Skip to content

andreped/GradientAccumulator

View on GitHub ↗Related repositories →

:dart: Gradient Accumulation for TensorFlow 2

active 2023-08-222025-08-02 (UTC)

Partial coverage20,514 / 26,270 hourly files (78%) · 2 absent upstream · 5,753 failed, retryable2023-08-152026-08-13 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
111
Pushes
33
Pull requests
12
Issues
13
Stars
9
Forks
5

Activity over time

Daily event counts in the loaded window

Line chart, 712 days from 2023-08-22 to 2025-08-02. Pushes: 33 total, peak 25 in a day. Pull requests: 12 total, peak 8 in a day. Issues: 13 total, peak 8 in a day. Comments: 31 total, peak 16 in a day. Stars: 9 total, peak 1 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
andreped76331120
dPys5013
codecov[bot]5005
21-10-42002
roebel1001
karim18711000

Recent activity

Latest issues, pull requests and releases

  • Issue#134karim18712024-06-30 23:08
    model saving issue at modelcheckpoint
  • Issue#133andreped2024-02-29 06:31
    raise ValueError('Optimizer must have a "lr" attribute.')
  • Issue comment#13321-10-42024-02-29 03:32
    raise ValueError('Optimizer must have a "lr" attribute.')
  • Issue comment#13321-10-42024-02-29 03:31
    raise ValueError('Optimizer must have a "lr" attribute.')
  • Issue comment#131andreped2024-02-11 17:51
    feat: add mixed-precision and agc to gradaccum optimizer
  • Issue comment#131dPys2024-02-11 06:31
    feat: add mixed-precision and agc to gradaccum optimizer
  • Issue comment#131andreped2024-02-08 10:53
    feat: add mixed-precision and agc to gradaccum optimizer
  • Issue comment#132roebel2024-02-04 22:29
    Review and potentially simplify the implementation
  • Issue comment#132andreped2024-02-04 14:24
    Review and potentially simplify the implementation
  • Issue comment#113andreped2024-01-22 00:26
    No mixed precision support with GradientAccumulateOptimizer?
  • Issue comment#131andreped2024-01-22 00:10
    feat: add mixed-precision and agc to gradaccum optimizer
  • Issue comment#113dPys2024-01-21 23:43
    No mixed precision support with GradientAccumulateOptimizer?
  • Pull request#131dPys2024-01-21 23:42
  • Issue comment#113andreped2024-01-21 23:27
    No mixed precision support with GradientAccumulateOptimizer?
  • Issue comment#113dPys2024-01-21 23:20
    No mixed precision support with GradientAccumulateOptimizer?
  • Pull request#130andreped2024-01-21 23:10
  • Pull request#130andreped2024-01-21 23:10
  • Issue comment#36andreped2024-01-21 22:47
    Model conversion from mixed precision to full precision fails
  • Issue#36andreped2024-01-21 22:47
    Model conversion from mixed precision to full precision fails
  • Issue comment#113andreped2024-01-21 22:44
    No mixed precision support with GradientAccumulateOptimizer?
  • Issue comment#96andreped2024-01-21 22:26
    Optimizer wrapper performance is dependent on tensorflow version
  • Issue#96andreped2024-01-21 22:26
    Optimizer wrapper performance is dependent on tensorflow version
  • Issue comment#93andreped2024-01-21 22:24
    Optimizer wrapper not compatible with tf==2.6
  • Issue#93andreped2024-01-21 22:24
    Optimizer wrapper not compatible with tf==2.6
  • Pull request#129andreped2024-01-21 22:22

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 9 stars here means stars gained during the window, not the repo's star count.