory-1.7B** trained from Qwen/Qwen3-1.7B using SFT and RL on synthesized tool-use trajectories.
 ## Key Features
@@ -83,14 +83,12 @@ This model is the official **EnvFactory-1.7B** trained from Qwen/Qwen3-1.7B usin
 ## Performance
-EnvFactory models achieve significant improvements on tool-use benchmarks:
-| Benchmark | Improvement |
-|-----------|-------------|
-| BFCLv3 | +15% |
-| MCP-Atlas | +8.6% |
-| τ²-Bench | +6% |
-| VitaBench | +6% |
 ## Usage

 We propose **EnvFactory**, a fully automated framework that addresses the challenges of equipping LLMs with tool-use capabilities via Agentic Reinforcement Learning (Agentic RL). EnvFactory autonomously explores and verifies stateful, executable tool environments from authentic resources, and synthesizes natural multi-turn trajectories through topology-aware sampling and calibrated refinement, producing grounded queries with implicit intents.
+This model is the official **EnvFactory-1.7B** trained from [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B) using SFT and RL on synthesized tool-use trajectories.
 ## Key Features
 ## Performance
+Results on tool-use benchmarks compared to the base model:
+| Model | BFCL Single Turn | BFCL Multi Turn | MCP-Atlas Pass Rate | MCP-Atlas Mean Cov. | τ²-Bench Avg. | VitaBench Avg. | Overall Avg. |
+|-------|------------------|-----------------|---------------------|---------------------|---------------|----------------|--------------|
+| Qwen3-1.7B (Base) | 79.48 | 16.75 | 1.03 | 6.25 | 14.61 | 1.33 | 16.27 |
+| **EnvFactory-1.7B** | 78.44 | 28.38 | 3.09 | 9.64 | 15.11 | 7.33 | 19.74 |
 ## Usage