AffectGPT-RL

Unofficial mirror of the original authors' released Stage-2 checkpoints. Model weights are unchanged. This repository is maintained by mcflurryshuoz and is not an official author release.

Original project: https://github.com/zeroQiaoba/AffectGPT/tree/master/AffectGPT-RL
Original checkpoint download: https://pan.baidu.com/s/11LfKslZA2KMoLCRwkKtnVA?pwd=45jk
Base language model: Qwen2.5-7B-Instruct, with CLIP ViT-L/14 and chinese-hubert-large components.

This repository contains all 12 original .pth checkpoints, for steps 5 through 60. These are project checkpoints, not a standalone Transformers model. Use the original code, base models and Stage-2 configuration for inference; checkpoint filenames are preserved.

The upstream README describes the release as Apache 2.0 and the service as a research preview for non-commercial use only. See UPSTREAM_TERMS.md and the original README. No alternative licensing terms are assigned by this mirror.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for mcflurryshuoz/AffectGPT-RL

Base model

Qwen/Qwen2.5-7B

Finetuned

Qwen/Qwen2.5-7B-Instruct

Finetuned

(3,258)

this model