leonMW/DeepSeek-R1-Distill-Qwen-1.5B-GSPO-Basic-Easy

Hugging Face
TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:1.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Sep 1, 2025Architecture:Transformer Featherless Exclusive Warm

The leonMW/DeepSeek-R1-Distill-Qwen-1.5B-GSPO-Basic-Easy is a 1.5 billion parameter language model, fine-tuned from leonMW/DeepSeek-R1-Distill-Qwen-1.5B-GSPO-Basic. It utilizes the GRPO (Generalized Reinforcement Learning from Policy Optimization) method, as introduced in the DeepSeekMath paper, for its training procedure. This model is specifically optimized for enhanced reasoning capabilities, particularly in mathematical contexts, making it suitable for tasks requiring logical inference and problem-solving. It has a context length of 32768 tokens.

Loading preview...

Model Overview

This model, leonMW/DeepSeek-R1-Distill-Qwen-1.5B-GSPO-Basic-Easy, is a 1.5 billion parameter language model derived from leonMW/DeepSeek-R1-Distill-Qwen-1.5B-GSPO-Basic. It has been fine-tuned using the TRL (Transformer Reinforcement Learning) library.

Key Training Methodology

A core differentiator of this model is its training with GRPO (Generalized Reinforcement Learning from Policy Optimization). This method was originally introduced in the research paper "DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models" (arXiv:2402.03300). The application of GRPO suggests an emphasis on improving reasoning and problem-solving abilities, particularly in domains that benefit from structured logical progression.

Capabilities and Use Cases

Given its foundation and training methodology, this model is well-suited for:

  • Reasoning-intensive tasks: Its GRPO-based training implies an optimization for tasks requiring logical inference.
  • Mathematical problem-solving: Drawing from the DeepSeekMath paper's methodology, it is likely to perform well in mathematical reasoning contexts.
  • General text generation: As a fine-tuned language model, it can handle various text generation prompts, as demonstrated by the quick start example.

Technical Details

The model leverages a substantial context length of 32768 tokens, allowing it to process and generate longer sequences of text while maintaining coherence. The training utilized specific versions of key frameworks:

  • TRL: 0.23.1
  • Transformers: 4.57.1
  • Pytorch: 2.8.0
  • Datasets: 4.4.1
  • Tokenizers: 0.22.1

Finetunes
1 models