search

bmaltais / kohya_ss

Apache License 2.0
9.56k stars 1.23k forks source link

AttributeError: 'NoneType' object has no attribute 'dtype' #1910

Closed Sfever closed 5 months ago

Sfever commented 9 months ago

Console Logs: [Dataset 0] loading image sizes. 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 160/160 [00:00<00:00, 13898.49it/s] prepare dataset preparing accelerator loading model for process 0/1 load Diffusers pretrained models: stabilityai/stable-diffusion-xl-refiner-1.0, variant=fp16 Loading pipeline components...: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:03<00:00, 1.31it/s] Traceback (most recent call last): File "D:\Administrator\Documents\kohya_ss\sdxl_train_network.py", line 184, in trainer.train(args) File "D:\Administrator\Documents\kohya_ss\train_network.py", line 229, in train model_version, text_encoder, vae, unet = self.load_target_model(args, weight_dtype, accelerator) File "D:\Administrator\Documents\kohya_ss\sdxl_train_network.py", line 42, in load_target_model ) = sdxl_train_util.load_target_model(args, accelerator, sdxl_model_util.MODEL_VERSION_SDXL_BASE_V1_0, weight_dtype) File "D:\Administrator\Documents\kohya_ss\library\sdxl_train_util.py", line 34, in load_target_model ) = _load_target_model( File "D:\Administrator\Documents\kohya_ss\library\sdxl_train_util.py", line 101, in _load_target_model if text_encoder1.dtype != torch.float32: AttributeError: 'NoneType' object has no attribute 'dtype' Traceback (most recent call last): File "C:\Users\Administrator\AppData\Local\Programs\Python\Python310\lib\runpy.py", line 196, in _run_module_as_main return _run_code(code, main_globals, None, File "C:\Users\Administrator\AppData\Local\Programs\Python\Python310\lib\runpy.py", line 86, in _run_code exec(code, run_globals) File "D:\Administrator\Documents\kohya_ss\venv\Scripts\accelerate.exe__main__.py", line 7, in File "D:\Administrator\Documents\kohya_ss\venv\lib\site-packages\accelerate\commands\accelerate_cli.py", line 47, in main args.func(args) File "D:\Administrator\Documents\kohya_ss\venv\lib\site-packages\accelerate\commands\launch.py", line 1017, in launch_command simple_launcher(args) File "D:\Administrator\Documents\kohya_ss\venv\lib\site-packages\accelerate\commands\launch.py", line 637, in simple_launcher raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd) subprocess.CalledProcessError: Command '['D:\Administrator\Documents\kohya_ss\venv\Scripts\python.exe', './sdxl_train_network.py', '--pretrained_model_name_or_path=stabilityai/stable-diffusion-xl-refiner-1.0', '--train_data_dir=D:/dataset/processes/prepared/img/', '--resolution=2048,2048', '--output_dir=D:/dataset/processes/prepared/model/', '--logging_dir=D:/dataset/processes/prepared/log/', '--network_alpha=32', '--training_comment=3 repeats. More info: https://civitai.com/articles/1771', '--save_model_as=safetensors', '--network_module=networks.lora', '--text_encoder_lr=3e-05', '--unet_lr=3e-05', '--network_dim=32', '--network_weights=D:/Administrator/Documents/stable-diffusion-webui/models/Lora/(v1)reg-any4.5-shenhe-000010-nsfw.safetensors', '--dim_from_weights', '--output_name=last', '--lr_scheduler_num_cycles=50', '--no_half_vae', '--learning_rate=3e-05', '--lr_scheduler=linear', '--train_batch_size=3', '--max_train_steps=4000', '--save_every_n_epochs=1', '--mixed_precision=fp16', '--save_precision=fp16', '--caption_extension=.txt', '--cache_latents', '--cache_latents_to_disk', '--optimizer_type=AdamW', '--max_grad_norm=1', '--max_train_epochs=50', '--max_data_loader_n_workers=0', '--caption_dropout_rate=0.05', '--bucket_reso_steps=64', '--min_snr_gamma=5', '--gradient_checkpointing', '--xformers', '--noise_offset=0.0']' returned non-zero exit status 1.

Not Sure What's happening I'm training Lora

Sfever commented 9 months ago

I've tried again, seems Dreambooth refuses to work as well both using official SDXL-refiner model

piotr-sikora-v commented 9 months ago

I have same error using docker image on runpod.io

muneeb2001 commented 5 months ago

@bmaltais I am still facing this issue while training LORA model: stabilityai/stable-diffusion-xl-refiner-1.0

File "/home/muneeb/Desktop/projects/proptexx/lora-training/kohya_update/kohya_ss/sd-scripts/sdxl_train_network.py", line 185, in <module>
    trainer.train(args)
  File "/home/muneeb/Desktop/projects/proptexx/lora-training/kohya_update/kohya_ss/sd-scripts/train_network.py", line 234, in train
    model_version, text_encoder, vae, unet = self.load_target_model(args, weight_dtype, accelerator)
  File "/home/muneeb/Desktop/projects/proptexx/lora-training/kohya_update/kohya_ss/sd-scripts/sdxl_train_network.py", line 44, in load_target_model
    ) = sdxl_train_util.load_target_model(args, accelerator, sdxl_model_util.MODEL_VERSION_SDXL_BASE_V1_0, weight_dtype)
  File "/home/muneeb/Desktop/projects/proptexx/lora-training/kohya_update/kohya_ss/sd-scripts/library/sdxl_train_util.py", line 40, in load_target_model
    ) = _load_target_model(
  File "/home/muneeb/Desktop/projects/proptexx/lora-training/kohya_update/kohya_ss/sd-scripts/library/sdxl_train_util.py", line 106, in _load_target_model
    if text_encoder1.dtype != torch.float32:
AttributeError: 'NoneType' object has no attribute 'dtype'
lastYoueven commented 5 months ago

I have same error using docker image on runpod.io