Poor performance on DDTI dataset
ShadiDorosti opened this issue · comments
Hello. Thanks for the great work. I tried to replicate the results on Thyroid DDTI dataset, but I got very poor results. The average Dice score is 0.61 and IoU is 0.47. Could you please share your insight on this issue? and I trained the model for 395k iterations, and tested using the 150k checkpoint, 185k checkpoint, and 395k checkpoint, with "ema" checkpoints. The results I got are almost the same for all three checkpoints.
I am using these parameters to train the model:
--image_size 256 --num_channels 128 --class_cond False --num_res_blocks 2 --num_heads 1 --learn_sigma True --use_scale_shift_norm False --attention_resolutions 16 --diffusion_steps 1000 --noise_schedule linear --rescale_learned_sigmas False --rescale_timesteps False --lr 1e-4 --batch_size 8
Thanks.
@ShadiDorosti 你好我想问下你怎么生成彩色背景图片的？
Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑🤝🧑👫🧑🏿🤝🧑🏻👩🏾🤝👨🏿👬🏿
@ShadiDorosti Hello, I would like to ask how you generate a colored background image?