Skip to content

Issues: pytorch-labs/attention-gym

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

55 Open 39 Closed
55 Open 39 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

Feature Request: Support for Dynamic Bias Tensor in FlexAttention Without Recompilation
#123 opened Feb 18, 2025 by pengzhangzhi
BlockMask Value Error
#122 opened Feb 17, 2025 by Leo-T-Zang
1
Use case against efficient SDPA backend
#121 opened Feb 13, 2025 by TParcollet
2
FlexAttention customizability for softmax
#117 opened Feb 9, 2025 by veritas9872
Sigmoid attention?
#116 opened Feb 9, 2025 by lhallee
Error when using flex attention and F.sdpa together.
#115 opened Feb 8, 2025 by wuyushuwys
8
Does Flex attention API accepts a customized attention mask?
#112 opened Feb 6, 2025 by jiagaoxiang
1
Non-elementwise score_mod
#111 opened Jan 30, 2025 by AmoghDabholkar
1
Dynamic mask block sizes during inference
#109 opened Jan 30, 2025 by windsornguyen
3
Can FlexAttention Optimize Masks for Large Table Constraints?
#106 opened Jan 15, 2025 by RaphaelMouravieff
1
FlexAttention uses much more GPU memory than FlashAttention-2
#101 opened Jan 9, 2025 by ChenlongDeng
Building a composite mask with attention_mask from tokenizers
#98 opened Jan 2, 2025 by lhallee
Illegal memory access on backward when there are unused block masks (nightly build)
#96 opened Dec 28, 2024 by timt51
2
FlexAttention slower than eager in HF transformers
#95 opened Dec 27, 2024 by staghado
5
Doc mask returns negative sparsity
#93 opened Dec 21, 2024 by staghado
1
question about masking
#92 opened Dec 18, 2024 by esason
2
Short vs long sequences performance question Further information is requested
#89 opened Dec 12, 2024 by francoishernandez
1
[Inquiry] Document Masking and Assigning Different Weights
#88 opened Dec 12, 2024 by yeahjack
7
flexattn with qwen2
#81 opened Nov 18, 2024 by NonvolatileMemory
6
Flex attention with dropout
#77 opened Nov 13, 2024 by zbh2047
3
Flex attention - gaps in profiler
#76 opened Nov 11, 2024 by tugot17
7
Rope2d
#75 opened Nov 11, 2024 by bhack
7
How to implement Bidirectional Alibi with padding using flex attention?
#74 opened Nov 7, 2024 by sphmel
4
Is there any chance to call backward function dircetly instead of using pytorch autograd mechanism?
#73 opened Nov 7, 2024 by MayDomine
3
Block Size when Q_LEN and KV_LEN are different
#71 opened Nov 4, 2024 by johng149
ProTip! Follow long discussions with comments:>50.