1# Interactive chat (recommended)
2uvx --from mlx-lm mlx_lm.chat --model Plurigrid/DR-Tulu-8B-MLX-4bit
3
4# Generate text
5uvx --from mlx-lm mlx_lm.generate --model Plurigrid/DR-Tulu-8B-MLX-4bit --prompt "What is category theory?" --max-tokens 500
1from mlx_lm import load, generate
2
3# Load the 4-bit quantized model
4model, tokenizer = load("Plurigrid/DR-Tulu-8B-MLX-4bit")
5
6prompt = "Explain quantum computing step by step."
7
8# Apply chat template if available
9if tokenizer.chat_template is not None:
10 messages = [{"role": "user", "content": prompt}]
11 prompt = tokenizer.apply_chat_template(
12 messages, add_generation_prompt=True
13 )
14
15# Generate response with DR-Tulu reasoning
16response = generate(model, tokenizer, prompt=prompt, verbose=True)
17print(response)
1# Install MLX-LM
2pip install mlx-lm
3# or with uv
4uv add mlx-lm
This is the RL checkpoint of DR Tulu, an open deep research agent trained on top of
rl-research/DR-Tulu-SFT-8B.
For more details on DR Tulu, please
read our paper!
1messages = [
2 {"role": "user", "content": "What is category theory?"},
3 {"role": "assistant", "content": "Category theory is a mathematical framework..."},
4 {"role": "user", "content": "How does it apply to computer science?"}
5]
6
7if tokenizer.chat_template is not None:
8 formatted_prompt = tokenizer.apply_chat_template(messages, add_generation_prompt=True)
9 response = generate(model, tokenizer, prompt=formatted_prompt, max_tokens=1000)
1research_prompt = """
2Analyze the relationship between quantum mechanics and information theory.
3Think step by step and provide a comprehensive analysis.
4"""
5
6response = generate(model, tokenizer, prompt=research_prompt, max_tokens=1500, verbose=True)
This model is licensed under
Apache 2.0. It is intended for research and educational use in accordance with
Ai2's Responsible Use Guidelines.
1@article{drtulu,
2 title = {{DR Tulu: Reinforcement Learning with Evolving Rubrics for Deep Research}},
3 author = {{Rulin Shao, Akari Asai, Shannon Shen, Hamish Ivison, Varsha Kishore, Jingming Zhuo, Xinran Zhao, Molly Park, Sam Finlayson, David Sontag, Tyler Murray, Sewon Min, Pradeep Dasigi, Luca Soldani, Faeze Brahman, Scott Yih, Sherry Tongshuang Wu, Luke Zettlemoyer, Yoon Kim, Hanna Hajishirzi, Pang Wei Koh}},
4 year = {2025},
5}
