NVIDIA/kvpress issues and pull requests

#49 - Fix distributed inference

Pull Request - State: closed - Opened by SimJeg 4 days ago

#48 - Fix style errors

Pull Request - State: closed - Opened by maxjeblick 6 days ago

#47 - Add epsilon

Pull Request - State: closed - Opened by SimJeg 6 days ago - 3 comments

#46 - Integrating CriticalKV with Existing Methods

Pull Request - State: closed - Opened by FFY0 6 days ago - 9 comments

#45 - Request for New Press Support: Integrating CriticalKV with Existing Methods

Issue - State: closed - Opened by FFY0 11 days ago - 11 comments
Labels: feature request

#44 - SwiftKV implementation

Issue - State: closed - Opened by msharmavikram 13 days ago - 1 comment
Labels: feature request

#43 - Questions and Suggestions on Prompt Design

Issue - State: closed - Opened by giulio98 24 days ago - 3 comments
Labels: feature request

#42 - Update README

Pull Request - State: closed - Opened by SimJeg 27 days ago

#41 - Update README

Pull Request - State: closed - Opened by SimJeg 27 days ago

#40 - Add chunk press

Pull Request - State: closed - Opened by maxjeblick 28 days ago - 2 comments

#39 - Support transformers 4.48

Pull Request - State: closed - Opened by SimJeg about 1 month ago

#38 - AdaKVPress

Pull Request - State: closed - Opened by SimJeg about 1 month ago - 7 comments

#37 - Add infinitebench benchmark

Pull Request - State: closed - Opened by maxjeblick about 1 month ago

#36 - Fix transformers <4.48

Pull Request - State: closed - Opened by maxjeblick about 1 month ago

#35 - Breaking changes from transformers v4.48

Issue - State: closed - Opened by SimJeg about 1 month ago
Labels: bug

#34 - How can KVPress be used without pipeline abstraction?

Issue - State: closed - Opened by sleepingcat4 about 1 month ago - 4 comments

#33 - Fix pipeline for sdpa

Pull Request - State: closed - Opened by SimJeg 2 months ago

#32 - Add Infllm

Issue - State: closed - Opened by crazyofapple 2 months ago - 2 comments
Labels: feature request, not planned

#31 - Add KeyRerotationPress

Pull Request - State: closed - Opened by maxjeblick 2 months ago - 2 comments

#30 - Fix quantized cache

Pull Request - State: closed - Opened by maxjeblick 2 months ago

#29 - Add ComposedPress and compress method

Pull Request - State: closed - Opened by SimJeg 2 months ago

#29 - Add ComposedPress and compress method

Pull Request - State: closed - Opened by SimJeg 2 months ago

#28 - Add SimLayerKVPress

Pull Request - State: closed - Opened by SimJeg 2 months ago - 7 comments

#27 - QuantoQuantizedCache fails with number of questions > 1

Issue - State: closed - Opened by maxjeblick 2 months ago
Labels: bug

#26 - Bridge to `vllm`

Issue - State: closed - Opened by toilaluan 2 months ago - 1 comment
Labels: feature request, not planned

#26 - Bridge to `vllm`

Issue - State: open - Opened by toilaluan 2 months ago - 1 comment
Labels: feature request, not planned

#25 - Head-Specific KV Cache Compression Feature (Ada-SnapKV, AdaKV)

Pull Request - State: open - Opened by FFY0 2 months ago - 3 comments

#25 - Head-Specific KV Cache Compression Feature (Ada-SnapKV, AdaKV)

Pull Request - State: closed - Opened by FFY0 2 months ago - 9 comments

#24 - Refactor press-scorer

Pull Request - State: closed - Opened by maxjeblick 3 months ago - 2 comments

#24 - Refactor press-scorer

Pull Request - State: closed - Opened by maxjeblick 3 months ago - 2 comments

#23 - Unknown task kv-press-text-generation

Issue - State: closed - Opened by Dominic789654 3 months ago - 1 comment
Labels: bug

