fast_finetune in Pytorch failed with complaining something related to grad_fn

Xilinx / Vitis-AI

Vitis AI is Xilinx’s development stack for AI inference on Xilinx hardware platforms, including both edge devices and Alveo cards.

https://www.xilinx.com/ai

Apache License 2.0

1.45k stars 627 forks source link

fast_finetune in Pytorch failed with complaining something related to grad_fn #1173

Open eerobert opened 1 year ago

eerobert commented 1 year ago

hi,

continue to explore vitis-ai on pytorfch.

i come across with fast_finetuning my model, but it crashed by reporting that //RuntimeError: element 0 of tensors does not require grad and does not have a grad_fn//

the model can go thru normal torch_quantizer

any comments on the above issue?

kewang-xlnx commented 1 year ago

Hi @eerobert, which version of Vitis-AI was used? Could you please provide more model information and error message?

Ruiii1 commented 3 months ago

@kewang-xlnx hello.sir, i have encounted this problem,can you help me??