Discussions

• 1 You must be logged in to vote

  🙏

  Where do I find the .whl file to install it?

  Centurax asked Apr 4, 2023 in Q&A · Answered

  

  

  

  

  14
• 1 You must be logged in to vote

  🙏

  Wrong pytorch

  Creepybits asked Mar 20, 2024 in Q&A · Closed · Unanswered

  

  

  3
• 2 You must be logged in to vote

  💡

  Interested in Developing MPS Support

  theadamsabra started Jan 29, 2024 in Ideas

  

  

  

  

  9
• 1 You must be logged in to vote

  💬

  xformers pypi project capacity is full

  rltgjqmcpgjadyd started Jan 20, 2024 in General · Closed

  

  

  2
• 5 You must be logged in to vote

  🙏

  How to implement a Custom Block Diagonal Mask?

  TamirFriedman-RecoLabs asked Dec 4, 2023 in Q&A · Unanswered

  

  

  1
• 1 You must be logged in to vote

  💡

  Dilated Attention planned or already implemented?

  Coluding started Nov 19, 2023 in Ideas

  

  

  10
• 1 You must be logged in to vote

  💬

  Is there a way to create FLASH attention in xFormerConfig?

  YoelShoshan started Nov 6, 2023 in General

  

  

  3
• 2 You must be logged in to vote

  🙏

  Confused about documentation example

  R-seny asked Oct 27, 2023 in Q&A · Answered

  

  

  2
• 1 You must be logged in to vote

  🙏

  fastchat calling on xformers?

  nbollman asked Oct 18, 2023 in Q&A · Unanswered

  

  

  0
• 1 You must be logged in to vote

  🙏

  When will Flash-Attentionv2 be available for windows?

  rltgjqmcpgjadyd asked Aug 26, 2023 in Q&A · Closed · Answered

  

  

  

  

  3
• 1 You must be logged in to vote

  💡

  Why we use q,k,k_max = (64, 64, 32) / (64, 128,128) / (64, 128, 2e16) for sm80?

  ZhangDY-6483 started Sep 4, 2023 in Ideas

  

  

  4
• 1 You must be logged in to vote

  🙏

  cuda 12.1

  zethfoxster asked Aug 18, 2023 in Q&A · Unanswered

  

  

  1
• 1 You must be logged in to vote

  🙏

  cant deinstall xformers 0.0.16

  Gugu32 asked May 2, 2023 in Q&A · Unanswered

  

  

  1
• 1 You must be logged in to vote

  💬

  flshattF is not supported because: max(query.shape[-1] != value.shape[-1]) > 128

  liuchenbaidu started Apr 14, 2023 in General

  

  

  1
• 1 You must be logged in to vote

  🙌

  RTX 3060 vs RTX 3090 Benchmarks - Tested Torch 1.13, Torch 2, cudNN 8.8.0.1, xFormers, OPT-SDP-Attention, DreamBooth, IT/s, NansException all NaNs Solution, Watt Usage, Dual Cards Performance

  FurkanGozukara started Apr 11, 2023 in Show and tell

  

  

  0
• 1 You must be logged in to vote

  🙏

  Is there a way to make a decoder only model?

  antoni4040 asked Mar 31, 2023 in Q&A · Unanswered

  

  

  0
• 1 You must be logged in to vote

  🙏

  fatal: Failed to recurse into submodule path 'third_party/flash-attention'

  DoDoExt asked Mar 30, 2023 in Q&A · Answered

  

  

  1
• 1 You must be logged in to vote

  🙏

  CUDA role in Memory Efficient Attention

  BillyGun27 asked Mar 13, 2023 in Q&A · Answered

  

  

  1
• 1 You must be logged in to vote

  🙏

  Trying to modify microGPT to use BERT architecture

  phshah95 asked Feb 3, 2023 in Q&A · Unanswered

  

  

  0
• 1 You must be logged in to vote

  💬

  Is it possible to run BlockSparseAttention with cuda 11.0-11.2?

  fedorovgv started Jan 17, 2023 in General

  

  

  1
• 1 You must be logged in to vote

  🙏

  cpp_lib.json not found after new build of xformers

  aifartist asked Jan 13, 2023 in Q&A · Answered

  

  

  1
• 1 You must be logged in to vote

  🙏

  How can I get built wheel after "pip install -v -U git+https://github.com/facebookresearch/xformers.git@main#egg=xformers"

  terrificdm asked Jan 12, 2023 in Q&A · Unanswered

  

  

  2
• 2 You must be logged in to vote

  🙏

  Remove the /models subfolder ?

  blefaudeux asked Nov 1, 2021 in Q&A · Answered

  

  

  2
• 1 You must be logged in to vote

  🙏

  are the inputs always assumed to be tokens?

  kashif asked Nov 29, 2021 in Q&A · Answered

  

  

  2