Skip to content

[ICML 2025 Spotlight] ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inference

active 2025-04-252026-06-08 (UTC)

Complete coverage26,681 / 26,681 hourly files (100%) · 2 absent upstream2023-08-152026-08-30 (UTC)
Events
117
Pushes
3
Pull requests
0
Issues
15
Stars
77
Forks
3

Activity over time

Daily event counts in the loaded window

Line chart, 410 days from 2025-04-25 to 2026-06-08. Pushes: 3 total, peak 2 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 15 total, peak 2 in a day. Comments: 19 total, peak 3 in a day. Stars: 77 total, peak 4 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.

Top contributors

Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity

ContributorContributionsPushesPRsComments
preminstrel133010
YEOMJINSEOP10002
vladislavkruglikov5004
SourishW2001
StepyHan1000
GWS04281000
souvikku1000
yuzhenmao1001
KevinZeng081001
designeryuan1000
m-alqblawi1000

Recent activity

Latest issues, pull requests and releases

  • Issue#22souvikku2026-05-21 02:06
    [vLLM support?]
  • Issue#21YEOMJINSEOP2025-12-18 03:57
    [QUESTION] Why Qwen 2 only supports NIAH, not includng RULER?
  • Issue#21YEOMJINSEOP2025-12-18 01:18
    [QUESTION] Why Qwen 2 only supports NIAH, not includng RULER?
  • Issue#18YEOMJINSEOP2025-09-27 03:38
    [QUESTION] Reproduing Accuracy Evaluation Experiment
  • Issue comment#18YEOMJINSEOP2025-09-27 03:38
    [QUESTION] Reproduing Accuracy Evaluation Experiment
  • Issue comment#18preminstrel2025-09-26 18:26
    [QUESTION] Reproduing Accuracy Evaluation Experiment
  • Issue comment#18preminstrel2025-09-26 06:17
    [QUESTION] Reproduing Accuracy Evaluation Experiment
  • Issue comment#18preminstrel2025-09-26 02:22
    [QUESTION] Reproduing Accuracy Evaluation Experiment
  • Issue#20GWS04282025-09-25 14:49
    [QUESTION] Issue with e2e.py on A100 GPU
  • Issue#15YEOMJINSEOP2025-09-24 05:49
    [QUESTION] Were Prefill_Local and Outlier_Chunk Budget considered in Quest too?
  • Issue#19YEOMJINSEOP2025-09-24 00:37
    [QUESTION] Theoretical Equivalent Bandwidth for ShadowKV
  • Issue#18YEOMJINSEOP2025-09-23 07:36
    [QUESTION] Reproduing Accuracy Evaluation Experiment
  • Issue comment#5SourishW2025-09-16 22:35
    [QUESTION] Multi-turn Conversation Capability
  • Issue comment#13vladislavkruglikov2025-09-16 08:20
    ShadowKVCache and ShadowKVCache_CPU produce different logits
  • Issue#17SourishW2025-09-15 15:48
    [QUESTION] Cache Policy Code
  • Issue comment#14preminstrel2025-09-02 06:33
    [QUESTION] Reproduction of RULER Experiments
  • Issue#15YEOMJINSEOP2025-09-02 00:48
    [QUESTION] Dose Prefill_Local and Outlier_Chunk Budget is considered in Quest too?
  • Issue comment#13vladislavkruglikov2025-09-01 14:12
    Prefill SVD latency does not match paper ablation
  • Issue comment#13vladislavkruglikov2025-09-01 13:18
    Prefill SVD latency does not match paper ablation
  • Issue comment#13preminstrel2025-08-29 18:04
    Prefill SVD latency does not match paper ablation
  • Issue comment#13vladislavkruglikov2025-08-29 12:11
    Prefill SVD latency does not match paper ablation
  • Issue comment#13preminstrel2025-08-28 17:56
    Prefill SVD latency does not match paper ablation
  • Issue#13vladislavkruglikov2025-08-28 14:55
    Prefill SVD latency does not match paper ablation
  • Issue comment#12YEOMJINSEOP2025-08-19 03:51
    [QUESTION] About RULER Benchmark Scoring
  • Issue#12YEOMJINSEOP2025-08-19 02:36
    [QUESTION] About RULER Benchmark Scoring

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 77 stars here means stars gained during the window, not the repo's star count.