Google Colab Pro - xfromers needs to be reinstalled

[SDPDVADER]

WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for:
PyTorch 2.1.0+cu121 with CUDA 1202 (you have 2.2.1+cu121)
Python 3.10.12 (you have 3.10.12)
Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers)
Memory-efficient attention, SwiGLU, sparse and more won't be available.
Set XFORMERS_MORE_DETAILS=1 for more details

[HaikenEdge]

For reference, have this on Google Colab Pro as well.

[FackJox]

Also getting this issue, have tried some old fixes/workarounds resulting in the same error

[KyotoLyon]

same error here

[Maximb1985]

I have the same error:(

[Maximb1985]

Found a temporary solution in old records. Pardon my English. You need to insert this text into
"Connect Google Drive":

!pip install lmdb
!pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download.pytorch.org/whl/cu118

[HaikenEdge]

Found a temporary solution in old records. Pardon my English. You need to insert this text into "Connect Google Drive":

!pip install lmdb !pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download.pytorch.org/whl/cu118

Looks like a working hotfix.

[KyotoLyon]

Found a temporary solution in old records. Pardon my English. You need to insert this text into "Connect Google Drive":

!pip install lmdb !pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download.pytorch.org/whl/cu118

I used the code and it seems to be working

[Ptentacle1]

This didn't work for me. Everything loads fine, but as soon as the webui is launched and I try to generate anything, the gradio connection just gets killed and I get the "connection errored out." Then if I try to relaunch SD I get a miles long list of error messages.

[DjNastyMagic]

NotImplementedError: No operator found for memory_efficient_attention_forward with inputs: query : shape=(2, 6144, 8, 40) (torch.float32) key : shape=(2, 6144, 8, 40) (torch.float32) value : shape=(2, 6144, 8, 40) (torch.float32) attn_bias : <class 'NoneType'> p : 0.0 decoderF is not supported because: xFormers wasn't build with CUDA support attn_bias type is <class 'NoneType'> operator wasn't built - see python -m xformers.info for more info flshattF@0.0.0 is not supported because: xFormers wasn't build with CUDA support requires device with capability > (8, 0) but your GPU has capability (7, 5) (too old) dtype=torch.float32 (supported: {torch.bfloat16, torch.float16}) operator wasn't built - see python -m xformers.info for more info tritonflashattF is not supported because: xFormers wasn't build with CUDA support requires device with capability > (8, 0) but your GPU has capability (7, 5) (too old) dtype=torch.float32 (supported: {torch.bfloat16, torch.float16}) operator wasn't built - see python -m xformers.info for more info triton is not available requires GPU with sm80 minimum compute capacity, e.g., A100/H100/L4 Only work on pre-MLIR triton for now cutlassF is not supported because: xFormers wasn't build with CUDA support operator wasn't built - see python -m xformers.info for more info smallkF is not supported because: max(query.shape[-1] != value.shape[-1]) > 32 xFormers wasn't build with CUDA support operator wasn't built - see python -m xformers.info for more info unsupported embed per head: 40
Time taken: 1.1 sec.

[osm14wave]

Found a temporary solution in old records. Pardon my English. You need to insert this text into "Connect Google Drive":

!pip install lmdb !pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download.pytorch.org/whl/cu118

Unfortunately I got an error. Maybe we need to update the text? 118→121

[momoci0704]

I got the same error.

[A3734]

me too (TT

[HaikenEdge]

This didn't work for me. Everything loads fine, but as soon as the webui is launched and I try to generate anything, the gradio connection just gets killed and I get the "connection errored out." Then if I try to relaunch SD I get a miles long list of error messages.

Don't know what to say; it's working fine for me with the hotfix.

[olowwwlolololo]

Hello. Please share screenshots of your notebook with me. I don't understand what to do with the provided code

[TooAiArt]

Hello. Please share screenshots of your notebook with me. I don't understand what to do with the provided code

In the connect google drive cell I have put under "--#@markdown ---":

!pip install lmdb
!pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download. pytorch.org/whl/cu118

and at the moment it works for me

[olowwwlolololo]

Hello. Please share screenshots of your notebook with me. I don't understand what to do with the provided code

In the connect google drive cell I have put under "--#@markdown ---":

!pip install lmdb !pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download. pytorch.org/whl/cu118

and at the moment it works for me

I still need a screenshot. 🗿

[A3734]

I think this method is fine for now. Please insert the code before "Connect Google Drive" like in the screenshot.

!pip install lmdb
!pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download.pytorch.org/whl/cu118

[momoci0704]

Installing collected packages: lmdb
Successfully installed lmdb-1.4.1
ERROR: Invalid requirement: '//download.pytorch.org/whl/cu118'
Hint: It looks like a path. File '//download.pytorch.org/whl/cu118' does not exist.

