-
Notifications
You must be signed in to change notification settings - Fork 6
Open
Description
Problem Description
madengine/src/madengine/tools/run_models.py
Line 593 in eca77f7
| + " --pull -f " |
Example command
madengine run --tags pyt_deepspeed_megatron_gpt2 --live-output --additional-context "{'guest_os':'UBUNTU', 'docker_build_arg':{'BASE_DOCKER':'compute-artifactory.amd.com:5000/rocm-plus-docker/framework/compute-rocm-rel-6.4:114_ubuntu22.04_py3.10_pytorch_release-2.6_6b7dbfa'}}" --timeout 0 --clean-docker-cache
The command fails with the following logs if replacing the BASE_DOCKER argument with any local image.
#2 [internal] load metadata for docker.io/library/swdev-542227:latest
#2 ERROR: pull access denied, repository does not exist or may require authorization: server message: insufficient_scope: authorization failed ------
> [internal] load metadata for docker.io/library/swdev-542227:latest:
------
pyt_deepspeed_megatron_gpt2.ubuntu.amd.Dockerfile:3
--------------------
1 | # CONTEXT {'gpu_vendor': 'AMD', 'guest_os': 'UBUNTU'}
2 | ARG BASE_DOCKER=rocm/pytorch:latest
3 | >>> FROM $BASE_DOCKER
4 | ENV WORKSPACE_DIR=/workspace
5 | RUN mkdir -p $WORKSPACE_DIR
--------------------
ERROR: failed to solve: swdev-542227:latest: failed to resolve source metadata for docker.io/library/swdev-542227:latest: pull access denied, repository does not exist or may require authorization: server message: insufficient_scope: authorization failed
Confirmed removing --pull can fix the problem.
Operating System
Does not matter
CPU
Does not matter
GPU
Does not matter
ROCm Version
Does not matter
ROCm Component
No response
Steps to Reproduce
No response
(Optional for Linux users) Output of /opt/rocm/bin/rocminfo --support
No response
Additional Information
No response
Metadata
Metadata
Assignees
Labels
No labels