search

triton-inference-server / fastertransformer_backend

BSD 3-Clause "New" or "Revised" License
411 stars 134 forks source link

Fix docker build to work with Triton 23.05 image #153

Closed samiur closed 1 year ago

samiur commented 1 year ago

This PR adds CUDA_FLAGS that fix the build when using the 23.05 triton base image.

It also optimizes the docker commands to create fewer, smaller layers, reducing the overall image size.

samiur commented 1 year ago

Ping @jbkyang-nvi @Tabrizian

Tabrizian commented 1 year ago

@samiur Thanks for the contribution! Are you able to sign the CLA as instructed here: https://github.com/triton-inference-server/server/blob/main/CONTRIBUTING.md#contributor-license-agreement-cla