I get this error message

[A3734]

How about this code?
!pip install -q torch==2.1.2+cu121 torchvision==0.16.2+cu121 torchaudio==2.1.2+cu121 torchtext==0.16.2 torchdata==0.7.1 --extra-index-url https://download.pytorch.org/whl/cu121 -U

[dymension]

Found a temporary solution in old records. Pardon my English. You need to insert this text into "Connect Google Drive":

!pip install lmdb !pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download.pytorch.org/whl/cu118

Works great, thanks Maxim

[olowwwlolololo]

I think this method is fine for now. Please insert the code before "Connect Google Drive" like in the screenshot.

!pip install lmdb !pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download.pytorch.org/whl/cu118

Thank you so much for the screenshot! I didn't understand at all what to do with the provided code, but your screenshot helped me a lot.

[dymension]

[Azura-13]

Same

[aifantasyglam]

Found a temporary solution in old records. Pardon my English. You need to insert this text into "Connect Google Drive":

!pip install lmdb !pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download.pytorch.org/whl/cu118

It works, thanks a lot

[osm14wave]

Works great, thanks Maxim

How about this code? !pip install -q torch==2.1.2+cu121 torchvision==0.16.2+cu121 torchaudio==2.1.2+cu121 torchtext==0.16.2 torchdata==0.7.1 --extra-index-url https://download.pytorch.org/whl/cu121 -U

It works, thanks a lot!

[momoci0704]

Putting the text in front of or in the "Connect Google Drive" does not change the error.
Thanks a lot for your help.

[aifantasyglam]

Putting the text in front of or in the "Connect Google Drive" does not change the error. Thanks a lot for your help.

put the code befor the "start stable diffusion"

[momoci0704]

The image can now be generated.
Thank you all!

[TheLastBen]

fixed, update to the latest notebook and you won't need to install any additional package

[marsellerai]

Use rundiffusion.com. I am happy with that.

[LIQUIDMIND111]

!pip install lmdb
!pip install torch==2.1.0+cu118 torchvision==0.16.0+cu118 torchaudio==2.1.0 torchtext==0.16.0+cpu torchdata==0.7.0 --index-url https://download. pytorch.org/whl/cu118

how do i update to the latest note book? i thought if you click on the collab section it will take you to the latest version of dreambooth? or where do i have to click then?

[TheLastBen]

Simply click on the thumbnail on the main page of the repo

[aifantasyglam]

Hello everyone, there is a new update that is required, what part of the code should I modify...?

Message:
WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for:
PyTorch 2.2.1+cu121 with CUDA 1201 (you have 2.1.0+cu118)
Python 3.10.13 (you have 3.10.12)
Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers)
Memory-efficient attention, SwiGLU, sparse and more won't be available.
Set XFORMERS_MORE_DETAILS=1 for more details

[aifantasyglam]

Hello everyone, there is a new update that is required, what part of the code should I modify...?

Message: WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for: PyTorch 2.2.1+cu121 with CUDA 1201 (you have 2.1.0+cu118) Python 3.10.13 (you have 3.10.12) Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers) Memory-efficient attention, SwiGLU, sparse and more won't be available. Set XFORMERS_MORE_DETAILS=1 for more details

Hello again everyone, according the documentation, using only this line is enough to solve the issue:
!pip3 install -U xformers --index-url https://download.pytorch.org/whl/cu121

and is true, I tried and it worked

[HaikenEdge]

Hello everyone, there is a new update that is required, what part of the code should I modify...?
Message: WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for: PyTorch 2.2.1+cu121 with CUDA 1201 (you have 2.1.0+cu118) Python 3.10.13 (you have 3.10.12) Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers) Memory-efficient attention, SwiGLU, sparse and more won't be available. Set XFORMERS_MORE_DETAILS=1 for more details

Hello again everyone, according the documentation, using only this line is enough to solve the issue: !pip3 install -U xformers --index-url https://download.pytorch.org/whl/cu121

and is true, I tried and it worked

Fixed for now; I imagine @TheLastBen will make an update to the notebook to resolve the issue in due time.

[aifantasyglam]

Hello everyone, there is a new update that is required, what part of the code should I modify...?
Message: WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for: PyTorch 2.2.1+cu121 with CUDA 1201 (you have 2.1.0+cu118) Python 3.10.13 (you have 3.10.12) Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers) Memory-efficient attention, SwiGLU, sparse and more won't be available. Set XFORMERS_MORE_DETAILS=1 for more details

Hello again everyone, according the documentation, using only this line is enough to solve the issue: !pip3 install -U xformers --index-url https://download.pytorch.org/whl/cu121
and is true, I tried and it worked

Fixed for now; I imagine @TheLastBen will make an update to the notebook to resolve the issue in due time.

I guess so

[TheLastBen]

fixed

[aifantasyglam]

fixed

Thanks a lot