Fixed some MobileVLM's inference bugs. Added more tests on different devices. #33
bench.yml
on: pull_request
bench-server-baseline
12m 40s
Annotations
1 error and 2 warnings
bench-server-baseline
Error setting status:
Resource not accessible by integration
Request object:
{
"context": "bench-server-baseline",
"description": "{\"req\":{\"p90\":25863.19,\"avg\":9281.61},\"pp\":{\"p90\":732.4,\"avg\":241.61},\"tg\":{\"p90\":264.26,\"avg\":98.9}}",
"state": "success",
"owner": "ggerganov",
"repo": "llama.cpp",
"sha": "4ab46218c15d7921eaf25f76d6d965f56b5ace14",
"target_url": ""
}
|
bench-server-baseline
Node.js 16 actions are deprecated. Please update the following actions to use Node.js 20: actions/checkout@v3, Sibz/github-status-action@v1. For more information see: https://github.blog/changelog/2023-09-22-github-actions-transitioning-from-node-16-to-node-20/.
|
bench-server-baseline
The following actions uses node12 which is deprecated and will be forced to run on node16: Sibz/github-status-action@v1. For more info: https://github.blog/changelog/2023-06-13-github-actions-all-actions-will-run-on-node16-instead-of-node12-by-default/
|
Artifacts
Produced during runtime
Name | Size | |
---|---|---|
benchmark-results
|
189 KB |
|