Skip to content

jaymody/speculative-sampling

View on GitHub ↗Related repositories →

Simple implementation of Speculative Sampling in NumPy for GPT-2.

active 2023-08-152025-12-23 (UTC)

Complete coverage26,684 / 26,684 hourly files (100%) · 2 absent upstream2023-08-152026-08-30 (UTC)
Events
98
Pushes
2
Pull requests
1
Issues
12
Stars
65
Forks
7

Activity over time

Daily event counts in the loaded window

Line chart, 862 days from 2023-08-15 to 2025-12-23. Pushes: 2 total, peak 2 in a day. Pull requests: 1 total, peak 1 in a day. Issues: 12 total, peak 2 in a day. Comments: 8 total, peak 2 in a day. Stars: 65 total, peak 3 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
jaymody11214
sdtblck3002
StudyingLover3001
MeWannaSleep2001
Nero01131000
lawo1231000
al0931000
leekum20181000

Recent activity

Latest issues, pull requests and releases

  • Issue#10al0932025-02-12 15:25
    `min` in `np.random.random() < min(1, q[i][j] / p[i][j]):` can be removed ?
  • Issue comment#9jaymody2024-09-24 19:26
    Hi, a practical question about speculative decoding
  • Issue#9jaymody2024-09-24 19:26
    Hi, a practical question about speculative decoding
  • Issue#9leekum20182024-09-22 05:34
    Hi, a practical question about speculative decoding
  • Issue comment#3MeWannaSleep2023-10-15 02:40
    Question about the code
  • Issue comment#8jaymody2023-09-27 16:25
    About the batch setting
  • Issue#8jaymody2023-09-27 16:25
    About the batch setting
  • Issue#8Nero01132023-09-27 08:45
    About the batch set
  • Issue comment#7StudyingLover2023-09-05 15:05
    No Python 3.9.10 in conda
  • Issue#7StudyingLover2023-09-05 15:05
    No Python 3.9.10 in conda
  • Issue comment#7jaymody2023-09-05 14:34
    No Python 3.9.10 in conda
  • Issue#7StudyingLover2023-09-05 01:44
    No Python 3.9.10 in conda
  • Issue#6lawo1232023-08-29 02:48
    Can this decoding method be directly used in the question answering task? For example chatgpt
  • Issue#5jaymody2023-08-22 14:42
    Adjusting distribution can result in NaNs if q[i] == p[i]
  • Issue comment#5sdtblck2023-08-22 02:19
    Adjusting distribution can result in NaNs if q[i] == p[i]
  • Issue comment#5sdtblck2023-08-20 14:52
    Adjusting distribution can result in NaNs if q[i] == p[i]
  • Issue#5sdtblck2023-08-20 14:49
    Adjusting distribution can result in NaNs if q[i] == p[i]
  • Issue#3jaymody2023-08-17 18:45
    Question about the code
  • Pull request#4jaymody2023-08-17 18:44
  • Issue comment#3jaymody2023-08-17 18:40
    Question about the code
  • Issue#3MeWannaSleep2023-08-15 09:22
    Question about the code

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 65 stars here means stars gained during the window, not the repo's star count.