wen

Results 21 comments of wen

it doesnt work well on Asian faces

How do you fix it？

单机多卡全量微调时，学习率和梯度范数为0，并且爆显存

请问全量微调跟训练pretrain有区别吗

stable diffusion machine memory leak

@patrickvonplaten I have no idea about this problem.Can you give me some suggestions？And do you have a plan to solve this problem？

[BUG] Permormance drop while training with MoE

how to get Model TFLOPS/GPU?

how to pretrain llama2?

ok

DataChunkRecipe is not working when used in litgpt's TinyLlama pretraining example

ok

DataChunkRecipe is not working when used in litgpt's TinyLlama pretraining example

I will to make a PR to LitGPT

Exception ignored in: <function AtomicFunction.__del__ at 0x71926a728940>

I also encountered the same problem, is there any progress?

New error: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!

+1

upscaler a video, but the result is bad

@ckkelvinchan