费政聪 comments

Results 7 comments of


                                            费政聪

toy_data.json can't find

Actually, toy_data.json is the first 100 items in the train.json, which is used to debug code quickly. You can use train.json instead, off course, we will fix it later :)

Does it support Chinese?

Hi, i have tried to fine-tune LLaMA model for better Chinese text generation. Please refer to: https://github.com/feizc/MLE-LLaMA

Is there a way to fine-tune this model?

Hi, i have tried for instruct tuning on LLaMa, and provides scripts and alignment datasets : ) . You can reference for: https://github.com/feizc/MLE-LLaMA

About music generation with perceiver-ar model

> @feizc how are you approaching the problem of generating starting from a length that is less than the prefix? Actually, I use a fixed length of conditional context, i.e.,...

Hi, @a-r-r-o-w I try to convert the model ckpts with scripts: https://github.com/huggingface/diffusers/blob/mochi/scripts/convert_mochi_to_diffusers.py And create generation with pipeline in mochi branch: https://github.com/huggingface/diffusers/tree/mochi with code: import torch from diffusers import MochiPipeline from...

OSError: Error no file named model_index.json found in directory

Hi, sorry for the missing file upload. We have uploaded the file in [huggingface](https://huggingface.co/Skywork/SkyReels-A2/blob/main/model_index.json).

启动报错:huggingface_hub.errors.HFValidationError: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/maindata/data/shared/public/multimodal/ckpt/wan14B-compose'. Use `repo_type` argument if needed.

hi, the hyper-parameters here need to be set to the path of the downloaded model. https://github.com/SkyworkAI/SkyReels-A2/blob/main/app.py#L35