Yuhao Xu
Yuhao Xu
@Aaron2117 300 epochs sounds too many. We only trained for 36000 steps (~42 epochs)
> @Aaron2117 The results shown above are all from the test set. Our team believes that the training hyperparameters of ckpt released by the author are very different from those...
> In VITON-HD dataset, I also faced the problem that the results on training data are good but no good on tested data. Whether the open source model provided by...
@JcandZero 是个好想法和好提议 等忙完这段联系你哈
Sorry, we've only tested our code on linux
It costs less than 20GB memory on my RTX4090 for 4 samples. Could you try it again?
Please check your conda environment, like torch and diffusers versions. Different versions may cause different memory usage.
> torch 2.0.1 diffusers 0.24.0 > > can I ask which version of diffusers you're using? Same. Are you running on linux? I have mentioned the required environments in readme....
Maybe you can try this pip list, some versions differ from those in my repo though: ``` absl-py 2.0.0 accelerate 0.24.1 aiofiles 23.2.1 altair 5.2.0 annotated-types 0.6.0 antlr4-python3-runtime 4.9.3 anyio...
T_T weird. Only costs 20 GB for 6 samples for me...