You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/source/getting_started/amd-installation.rst
+4-4Lines changed: 4 additions & 4 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -14,7 +14,7 @@ Requirements
14
14
* Python: 3.8 -- 3.11
15
15
* GPU: MI200s (gfx90a), MI300 (gfx942)
16
16
* Pytorch 2.0.1/2.1.1/2.2
17
-
* ROCm 5.7 ((Verified on 3.10) or ROCm 6.0 ((Verified on 3.9))
17
+
* ROCm 5.7 (Verified on python 3.10) or ROCm 6.0 (Verified on python 3.9)
18
18
19
19
Installation options:
20
20
@@ -97,22 +97,22 @@ You can build and install vLLM from source:
97
97
98
98
Build a docker image from `Dockerfile.rocm`, and launch a docker container.
99
99
100
-
The `Dokerfile.rocm` is designed to support both ROCm 5.7 and ROCm 5.8. It provides flexibility to customize the build of docker image using the following arguments:
100
+
The `Dokerfile.rocm` is designed to support both ROCm 5.7 and ROCm 6.0. It provides flexibility to customize the build of docker image using the following arguments:
101
101
102
102
* `BASE_IMAGE`: specifies the base image used when running `docker build`, specifically the PyTorch on ROCm base image. We have tested ROCm 5.7 and ROCm 6.0. The default is `rocm/pytorch:rocm5.7_ubuntu22.04_py3.10_pytorch_2.0.1`
103
103
* `FX_GFX_ARCHS`: specifies the GFX architecture that is used to build flash-attention, for example, `gfx90a;gfx942` for MI200 and MI300. The default is `gfx90a;gfx942`
104
104
* `FA_BRANCH`: specifies the branch used to build the flash-attention in `ROCmSoftwarePlatform's flash-attention repo <https://github.com/ROCmSoftwarePlatform/flash-attention>`_. The default is `3d2b6f5`
105
105
106
106
Their values can be passed in when running `docker build` with `--build-arg` options.
107
107
108
-
For example, to build docker image for vllm on ROCm 6.0, we can run:
108
+
For example, to build docker image for vllm on ROCm 6.0, you can run:
0 commit comments