Hi team,
I followed the environment setup steps provided in the README :
# Create a directory named 'ckpts' where the model will be saved
mkdir ckpts
# Download the model using Huggingface CLI
huggingface-cli download Tencent-Hunyuan/HunyuanDiT-v1.2 --local-dir ./ckpts
Then, I tried to start training with the following command:
PYTHONPATH=./:$PYTHONPATH sh hydit/train.sh --index-file dataset/porcelain/jsons/porcelain_mt.json --multireso --reso-step 64
But I got the following error in the logs:
[rank9]: File "/home/xuechao/0429/HunyuanDiT/hydit/utils/tools.py", line 176, in model_resume
[rank9]: raise FileNotFoundError(
[rank9]: FileNotFoundError: Cannot find model checkpoint from ./ckpts/t2i/model/pytorch_model_distill.pt
It seems that the expected file ./ckpts/t2i/model/pytorch_model_distill.pt is missing. I check the hugging face repo https://huggingface.co/Tencent-Hunyuan/HunyuanDiT-v1.2
there is no pytorch_model_distill.pt file provided. But the code requires this checkpoint to run.