Skip to content

EthenEthenEthen

LoginPlatformTry Ethen

Open Source Model Profile · Yuqian-Fu

SRFT-Qwen2.5-Math-7B

SRFT-Qwen2.5-Math-7B is a 7.62B-parameter Qwen2 math fine-tune from Yuqian-Fu. Its card documents the single-stage SRFT method under MIT licensing.

Publisher
Yuqian-Fu
Task
text-generation
Model type
qwen2
License
mit
Library
transformers
Publication status
Accepted · not indexed

Model overview

SRFT-Qwen2.5-Math-7B is published by Yuqian-Fu as a Qwen2 text-generation model. The captured configuration identifies Qwen2ForCausalLM with model type qwen2, and Safetensors metadata reports 7,615,616,512 parameters. According to the model card, it applies Supervised Reinforcement Fine-Tuning, while hub tags record a math dataset and a Qwen2.5-Math-7B finetune marker.

Recorded capabilities

Qwen2 math fine-tune record

Captured configuration records Qwen2ForCausalLM with model type qwen2 and about 7.62B parameters, with hub tags marking a Qwen2.5-Math-7B finetune.

SRFT method note

According to the model card, SRFT is a single-stage method unifying fine-tuning paradigms through entropy-aware weighting.

MIT Transformers record

Card data records MIT licensing with Transformers and safetensors support and conversational text-generation markers.

Use cases in the source record

  • Math-focused text-generation experiments using a Qwen2 checkpoint associated with the documented SRFT method and a math dataset tag.

Limitations and unknowns

  • No context-window value was extracted from this record.
  • No evaluation results were extracted from this record.
  • No training hyperparameters, hardware, or schedule facts were extracted beyond the method description and dataset tag.
  • Provider state is historical snapshot data and should be refreshed before being presented as current.

Source and provenance

Source: Yuqian-Fu/SRFT-Qwen2.5-Math-7B

Captured: Unknown. Processed: 2026-09-07T19:35:53.208015+00:00.

📄 Introduction Supervised Reinforcement Fine-Tuning (SRFT) is a single-stage method that unifies both fine-tuning paradigms through entropy-aware weighting mechanisms. Paper: arXiv Project Website: SRFT

F001F002F003F004F005F006F007F008F009F010F011