Skip to content

SonyResearch/micro_diffusion

View on GitHub ↗Related repositories →

Official repository for our work on micro-budget training of large-scale diffusion models.

active 2024-07-232026-05-18 (UTC)

Partial coverage15,874 / 19,727 hourly files (80%) · 2 absent upstream · 3,852 failed, retryable2024-05-122026-08-11 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
729
Pushes
3
Pull requests
5
Issues
14
Stars
662
Forks
26

Activity over time

Daily event counts in the loaded window

Line chart, 665 days from 2024-07-23 to 2026-05-18. Pushes: 3 total, peak 2 in a day. Pull requests: 5 total, peak 2 in a day. Issues: 14 total, peak 4 in a day. Comments: 18 total, peak 3 in a day. Stars: 662 total, peak 140 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

Recent activity

Latest issues, pull requests and releases

  • Pull request#6NielsRogge2025-12-14 19:29
  • Pull request#15yuasosnin2025-10-30 17:00
  • Pull request#15yuasosnin2025-10-30 17:00
  • Issue#14SwayStar1232025-05-04 11:57
    Swiglu?
  • Pull request#13nano2562025-03-08 16:04
  • Pull request#13nano2562025-03-08 16:01
  • Issue#12Vishal-S-P2025-03-01 05:23
    CIFAR-Captions
  • Issue comment#6NielsRogge2025-02-16 11:08
    Make sure Hugging Face download stats work, better discoverability
  • Issue#11ascust2025-02-13 02:41
    Evaluation code
  • Issue comment#10LuciferZap2025-02-07 06:03
    Observations on Training Efficiency with Different Masking Strategies
  • Issue comment#10LuciferZap2025-02-07 06:01
    Observations on Training Efficiency with Different Masking Strategies
  • Issue comment#10xiankgx2025-02-04 06:07
    Observations on Training Efficiency with Different Masking Strategies
  • Issue comment#10LuciferZap2025-01-29 03:50
    Observations on Training Efficiency with Different Masking Strategies
  • Issue#10LuciferZap2025-01-29 03:11
    Observations on Training Efficiency with Different Masking Strategies
  • Issue#9yja12025-01-20 03:31
    when use streaming data ,how about read speed when training, just save on local. compare to just use pytorch dataloader
  • Issue comment#6NielsRogge2025-01-18 09:00
    Make sure Hugging Face download stats work, better discoverability
  • Issue#7nitinmukesh2025-01-15 05:55
    12 minute to generate one 512 x 512 image
  • Issue comment#5vsehwag-sony2025-01-12 20:36
    Release the code?
  • Issue#5vsehwag-sony2025-01-12 20:36
    Release the code?
  • Issue#3vsehwag-sony2025-01-12 20:34
    does this method support video generation, such as dit base
  • Issue comment#3vsehwag-sony2025-01-12 20:34
    does this method support video generation, such as dit base
  • Issue#2vsehwag-sony2025-01-12 20:23
    Code release suggestion
  • Issue comment#1vsehwag-sony2025-01-12 20:21
    code release?
  • Issue#1vsehwag-sony2025-01-12 20:21
    code release?
  • Issue#5odusseys2024-12-03 06:36
    Release the code?

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 662 stars here means stars gained during the window, not the repo's star count.