Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 27, 2025

• Add 'model_name' label to metrics and add MetricsServerConfig (#209)

  Show description for f4d7753
  Bslabe123

  authored
  f4d7753

  Copy full SHA for f4d7753

Commits on Jan 13, 2025

• Update README.md (#207)

  Show description for 1874be7
  

  qihqi

  and

  lsy323

  authored
  1874be7

  Copy full SHA for 1874be7

Commits on Jan 2, 2025

• Raise error if weights are not loaded (#206)

  qihqi

  authored
  ec66526

  Copy full SHA for ec66526

Commits on Dec 16, 2024

• Llama 3.1 RoPE scaling (#205)

  Show description for bb174b6
  tengomucho

  authored
  bb174b6

  Copy full SHA for bb174b6

Commits on Dec 13, 2024

• Update jax version to 0.4.37 (#204)

  Show description for 08e4977
  qihqi

  authored
  08e4977

  Copy full SHA for 08e4977

• fix: correct error message (#193)

  Show description for 3e46b8e
  tengomucho

  authored
  3e46b8e

  Copy full SHA for 3e46b8e

• Add Dockerfile and entrypoint script for the jetstream-pytorch-server (#203)

  Show description for 26908e9
  vivianrwu

  authored
  26908e9

  Copy full SHA for 26908e9

Commits on Dec 9, 2024

• update branch tag in tutorial (#202)

  sixiang-google

  authored
  a0449f7

  Copy full SHA for a0449f7

• Add `prometheus_port` flag to CLI (#201)

  Show description for d64d941
  Bslabe123

  authored
  d64d941

  Copy full SHA for d64d941

Commits on Nov 7, 2024

• Add jax compilation cache config (#198)

  vivianrwu

  authored
  a12698d

  Copy full SHA for a12698d

Commits on Nov 6, 2024

• Add model warmup flag into cli (#197)

  Show description for fe22a9f
  vivianrwu

  authored
  fe22a9f

  Copy full SHA for fe22a9f

Commits on Oct 23, 2024

• Fix: correct quantization name filtering (#196)

  Show description for 02927c9
  tengomucho

  authored
  02927c9

  Copy full SHA for 02927c9

Commits on Oct 17, 2024

• feat: add quantize exclude layer flag (#194)

  Show description for 664c124
  tengomucho

  authored
  664c124

  Copy full SHA for 664c124

Commits on Oct 15, 2024

• add local tokenizer option for automated testing without hf token (#192)

  sixiang-google

  authored
  e8f5f00

  Copy full SHA for e8f5f00

Commits on Oct 14, 2024

• Add an option to not quantize embedding layer when doing quantization.

  Show description for 5704b30
  qihqi

  committed
  5704b30

  Copy full SHA for 5704b30

Commits on Oct 9, 2024

• Delete convert_checkpoints and helper classmethods. (#190)

  qihqi

  authored
  3afdc43

  Copy full SHA for 3afdc43

Commits on Oct 8, 2024

• Make jpt the default cli - remove other entry point scripts (#188)

  Show description for f32f9c6
  qihqi

  authored
  f32f9c6

  Copy full SHA for f32f9c6

Commits on Oct 7, 2024

• Fix ray recompilation and accuracy (#189)

  sixiang-google

  authored
  d84ae15

  Copy full SHA for d84ae15

Commits on Oct 2, 2024

• Add model warmup and jax compilation cache flags (#187)

  Show description for f2e5181
  vivianrwu

  authored
  f2e5181

  Copy full SHA for f2e5181

Commits on Sep 28, 2024

• Fix too many positional arguments lint error (#186)

  Show description for 5d6ad62
  FanhaiLu1

  authored
  5d6ad62

  Copy full SHA for 5d6ad62

Commits on Sep 24, 2024

• Switch to NP from Jax to improve attention manager performance (#184)

  Show description for f1fa78c
  FanhaiLu1

  authored
  f1fa78c

  Copy full SHA for f1fa78c

Commits on Sep 13, 2024

• Add offline perf ci (#181)

  Show description for 5b8823e
  qihqi

  authored
  5b8823e

  Copy full SHA for 5b8823e

Commits on Sep 10, 2024

• Support End To End PagedAttention in JetStream (#180)

  Show description for 33348d2
  FanhaiLu1

  authored
  33348d2

  Copy full SHA for 33348d2

• Update README for new CLI (#178)

  Show description for ec4ac8f
  qihqi

  authored
  ec4ac8f

  Copy full SHA for ec4ac8f

Commits on Aug 30, 2024

• Update Jetstream, add optional sampler args. (#177)

  Show description for 321f5aa
  qihqi

  authored
  321f5aa

  Copy full SHA for 321f5aa

• Add gemma support in better cli (#176)

  Show description for 7cbd9ec
  qihqi

  authored
  7cbd9ec

  Copy full SHA for 7cbd9ec

Commits on Aug 27, 2024

• Use kwargs to simplify the call sites a bit (#175)

  Show description for df92015
  

  yixinshi

  and

  Yixin Shi

  authored
  df92015

  Copy full SHA for df92015

• Add mixtral support to new CLI (#174)

  Show description for 9c555a8
  qihqi

  authored
  9c555a8

  Copy full SHA for 9c555a8

Commits on Aug 20, 2024

• Fix the performance regression with ragged attention on for llama2 7b. (#172)

  Show description for 7307541
  wang2yn84

  authored
  7307541

  Copy full SHA for 7307541

• fix ray engine crashes on multihost (#170)

  Show description for a69ca92
  sixiang-google

  authored
  a69ca92

  Copy full SHA for a69ca92

• Replace repeat kv with proper GQA handling. (#171)

  Show description for 7092cc5
  wang2yn84

  authored
  7092cc5

  Copy full SHA for 7092cc5

Commits on Aug 6, 2024

• Add a script to measure speed of basic ops (#168)

  Show description for 1e08833
  qihqi

  authored
  1e08833

  Copy full SHA for 1e08833

• Add page attention manager and kvcache manager (#167)

  Show description for eb360ee
  FanhaiLu1

  authored
  eb360ee

  Copy full SHA for eb360ee

• Optimize cache update. (#151)

  Show description for ee040a4
  

  3 peopleauthored
  ee040a4

  Copy full SHA for ee040a4

Commits on Aug 2, 2024

• Fixed exhausted bug between head and workers (#163)

  Show description for dcc7b27
  FanhaiLu1

  authored
  dcc7b27

  Copy full SHA for dcc7b27

Pagination

Previous

Next