Generating Images from Text (Training - Create Embedding)

[deicool]

Hello,

I am trying to run “Stable Diffusion WebUI” on Windows 10 (Legion GPU).

(https://github.com/AUTOMATIC1111/stable-diffusion-webui)

While I try to train (Create Embedding, Provided a Custom Image Directory), I get the following error, any ideas?:

Calculating sha256 for D:\Ganu\AIImage\huggingface\test\webui\embeddings\Test-31stDec20204-3.pt: c900b2fed73e52a66f6fa2a52c2456ac66a8cd252546b5287bf07d9bfc2b9107
Training at rate of 0.005 until step 100000
Preparing dataset…
100%|¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦¦| 1/1 [00:15<00:00, 15.11s/it]
D:\Ganu\AIImage\huggingface\test\webui\modules\textual_inversion\textual_inversion.py:493: FutureWarning: torch.cuda.amp.GradScaler(args...) is deprecated. Please use torch.amp.GradScaler('cuda', args...) instead.
scaler = torch.cuda.amp.GradScaler()
0%| | 0/100000 [00:00<?, ?it/s]*** Error training embedding
Traceback (most recent call last):
File “D:\Ganu\AIImage\huggingface\test\webui\modules\textual_inversion\textual_inversion.py”, line 555, in train_embedding
scaler.scale(loss).backward()
File “D:\Ganu\AIImage\huggingface\test\system\python\lib\site-packages\torch_tensor.py”, line 581, in backward
torch.autograd.backward(
File “D:\Ganu\AIImage\huggingface\test\system\python\lib\site-packages\torch\autograd_init_.py”, line 347, in backward
_engine_run_backward(
File “D:\Ganu\AIImage\huggingface\test\system\python\lib\site-packages\torch\autograd\graph.py”, line 825, in _engine_run_backward
return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass
RuntimeError: Found dtype Float but expected Half

[missionfloyd]

The train tab is severely outdated. Use this instead.
https://github.com/bmaltais/kohya_ss

[deicool]

Can I please use "Kohya_SS" if I have only about 100 custom images?