symmetric quantization to FBGEMM prefill token-wise FP8 (fixed) #5991
ColinPepplerbuild_wheels_genai_linux_x86.yml
on: pull_request
generate-matrix
/
generate
8s
Matrix: pytorch/FBGEMM / build
Matrix: pytorch/FBGEMM / upload / upload
Annotations
1 warning
|
filter-matrix
The `python-version` input is not set. The version of Python currently in `PATH` will be used.
|
Artifacts
Produced during runtime
| Name | Size | Digest | |
|---|---|---|---|
|
pytorch_FBGEMM__3.10_cu126_x86_64
|
17.6 MB |
sha256:1250f52163fa66df82f4a507461c5a6e0a4604ae247c318d2ff73cbe1cb0db73
|
|
|
pytorch_FBGEMM__3.10_cu128_x86_64
|
49.7 MB |
sha256:e5eb292c8e53351345840c508297d65d04f0e913de23d3e05aab23aec72448ee
|
|
|
pytorch_FBGEMM__3.10_cu130_x86_64
|
47.6 MB |
sha256:fa48d48b4abd73d06a393ea87ac47d5594fb1a9fab910cbe1c0d17807ff7675b
|
|
|
pytorch_FBGEMM__3.10_rocm6.3_x86_64
|
11.2 MB |
sha256:4995a60262c5cfefdca4c53d91ba80ba95d28d2e4188e1eb26cc7eb74a5756e3
|
|
|
pytorch_FBGEMM__3.10_rocm6.4_x86_64
|
11.2 MB |
sha256:72ef9bf00cb75657f6750cb9953aa4f4ad1fb4fdda45cd86de9db4e28705b3c9
|
|