Qw Q 32 B Preview Idea Whiz v1

Introduction

The QwQ-32B-Preview-IdeaWhiz-v1 model is a fine-tuned version optimized for scientific creativity and step-by-step reasoning. It leverages the LiveIdeaBench dataset to enhance its capabilities in generating novel scientific ideas and hypotheses.

Architecture

• Base Model: QwQ-32B-Preview
• Training Dataset: LiveIdeaBench
• Focus: Scientific creativity and idea generation
• Reasoning Style: o1-style step-by-step reasoning

Training

The model was trained using the LiveIdeaBench dataset, which is designed to evaluate the scientific creativity and idea generation capabilities of large language models (LLMs) with minimal context. For more information, see the associated paper on arXiv: 2412.17596.

Guide: Running Locally

1. Installation: Ensure you have Python and the required packages installed. Use pip install transformers to get the Hugging Face Transformers library.

2. Model Loading:

   from transformers import AutoModelForCausalLM, AutoTokenizer
   
   model_name = "6cf/QwQ-32B-Preview-IdeaWhiz-v1"
   model = AutoModelForCausalLM.from_pretrained(
       model_name,
       torch_dtype="auto",
       device_map="auto"
   )
   tokenizer = AutoTokenizer.from_pretrained(model_name)
   

3. Running Inference:

   prompt = "Your scientific prompt here"
   messages = [{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt}]
   text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
   model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
   
   generated_ids = model.generate(**model_inputs, max_new_tokens=4096)
   response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
   

4. Hardware Recommendations: Running this model locally requires significant computational resources. It is advisable to use cloud GPU services such as AWS, GCP, or Azure for efficient processing.

License

This model is licensed under the Apache License 2.0. You can freely use, modify, and distribute it, but must include the original license and notices.