4 bit models on hopper support
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1.8k
- Forks
- 124
- PR merge metrics
- No merged PRs in 30d
Description
We would like to run the GLM 5.2 model in int4/mxfp4 on a single node with 8×H20 GPUs. Could you let us know whether Tilert currently has plans to support 4-bit inference on the Hopper architecture?
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
The issue names GLM 5.2, int4/mxfp4, Hopper, and an 8×H20 single-node target, but no files, tests, or entry points. Start by checking TileRT's current support and documentation for 4-bit inference on Hopper; done would be a clear status or implementation plan for that configuration.
Written by the indexing model from the issue text.
Assessment
- Domain
- ai, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100