Docker build for stable diffusion fails

[shrutiramesh1988]

Docker build for stable diffusion failed as shown below (tried NVIDIA's and Dell's implementations).
Could you please help.

docker build . -t nvidia_stablediffusion_pytorch_mlperf3.1 [+] Building 1.7s (3/3) FINISHED docker:default
=> [internal] load .dockerignore 0.0s
=> => transferring context: 2B 0.0s
=> [internal] load build definition from Dockerfile 0.0s
=> => transferring dockerfile: 719B 0.0s
=> ERROR [internal] load metadata for nvcr.io/ea-bignlp/ea-mm-sd-alpha/bignlp-mm-sd:23.09-py3 1.7s

[internal] load metadata for nvcr.io/ea-bignlp/ea-mm-sd-alpha/bignlp-mm-sd:23.09-py3:

---

Dockerfile:2

1 | ARG FROM_IMAGE_NAME=nvcr.io/ea-bignlp/ea-mm-sd-alpha/bignlp-mm-sd:23.09-py3
2 | >>> FROM ${FROM_IMAGE_NAME}
3 |
4 | RUN pip install --upgrade webdataset

ERROR: failed to solve: nvcr.io/ea-bignlp/ea-mm-sd-alpha/bignlp-mm-sd:23.09-py3: pulling from host nvcr.io failed with status code [manifests 23.09-py3]: 401 Unauthorized

[ruijietey]

I think your original error is that you should do docker login nvcr.io first, but I guess nvcr.io/ea-bignlp/ea-mm-sd-alpha/bignlp-mm-sd:23.09-py3 is also deprecated, so we should directly use NeMo container instead (https://catalog.ngc.nvidia.com/orgs/nvidia/containers/nemo)

However, I have issues with dependencies after using NeMo container (nvcr.io/nvidia/nemo:23.08, nvcr.io/nvidia/nemo:23.10) after running docker build with the updated Dockerfile

Hunk #1 succeeded at 614 (offset 4 lines).
patching file strategies/ddp.py
Hunk #1 FAILED at 191.
1 out of 1 hunk FAILED -- saving rejects to file strategies/ddp.py.rej
patching file trainer/connectors/logger_connector/result.py
Hunk #1 FAILED at 502.
Hunk #2 FAILED at 512.
2 out of 2 hunks FAILED -- saving rejects to file trainer/connectors/logger_connector/result.py.rej
The command '/bin/sh -c PL_ROOT=$(python -c "import pytorch_lightning; print(pytorch_lightning.file.replace('/init.py',''))"); patch -p3 -d${PL_ROOT} -i /source/lightning.v1.9.4.patch' returned a non-zero code: 1

[shrutiramesh1988]

I'm now facing the same issue with nemo container. Could anyone please let me know if this issue has been resolved.