OpenGVLab / Ask-Anything

[CVPR2024 Highlight][VideoChatGPT] ChatGPT with video understanding! And many more supported LMs such as miniGPT4, StableLM, and MOSS.
https://vchat.opengvlab.com/
MIT License
2.85k stars 230 forks source link

stage1无法eval #156

Closed LanXingXuan closed 2 months ago

LanXingXuan commented 3 months ago

在训练stage1过程中,evaluate报错

models/videochat2_qformer.py 中的返回的是None def get_text_encoder(self): return None

导致tasks/retrieval_utils.py的143行中text_encoder = model.get_text_encoder()读不到,造成227-245行 output = text_encoder( encoder_embeds=text_feats[topk_idx[j:]], attention_mask=text_atts[topk_idx[j:]], encoder_hidden_states=left_encoder_output, encoder_attention_mask=left_encoder_att, return_dict=True, mode="fusion", ) 运行时报错,进而stage1训练过程中无法eval

请问应该怎么修改代码?

riariam commented 2 months ago

你好,我也遇到了同样的问题,请问您解决了吗