como arrumar esse erro na hora de usar trainamento lora, por favor

[yoshyteru]

Traceback (most recent call last):
File "gradio\routes.py", line 394, in run_predict
File "gradio\blocks.py", line 1075, in process_api
File "gradio\blocks.py", line 884, in call_function
File "anyio\to_thread.py", line 31, in run_sync
File "anyio_backends_asyncio.py", line 937, in run_sync_in_worker_thread
File "anyio_backends_asyncio.py", line 867, in run
File "apps\stable_diffusion\scripts\train_lora_word.py", line 251, in lora_train
File "torch\nn\modules\module.py", line 1126, in to
device, dtype, non_blocking, convert_to_format = torch._C._nn._parse_to(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: Invalid device string: 'AMD Radeon RX 6600 => vulkan://0'

[cstueckrath]

backend implementation for training was not released upstream.
You can help yourself if you run it in Linux (WSL should work, too) on top of master with this guide:
(A implementação backend para formação não foi libertada a montante.
Pode ajudar-se a si próprio se a executar em Linux (WSL também deve funcionar) em cima do mestre com este guia:)
https://github.com/nod-ai/SHARK/tree/main/shark/examples/shark_training/stable_diffusion

[yoshyteru]

Hello, good evening, my system is windowns 11, what would be the correct way to install x formers, torch 2.0, I see a lot of limitation in relation to amd and windows for AI ART, this is very discouraging :( Thanks in advance Enviado do Email<https://go.microsoft.com/fwlink/?LinkId=550986> para Windows De: ***@***.***> Enviado:sexta-feira, 24 de março de 2023 05:57 Para: ***@***.***> ***@***.***>; ***@***.***> Assunto: Re: [nod-ai/SHARK] como arrumar esse erro na hora de usar trainamento lora, por favor (Discussion #1240) backend implementation for training was not released upstream. You can help yourself if you run it in Linux (WSL should work, too) on top of master with this guide: https://github.com/nod-ai/SHARK/tree/main/shark/examples/shark_training/stable_diffusion — Reply to this email directly, view it on GitHub<#1240 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AIAB526HIC253MAWCIRI2KDW5VOWFANCNFSM6AAAAAAWF3HVU4>. You are receiving this because you authored the thread.Message ID: ***@***.***>