microsoft / microsoft/TRELLIS.2

setup.sh has a few failures with conda activation, cuda versioning, flash attn and pillow-simd requirements

Open
#111 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
11.3k
Forks
1.4k
PR merge metrics
No merged PRs in 30d

Description

Minor issues on the install when run as as ./setup.sh with options:

--new-env 

conda is not initalized in shells so trellis gets installed in current env.

--basic    

pillow-simd installation can fail, some reqd pkgs, libjpeg-turbo zlib gxx_linux-64 are missing,

  • results in double pillow and pillow-simd installations
  • if using conda, links to both conda and system libs
--flash-attn 

errs on `No Module Named 'torch' https://github.com/Dao-AILab/flash-attention/issues/246

"$PLATFORM" = "cuda" with   --new-env  and   --flexgemm | --nvdiffrast  | --nvdiffrec 

If machine CUDA != torch cuda, these fail, as they require compiling against torch's cuda version.

Env. Ubuntu 24.04 + Conda + Nvidia A6000 + Cuda 13.0

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reading setup.sh and tracing the --new-env, --basic, --flash-attn, --flexgemm, --nvdiffrast, and --nvdiffrec branches. Reproduce the reported installation on Ubuntu 24.04 with Conda, an Nvidia A6000, and CUDA 13.0; done means each affected option handles environment activation, dependencies, and the Torch/CUDA version relationship without the reported failures.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, shell
Domain
build-system, devops, machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.