xformers

xformers copied to clipboard

# ❓ Questions and Help I am a CUDA programmer and want some practice in writing efficient transformer related ops for NVIDIA GPUs. Is there anywhere I could find a...

alexarmbr

How to support key_padding_mask

1

# ❓ Questions and Help Hi, I am trying to replace existing `torch.nn.functional.scaled_dot_product_attention` with `xformers.ops.memory_efficient_attention`. I want to use `xformers.ops.fmha.flash` and `xformers.ops.fmha.triton` as ops, but I am struggling with how...

YJYJLee

no performance advantage with xFormers

3

Hello! After 9 months I set up Automatic 1111 again from the scratch. But not like in the past, installing xFormers from github source didn't improve the image generation performance....

Marcophono2

# ❓ Questions and Help Given the optimized performance from BlockDiagonalMask on variable sequence length inputs, it is a great helper to speed up the training process. However, I cannot...

rosrad

import xformers.ops error

1

# ❓ Questions and Help pip install from source import xformers success but import xformers.ops error 

safehumeng

Can not execute the cifar VIT example with new pl - patch_embedding invoked multiple times

2

# 🐛 Bug ## Command I am trying to execute the VIT example [1] from the examples folder with new pytorch lightning and I get an error as below I...

jakubMitura14

FileNotFoundError while trying to build xformers from source

3

# ❓ Questions and Help I am trying to build and install xformers from source, since i have Cuda 12.3 and pytorch 2.2.0 nightly, I ran the command mentioned in...

T0biasCZe

Is this a mistake in requirement of llama_inference?

2

https://github.com/facebookresearch/xformers/blob/main/examples/llama_inference/requirements.txt torch>=2.2.0 this should be a mistake we should use FlashAttention >=2.2.0

pengjiao123

Which tag of xformer corresponds to pytorch1.9.0+cu111?

2

# ❓ Questions and Help Which tag of xformer corresponds to pytorch1.9.0+cu111?

iuhiyuh

# 🐛 Bug I was trying to run [mistral-7b ](https://github.com/mistralai/mistral-src)on Jeston ORIN and built triton (openAI) and xformers from source. However, when trying to run mistral-7b, I got following errors:...

cj401