LiweiPE comments

Repositories
Issues
Comments

Results 3 comments of


                                            LiweiPE

Add GPTQ Marlin 2:4 sparse structured support

Nice work! I would like to try with Qwen model series. May I know where can I find the tutorial to generate 2:4 sparse model?

多机多卡训练出现问题

> > NCCL WARN socketProgressOpt: Call to recv from 192.168.1.43 failed : Broken pipe。 [rank2]:[E ProcessGroupNCCL.cpp:577] [Rank 2] Some NCCL operations have failed or timed out. Due to the asynchronous...

[Question]: The eos_token of the Qwen 2.5 base model is inconsistent between config.json and tokenizer_config.json.

> Hi, all, > > If you are unfamilar how `transformers` handles stopping criteria for text generation (the `model.generate` method), please first check out [its official documentation](https://huggingface.co/docs/transformers/main_classes/text_generation). > > The...