Skip to content

JetStream is a throughput and memory optimized engine for LLM inference on XLA devices, starting with TPUs (and GPUs in future -- PRs welcome).

active 2024-03-132024-09-28 (UTC)

Partial coverage16,698 / 21,169 hourly files (79%) · 2 absent upstream · 4,470 failed, retryable2024-03-132026-08-12 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
914
Pushes
247
Pull requests
118
Issues
9
Stars
102
Forks
9

Activity over time

Daily event counts in the loaded window

Line chart, 200 days from 2024-03-13 to 2024-09-28. Pushes: 247 total, peak 21 in a day. Pull requests: 118 total, peak 6 in a day. Issues: 9 total, peak 2 in a day. Comments: 144 total, peak 14 in a day. Stars: 102 total, peak 7 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Top contributors

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

ContributorContributionsPushesPRsComments
JoeZijunZhou2621153945
FanhaiLu1145202635
morgandu9575152
vivianrwu270313
jwyang-google231750
yeandy234112
qihqi23556
vipannalla23106
Bslabe12319076
bhavya0115481
zhihaoshan-google14560
liurupeng11007
gangji8002
patemotter5003
allenwang284002
rwitten3001
dependabot[bot]2020
lsy3231010
hosseinsarshar1000
ImKeTT1001

Recent activity

Latest issues, pull requests and releases

  • Pull request#138zhihaoshan-google2024-09-26 21:00
  • Pull request#138vivianrwu2024-09-26 20:48
  • Issue#137hosseinsarshar2024-09-11 21:07
    Understanding the intuition behind `request-rate`
  • Pull request#136zhihaoshan-google2024-09-05 23:04
  • Pull request#133qihqi2024-08-27 22:31
  • Issue comment#133JoeZijunZhou2024-08-27 21:42
    Add an optional parameter for sampling in prefill / sample.
  • Pull request#132yeandy2024-08-23 23:16
  • Issue comment#120JoeZijunZhou2024-08-16 23:14
    when to support gpu?
  • Issue comment#131JoeZijunZhou2024-08-16 23:12
    Clean up Model Conversion Script
  • Issue#131yeandy2024-08-16 21:21
    Clean up Model Conversion Script
  • Pull request#126JoeZijunZhou2024-08-14 22:53
  • Pull request#130JoeZijunZhou2024-08-12 23:55
  • Pull request#130JoeZijunZhou2024-08-12 23:51
  • Pull request#129JoeZijunZhou2024-08-09 20:34
  • Pull request#127Bslabe1232024-08-07 22:22
  • Pull request#121JoeZijunZhou2024-08-07 20:50
  • Issue comment#126vivianrwu2024-08-07 20:13
    Manual model warmup to resolve AOT model warmup performance degradation
  • Pull request#126vivianrwu2024-08-07 04:17
  • Pull request#125Bslabe1232024-08-06 22:01
  • Pull request#123JoeZijunZhou2024-08-05 23:49
  • Issue comment#122JoeZijunZhou2024-08-05 23:07
    Performance optimized interleaved mode JetStream server
  • Pull request#124Bslabe1232024-07-31 22:07
  • Pull request#122JoeZijunZhou2024-07-26 10:53
  • Pull request#121Bslabe1232024-07-24 23:55
  • Pull request#119FanhaiLu12024-07-23 01:06

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 102 stars here means stars gained during the window, not the repo's star count.