#23 - Unknown task kv-press-text-generation

Issue - State: closed - Opened by Dominic789654 3 months ago - 1 comment
Labels: bug

#22 - Add simlayerkv

Pull Request - State: closed - Opened by dame-cell 3 months ago - 5 comments

#22 - Add simlayerkv

Pull Request - State: closed - Opened by dame-cell 3 months ago - 5 comments

#21 - Refactor press implementation

Issue - State: closed - Opened by maxjeblick 3 months ago - 2 comments
Labels: feature request

#21 - Refactor press implementation

Issue - State: closed - Opened by maxjeblick 3 months ago - 2 comments
Labels: feature request

#20 - Add ThinKPress

Pull Request - State: closed - Opened by SimJeg 3 months ago - 6 comments

#19 - Add SIMLAYERKV

Issue - State: closed - Opened by dame-cell 3 months ago - 3 comments
Labels: feature request

#18 - Add ThinK_Press

Pull Request - State: closed - Opened by yuhuixu1993 3 months ago - 8 comments

#17 - Handling Context Lengths Exceeding the Model's Maximum Context Window Size

Issue - State: closed - Opened by giulio98 3 months ago - 9 comments
Labels: feature request

#17 - Handling Context Lengths Exceeding the Model's Maximum Context Window Size

Issue - State: open - Opened by giulio98 3 months ago - 8 comments
Labels: feature request

#16 - Adjust Positional Embedding For Compression

Issue - State: closed - Opened by giulio98 3 months ago - 14 comments
Labels: feature request

#15 - Update pyproject.toml

Pull Request - State: closed - Opened by SimJeg 3 months ago

#15 - Update pyproject.toml

Pull Request - State: closed - Opened by SimJeg 3 months ago

#14 - Update issue templates

Pull Request - State: closed - Opened by SimJeg 3 months ago

#14 - Update issue templates

Pull Request - State: closed - Opened by SimJeg 3 months ago

#13 - Fix cache size calculation for quantized cache

Pull Request - State: closed - Opened by maxjeblick 3 months ago

#12 - Add TOVA press

Pull Request - State: closed - Opened by SimJeg 3 months ago - 5 comments

#12 - Add TOVA press

Pull Request - State: closed - Opened by SimJeg 3 months ago - 5 comments

#11 - Add Infinitebench benchmark

Issue - State: closed - Opened by maxjeblick 3 months ago
Labels: feature request

#11 - Add Infinitebench benchmark

Issue - State: open - Opened by maxjeblick 3 months ago
Labels: feature request

#10 - speed and memory plots

Pull Request - State: closed - Opened by maxjeblick 3 months ago - 9 comments

#9 - Add Activation Beacon

Issue - State: closed - Opened by toilaluan 3 months ago - 4 comments
Labels: feature request

#8 - docs: update zero_scrolls/README.md

Pull Request - State: closed - Opened by eltociear 3 months ago - 1 comment

#7 - Request for Head-Specific KV Cache Compression Feature

Issue - State: closed - Opened by FFY0 3 months ago - 10 comments
Labels: feature request, good first issue

#6 - Colab demo

Pull Request - State: closed - Opened by maxjeblick 3 months ago

#5 - add support for QuantizedCache

Pull Request - State: closed - Opened by SimJeg 3 months ago

#4 - fast-patching library to be compatible with colab and environments where latest transformers, tensorflow is installed.

Pull Request - State: closed - Opened by sleepingcat4 3 months ago - 5 comments

#3 - [BUG] pip install kvpress fails on colab

Issue - State: closed - Opened by maxjeblick 3 months ago - 1 comment

#2 - [NEW PRESS] Add TOVA

Issue - State: closed - Opened by hassidm 3 months ago - 2 comments

#1 - install poetry in workflows

Pull Request - State: closed - Opened by SimJeg 3 months ago

GitHub / NVIDIA/kvpress issues and pull requests