Commits

History for

Commits on Jun 3, 2024

fix typo (#98 )
morgandu
committedJun 3, 2024
Update benchmark config for xlml automation (#96 )
morgandu
committedJun 3, 2024

Commits on May 28, 2024

Ensure server warmup before benchmark (#91 )
JoeZijunZhou
committedMay 28, 2024

Commits on May 17, 2024

Allow tokenizer to customize stop_tokens (#84 )
qihqi
committedMay 17, 2024
Update benchmark script to easily test llama-3 (#83 )
bhavya01
committedMay 17, 2024

Commits on May 14, 2024

Update JetStream grpc proto to support I/O with text and token ids (#78 )
JoeZijunZhou
committedMay 14, 2024

Commits on May 8, 2024

Add metadata metrics (#77 )
yeandy
committedMay 8, 2024

Commits on May 7, 2024

Accuracy eval mlperf (#76 )
jwyang-google
committedMay 7, 2024

Commits on Apr 25, 2024

Fix float division by zero in benchmark (#62 )
FanhaiLu1
committedApr 25, 2024
Fix benchmark script type issue (#59 )
JoeZijunZhou
committedApr 25, 2024

Commits on Apr 24, 2024

Align Tokenizer in JetStream (#40 )
JoeZijunZhou
committedApr 24, 2024

Commits on Apr 19, 2024

Update server host default value (#43 )
JoeZijunZhou
committedApr 19, 2024
Reformat benchmark metrics (#42 )
yeandy
committedApr 19, 2024

Commits on Apr 9, 2024

add sample_idx for debugging (#32 )
morgandu
committedApr 9, 2024

Commits on Apr 4, 2024

Enable pylint linter and pyink formatter (#29 )
JoeZijunZhou
committedApr 4, 2024

Commits on Apr 3, 2024

Support JetStream MaxText user guide (#28 )
JoeZijunZhou
committedApr 3, 2024

Commits on Apr 2, 2024

refactor to sample before tokenize (#26 )
morgandu
committedApr 2, 2024

Commits on Mar 30, 2024

Add model ckpt conversion and AQT scripts for JetStream MaxText Serving (#23 )
JoeZijunZhou
committedMar 30, 2024
Update benchmark to run openorca dataset (#21 )
morgandu
committedMar 30, 2024

Commits on Mar 27, 2024

Fix starvation with async server and interleaving optimization (#13 )
JoeZijunZhou
committedMar 27, 2024

Commits on Mar 26, 2024

Allows more requests than available data. (#19 )
patemotter
committedMar 26, 2024

Commits on Mar 25, 2024

Adds filtering for sharegpt based on conversation starter. (#17 )
patemotter
committedMar 25, 2024

Commits on Mar 15, 2024

Replace token_list with generated_text in saved outputs (#12 )
FanhaiLu1
committedMar 15, 2024
Add option to warm up (#11 )
qihqi
committedMar 15, 2024

Commits on Mar 14, 2024

Fix output token drop issue (#9 )
JoeZijunZhou
committedMar 14, 2024
Set max_output_length by 1024 as default and use max_output_length in inference request (#10 )
FanhaiLu1
committedMar 14, 2024

Commits on Mar 8, 2024

Save request outputs and add eval accuracy support (#8 )
FanhaiLu1
committedMar 8, 2024

Commits on Mar 6, 2024

Support gracefully stopping orchestrator and server (#6 )
JoeZijunZhou
committedMar 6, 2024

Commits on Mar 5, 2024

First support necessary for MaxText (#5 )
rwitten
committedMar 5, 2024
Add benchmarks maximum threads config (#4 )
JoeZijunZhou
committedMar 5, 2024

Commits on Mar 4, 2024

Add online serving benchmark script
JoeZijunZhou
committedMar 4, 2024