NVIDIA / NVIDIA/TransformerEngine

can't install te v1.13 in databricks

Open
#1,782 3 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

build installation
Dominant language
Python
Stars
3.5k
Forks
831
Avg merge
3d 11h
Merged PRs (30d)
65

Description

While trying to install te in databricks (16.2 ml run time and l4 as the node type), the machine fails to build the wheels for te. initially it shown some error which indicates cuda.h file were missing. After figuring out this, it shows cudnn.h file was not found. I figured some of these error by manually copying and pasting the .h files. But still could not complete the installation. Can any one help me, in completing the installation of te in databricks.

Contributor guide

Open the contributing guide

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 reproducing the te v1.13 installation on Databricks 16.2 ML runtime with an L4 node, then inspect the wheel-build errors for missing cuda.h and cudnn.h. Done means the package installs successfully without manually copying header files.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
build-system, machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.