Yang Liu

@protossw512 Could you share the improvement in your github?

while script?

Thanks @pikapikazhou , could you please create a PR to fix this? @aditya-droid Sorry, I've switched to PyTorch 1 year ago and don't have a keras/tf environement.

Converting the SMPL pose to BVH was easy, and I got that done yesterday.

+1

Same here, sample code with LLaMA-30B

光是512维向量是不足以生成新图片的，需要微调整个模型。@HandsLing

@GunaSekhar-T finetune the model.

accelerate nailed it: 

You are rich. :) CuDNNLSTM is used to accelerate training, please help yourself to make it work in your environment. For example: 1.Open a browser, 2.Go to your favorite search...