Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

256x256_diffusion_uncond.pt是怎么训练出来的 #156

Open
123gjj opened this issue Nov 22, 2024 · 6 comments
Open

256x256_diffusion_uncond.pt是怎么训练出来的 #156

123gjj opened this issue Nov 22, 2024 · 6 comments

Comments

@123gjj
Copy link

123gjj commented Nov 22, 2024

256x256_diffusion_uncond.pt是怎么训练出来的?只能使用官方给的权重吗?

@shiheyuan2019
Copy link

256x256_diffusion_uncond.pt是怎么训练出来的?只能使用官方给的权重吗?

Hello, do you solve this? I want to fine-turn with my data. But the weight loaded is wrong.

@123gjj
Copy link
Author

123gjj commented Nov 29, 2024

目前还没有,请问你有思路吗

@shiheyuan2019
Copy link

shiheyuan2019 commented Nov 29, 2024 via email

@shiheyuan2019
Copy link

目前还没有,请问你有思路吗

MODEL_FLAGS="--image_size 256 --attention_resolutions 32,16,8 --num_channels 256 --num_head_channels 64 --num_res_blocks 2 --num_heads 4 --resblock_updown true --learn_sigma True --use_scale_shift_norm true --learn_sigma true --timestep_respacing 250 --use_fp16 false --use_kl false " DIFFUSION_FLAGS="--diffusion_steps 1000 --noise_schedule linear --rescale_learned_sigmas False" TRAIN_FLAGS="--lr 1e-4 --microbatch 4 --dropout 0.0"
export OPENAI_LOGDIR="model_result_Diff"
#export NCCL_P2P_DISABLE=1
#mpiexec -n 3 python scripts/image_train.py --resume_checkpoint model/256x256_diffusion_uncond.pt --data_dir datasets/train $MODEL_FLAGS $DIFFUSION_FLAGS $TRAIN_FLAGS
python scripts/image_train.py --resume_checkpoint model/256x256_diffusion_uncond.pt --data_dir datasets/train $MODEL_FLAGS $DIFFUSION_FLAGS $TRAIN_FLAGS
我使用的脚本,可以把模型使用自己的数据进行训练了

@123gjj
Copy link
Author

123gjj commented Dec 1, 2024

为什么我按照这样训练的权重大小只有300M,但是预训练权重256x256_diffusion_uncond.pt大小却有2.21G呢,请问你遇到这个问题了吗

1 similar comment
@123gjj
Copy link
Author

123gjj commented Dec 1, 2024

为什么我按照这样训练的权重大小只有300M,但是预训练权重256x256_diffusion_uncond.pt大小却有2.21G呢,请问你遇到这个问题了吗

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants