Open glynpu opened 9 months ago
It is weird because we experience no problem with inference code. I have tried to clone new repo and reinstall all of the required libs and checkpoint models.
root@6b3b847909fd:~/LFM# CUDA_VISIBLE_DEVICES=7 bash bash_scripts/run_test_cls.sh test_args/imnet_dit.txt
Argument file: test_args/imnet_dit.txt
MODEL_TYPE=DiT-B/2
EPOCH_ID=875
DATASET=imagenet_256
EXP=imnet_f8_ditb2
METHOD=dopri5
STEPS=0
CFG=1.5
starting in debug mode
Finish loading model
Inference
Samples are save at './samples_imagenet_256_dopri5_1e-05_1e-05_cfg1.5.jpg
Here is our generated samples with no changes in test args and codes:
Hi, Thanks for sharing this project.
Is my generated samples_imagenet_256_dopri5_1e-05_1e-05_cfg1.5.jpg correct, attached below?
I am trying to generate some images with the released model_875.pth, i.e. the one in following screenshot.
The command I am using is:
logs to terminal is
While the generate 'samples_imagenet_256_dopri5_1e-05_1e-05_cfg1.5.jpg' is mentioned above:
Models and their file structures are:
stabilityai/sd-vae-ft-mse models are downloaded from https://hf-mirror.com/stabilityai/sd-vae-ft-mse/tree/main and their md5sum values are:
md5sum value of model_875.pth is
The code I am using is the master version with no mocifications, i.e. commit "3e1c3fa142e3b13232dce791b6cabeb83402c348"
Is my generated samples_imagenet_256_dopri5_1e-05_1e-05_cfg1.5.jpg correct? Any suggestions to get the expected results?