[Bug]: AMD GPU docker image build No matching distribution found for torch==2.6.0.dev20241113+rocm6.2

[samos123]

Your current environment

Archlinux 13th Gen Intel(R) Core(TM) i9-13900HX environment to build the docker image

Model Input Dumps

No response

🐛 Describe the bug

Trying to build the AMD GPU docker image:

git checkout v0.6.6.post1
DOCKER_BUILDKIT=1 docker build -f Dockerfile.rocm -t substratusai/vllm-rocm:v0.6.6.post1 .

Results in following error:

1.147 Looking in indexes: https://pypi.org/simple, https://download.pytorch.org/whl/nightly/rocm6.2
1.717 ERROR: Could not find a version that satisfies the requirement torch==2.6.0.dev20241113+rocm6.2 (from versions: 1.7.1, 1.8.0, 1.8.1, 1.9.0, 1.9.1, 1.10.0, 1.10.1, 1.10.2, 1.11.0, 1.12.0, 1.12.1, 1.13.0, 1.13.1, 2.0.0, 2.0.1, 2.1.0, 2.1.1, 2.1.2, 2.2.0, 2.2.1, 2.2.2, 2.3.0, 2.3.1, 2.4.0, 2.4.1, 2.5.0, 2.5.1, 2.6.0.dev20241119+rocm6.2, 2.6.0.dev20241120+rocm6.2, 2.6.0.dev20241121+rocm6.2, 2.6.0.dev20241122+rocm6.2)
2.135 ERROR: No matching distribution found for torch==2.6.0.dev20241113+rocm6.2
------
Dockerfile.rocm:49
--------------------
  48 |     # Install torch == 2.6.0 on ROCm
  49 | >>> RUN --mount=type=cache,target=/root/.cache/pip \
  50 | >>>     case "$(ls /opt | grep -Po 'rocm-[0-9]\.[0-9]')" in \
  51 | >>>         *"rocm-6.2"*) \
  52 | >>>             python3 -m pip uninstall -y torch torchvision \
  53 | >>>             && python3 -m pip install --pre \
  54 | >>>                 torch==2.6.0.dev20241113+rocm6.2 \
  55 | >>>                 'setuptools-scm>=8' \
  56 | >>>                 torchvision==0.20.0.dev20241113+rocm6.2 \
  57 | >>>                 --extra-index-url https://download.pytorch.org/whl/nightly/rocm6.2;; \
  58 | >>>         *) ;; esac
  59 |     
--------------------
ERROR: failed to solve: process "/bin/sh -c case \"$(ls /opt | grep -Po 'rocm-[0-9]\\.[0-9]')\" in         *\"rocm-6.2\"*)             python3 -m pip uninstall -y torch torchvision             && python3 -m pip install --pre                 torch==2.6.0.dev20241113+rocm6.2                 'setuptools-scm>=8'                 torchvision==0.20.0.dev20241113+rocm6.2                 --extra-index-url https://download.pytorch.org/whl/nightly/rocm6.2;;         *) ;; esac" did not complete successfully: exit code: 1

Before submitting a new issue...

• Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

[samos123]

Trying with these changes:

--- a/Dockerfile.rocm
+++ b/Dockerfile.rocm
@@ -51,9 +51,9 @@ RUN --mount=type=cache,target=/root/.cache/pip \
         *"rocm-6.2"*) \
             python3 -m pip uninstall -y torch torchvision \
             && python3 -m pip install --pre \
-                torch==2.6.0.dev20241113+rocm6.2 \
+                torch==2.6.0.dev20241122+rocm6.2 \
                 'setuptools-scm>=8' \
-                torchvision==0.20.0.dev20241113+rocm6.2 \
+                torchvision \
                 --extra-index-url https://download.pytorch.org/whl/nightly/rocm6.2;; \
         *) ;; esac

which gets me past the previous error and some other errors.

Update: laptop now crashes when trying to build. Trying to set max_jobs to 1 and seeing if that helps.

Update 2: It's building very slowly but no longer crashing. Just need to wait now.

[hongxiayang]

Correct. it needs to cherry pick this PR: #12172