Skip to content

Official codebase for I-JEPA, the Image-based Joint-Embedding Predictive Architecture. First outlined in the CVPR paper, "Self-supervised learning from images with a joint-embedding predictive architecture."

active 2025-03-162026-06-14 (UTC)

Complete coverage26,568 / 26,568 hourly files (100%) · 2 absent upstream2023-08-152026-08-25 (UTC)
Events
205
Pushes
97
Pull requests
8
Issues
62
Stars
0
Forks
0

Activity over time

Daily event counts in the loaded window

Line chart, 456 days from 2025-03-16 to 2026-06-14. Pushes: 97 total, peak 10 in a day. Pull requests: 8 total, peak 4 in a day. Issues: 62 total, peak 9 in a day. Comments: 35 total, peak 5 in a day. Stars: 0 total, peak 0 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
kutayeroglu20297835

Recent activity

Latest issues, pull requests and releases

  • Issue comment#92kutayeroglu2026-05-03 16:06
    Come up with different masking methods
  • Issue comment#89kutayeroglu2026-04-20 12:31
    Train model with multinoise(ratio=0.5) enc scale [0.95, 1.0]
  • Issue#88kutayeroglu2026-04-11 08:41
    Inspect accuracy discrepancy between .err & .out logs in multi-GPU linear probing
  • Issue comment#84kutayeroglu2026-04-01 17:39
    Remove linear scaling of hyperparameters use same evaluation setting to see whether performance improves
  • Issue#83kutayeroglu2026-03-22 13:35
    Parser for calculating average iteration time
  • Issue#72kutayeroglu2026-03-20 12:04
    Add CLI arg: model_path for linear probing
  • Issue#81kutayeroglu2026-03-16 20:24
    Automate execution of linear probing
  • Issue comment#76kutayeroglu2026-03-06 19:37
    Make linear probing runnable on MN5
  • Issue#75kutayeroglu2026-03-01 12:17
    Get linear probing results for MN5: 37102468 and 37102565
  • Issue#74kutayeroglu2026-03-01 12:08
    Add `color_mask_ratio` to CLI, and passable with yaml
  • Issue#74kutayeroglu2026-03-01 12:08
    Add `color_mask_ratio` to CLI, and passable with yaml
  • Issue#40kutayeroglu2026-02-28 11:56
    Implement proposed masking strategy
  • Issue#72kutayeroglu2026-02-23 12:25
    Add CLI arg: model_path for linear probing
  • Issue#72kutayeroglu2026-02-23 12:25
    Add CLI arg: model_path for linear probing
  • Issue#72kutayeroglu2026-02-23 12:25
    Add CLI arg: model_path for linear probing
  • Issue#72kutayeroglu2026-02-23 12:25
    Add CLI arg: model_path for linear probing
  • Issue#72kutayeroglu2026-02-23 12:25
    Add CLI arg: model_path for linear probing
  • Pull request#71kutayeroglu2026-02-23 11:59
  • Pull request#71kutayeroglu2026-02-23 11:59
  • Pull request#71kutayeroglu2026-02-23 11:59
  • Issue comment#70kutayeroglu2026-02-20 10:02
    Inspect multiblock.py for implementation details
  • Issue comment#47kutayeroglu2026-02-20 08:06
    Why does training logs go under .err slurm outputs?
  • Issue comment#68kutayeroglu2026-02-20 08:03
    Get ImageNet1K on MN5
  • Issue#69kutayeroglu2026-02-18 07:38
    Get executable environemnt on MN5
  • Issue comment#59kutayeroglu2026-02-18 07:38
    Run main training and linear probing on TRUBA

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 0 stars here means stars gained during the window, not the repo's star count.