Yuxuan Zhang comments

Results 479 comments of


                                            Yuxuan Zhang

Collaboration with xDiT, where xDiT is a DiT parallel inference framework

Delighted to hear about your support! We also look forward to optimizations in parallel processing and welcome any exchanges. We eagerly anticipate faster and more cost-effective deployments in the future....

5B generation blank generations.

"Could you post the specific differences in the code (from cli_demo.py), as this will help us solve the problem."

TypeError: ChatGLMTokenizer._pad() got an unexpected keyword argument 'padding_side'

已经收到这个问题，预计节后投入人力解决

ModuleNotFoundError: No module named 'huggingface_hub.inference._text_generation'

这个不是TGI的模型

[Model Requests] Add support for GLM-4 series

Thank you for your awesome work! Any updates on this issue? Your efforts are greatly appreciated.

KeyError: 'ops' when full fine-tuning CogVideoX1.5-5B-SAT

SAT 1.5版本无法进行微调（现有代码），我们做了diffusers版本的，请关注 cogvideox-factory

8G显存运行 basic_demo 加载到29%退出，无报错

你这个是量化的，量化的时候吃内存，你这个闪退应该是内存不够，内存要16G，显存8G够

模型不能输出attention吗

output_attentions is not implemented. 目前模型没有支持这个功能

使用finetune_demo/finetune_hf.py的lora微调之后，用finetune_demo/inference_hf.py做推理了的时候加载权重后无反应

只有这个ckpt不正常吗

使用finetune_demo/finetune_hf.py的lora微调之后，用finetune_demo/inference_hf.py做推理了的时候加载权重后无反应

额，这个错误少见，我今天改了GLM4 GLM3都没遇到，复现不了。。。