Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 4, 2025

• CacheCraft: A Relevant Work on Chunk-Aware KV Cache Reuse for RAG (#126)

  Show description for 0faf3bf
  skejriwal44

  authored
  0faf3bf

  Copy full SHA for 0faf3bf

Commits on Mar 3, 2025

• Update DeepSeek/MLA Topics (#125)

  DefTruth

  authored
  eb7e0d0

  Copy full SHA for eb7e0d0

Commits on Mar 2, 2025

• Add DeepSeek Open Sources modules (#124)

  DefTruth

  authored
  9928022

  Copy full SHA for 9928022

Commits on Mar 1, 2025

• update the title of SageAttention2 and add SpargeAttn (#123)

  jt-zhang

  authored
  d32b3dd

  Copy full SHA for d32b3dd

• 🔥[MHA2MLA] Towards Economical Inference: Enabling DeepSeek’s Multi-Head Latent Attention in Any Transformer-based LLMs (#122)

  DefTruth

  authored
  cc97844

  Copy full SHA for cc97844

Commits on Feb 27, 2025

• Add our ICLR2025 work Dynamic-LLaVA (#121)

  Show description for 4cb8763
  Blank-z0

  authored
  4cb8763

  Copy full SHA for 4cb8763

Commits on Feb 24, 2025

• feat: flashMLA (#120)

  shaoyuyoung

  authored
  fe502d8

  Copy full SHA for fe502d8

Commits on Feb 19, 2025

• 🔥[DeepSeek-NSA] Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention (#119)

  DefTruth

  authored
  0525c4d

  Copy full SHA for 0525c4d

Commits on Feb 13, 2025

• Add Multi-head Latent Attention(MLA) topic (#118)

  DefTruth

  authored
  1ddf093

  Copy full SHA for 1ddf093

Commits on Jan 31, 2025

• 🔥🔥[Mooncake] Mooncake: A KVCache-centric Disaggregated Architecture for LLM Serving (#117)

  DefTruth

  authored
  d7914c0

  Copy full SHA for d7914c0

• 🔥🔥[DeServe] DESERVE: TOWARDS AFFORDABLE OFFLINE LLM INFERENCE VIA DECENTRALIZATION(#116)

  DefTruth

  authored
  bb1f117

  Copy full SHA for bb1f117

• 🔥🔥[KVDirect] KVDirect: Distributed Disaggregated LLM Inference (#115)

  DefTruth

  authored
  a523f1d

  Copy full SHA for a523f1d

• 🔥🔥[DistServe] DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving (#114)

  DefTruth

  authored
  301cc21

  Copy full SHA for 301cc21

Commits on Jan 24, 2025

• [feat] add deepseek-r1 (#113)

  Show description for b117b3c
  shaoyuyoung

  authored
  b117b3c

  Copy full SHA for b117b3c

Commits on Jan 23, 2025

• Update README.md

  DefTruth

  authored
  c03a56e

  Copy full SHA for c03a56e

Commits on Jan 16, 2025

• Update README.md

  DefTruth

  authored
  1ecf8ab

  Copy full SHA for 1ecf8ab

• Update README.md

  DefTruth

  authored
  e49ea48

  Copy full SHA for e49ea48

• Update README.md

  DefTruth

  authored
  c07a70d

  Copy full SHA for c07a70d

• Update README.md

  DefTruth

  authored
  295f408

  Copy full SHA for 295f408

• Update README.md

  DefTruth

  authored
  bd24577

  Copy full SHA for bd24577

Commits on Jan 15, 2025

• add MiniMax-01 in Trending LLM/VLM Topics and Long Context Attention (#112)

  Show description for 37adfda
  shaoyuyoung

  authored
  37adfda

  Copy full SHA for 37adfda

Commits on Jan 8, 2025

• Update README.md

  DefTruth

  authored
  7874901

  Copy full SHA for 7874901

Commits on Jan 6, 2025

• 🔥🔥[FFPA] FFPA: Yet another Faster Flash Prefill Attention with O(1) SRAM complexity for headdim > 256, ~1.5x faster than SDPA EA(@DefTruth) (#111)

  Show description for b8b3a43
  DefTruth

  authored
  b8b3a43

  Copy full SHA for b8b3a43

• 🔥🔥[SP: TokenRing] TokenRing: An Efficient Parallelism Framework for Infinite-Context LLMs via Bidirectional Communication (#110)

  DefTruth

  authored
  eb6fb10

  Copy full SHA for eb6fb10

Commits on Jan 3, 2025

• 🔥🔥🔥[DeepSeek-V3] DeepSeek-V3 Technical Report(@deepseek-ai) (#109)

  DefTruth

  authored
  12416b5

  Copy full SHA for 12416b5

Commits on Dec 27, 2024

• 🔥🔥[FLASH-ATTENTION RNG] Reducing the Cost of Dropout in Flash-Attention by Hiding RNG with GEMM

  DefTruth

  authored
  b5c9834

  Copy full SHA for b5c9834

Commits on Dec 22, 2024

• 🔥🔥[HADACORE] HADACORE: TENSOR CORE ACCELERATED HADAMARD TRANSFORM KERNEL (#108)

  DefTruth

  authored
  6ad7b30

  Copy full SHA for 6ad7b30

• 🔥[DynamicKV] DynamicKV: Task-Aware Adaptive KV Cache Compression for Long Context LLMs (#107)

  DefTruth

  authored
  7ab2d18

  Copy full SHA for 7ab2d18

• 🔥🔥[NITRO] NITRO: LLM INFERENCE ON INTEL® LAPTOP NPUS (#106)

  DefTruth

  authored
  34f9ce7

  Copy full SHA for 34f9ce7

• 🔥🔥[TurboAttention] TURBOATTENTION: EFFICIENT ATTENTION APPROXIMATION FOR HIGH THROUGHPUTS LLMS (#105)

  DefTruth

  authored
  5abc2c9

  Copy full SHA for 5abc2c9

Commits on Dec 8, 2024

• 🔥[BatchLLM] BatchLLM: Optimizing Large Batched LLM Inference with Global Prefix Sharing and Throughput-oriented Token Batching (#104)

  Show description for 32fdb84
  DefTruth

  authored
  32fdb84

  Copy full SHA for 32fdb84

• 🔥[ClusterKV] ClusterKV: Manipulating LLM KV Cache in Semantic Space for Recallable Compression (#103)

  Show description for 9bb3f6a
  DefTruth

  authored
  9bb3f6a

  Copy full SHA for 9bb3f6a

Commits on Dec 1, 2024

• 🔥[KV Cache Recomputation] Efficient LLM Inference with I/O-Aware Partial KV Cache Recomputation (#102)

  DefTruth

  authored
  9f548f6

  Copy full SHA for 9f548f6

Commits on Nov 28, 2024

• 🔥[Star-Attention: 11x~ speedup] Star Attention: Efficient LLM Inference over Long Sequences (#101)

  Show description for 7939ea2
  DefTruth

  authored
  7939ea2

  Copy full SHA for 7939ea2

Commits on Nov 25, 2024

• 🔥[SparseInfer] SparseInfer: Training-free Prediction of Activation Sparsity for Fast LLM Inference (#100)

  DefTruth

  authored
  40292d7

  Copy full SHA for 40292d7

Pagination

Previous

Next