Deepthink Reasoning 7 B

Introduction

Deepthink-Reasoning-7B is a fine-tuned version of the Qwen2.5-7B-Instruct base model, primarily designed for text generation tasks demanding deep reasoning, logical structuring, and problem-solving. It is optimized for applications in education, programming, and creative writing, providing accurate and contextually relevant outputs for complex queries.

Architecture

This model features advanced natural language processing capabilities, excelling in generating step-by-step solutions, creative content, and logical analyses. It supports long-context inputs up to 128K tokens and can generate outputs up to 8K tokens. It also offers multilingual support for over 29 languages, including Chinese, English, French, Spanish, and more. The model is enhanced for instruction following, generating long texts, understanding structured data, and role-play implementation.

Training

Deepthink-Reasoning-7B is trained with specialized expert models in coding and mathematics, resulting in significant improvements in these areas. It is also trained to better follow instructions and generate structured outputs like JSON, making it resilient to diverse system prompts.

Guide: Running Locally

1. Install Requirements: Ensure you have the necessary libraries by installing transformers and torch.
2. Load the Model: Use the following Python code snippet to load the tokenizer and model:

   from transformers import AutoModelForCausalLM, AutoTokenizer
   
   model_name = "prithivMLmods/Deepthink-Reasoning-7B"
   
   model = AutoModelForCausalLM.from_pretrained(
       model_name,
       torch_dtype="auto",
       device_map="auto"
   )
   tokenizer = AutoTokenizer.from_pretrained(model_name)
   

3. Generate Text: Prepare your input and generate responses using the model's capabilities.
4. Recommended Setup: For optimal performance, use cloud GPUs such as those offered by AWS, Google Cloud, or Azure.

License

Deepthink-Reasoning-7B is released under the CreativeML OpenRAIL-M license, which governs its use and distribution.