huggingface / huggingface/candle
Quantized models on multi-GPU
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 21.1k
- Forks
- 1.8k
- Avg merge
- 16h 42m
- Merged PRs (30d)
- 25
Description
I'm experimenting with the new implementation of CUDA acceleration for quantized models and wondering how to use sharded tensors in this context. I'm having a hard time adapting the ShardedVarBuilder to load like quantized_var_builder::VarBuilder::from_gguf.
Do you have any recommendations on the best approach in this case?
@LaurentMazare
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
Start by comparing the mentioned ShardedVarBuilder with quantized_var_builder::VarBuilder::from_gguf and the new CUDA acceleration implementation. Determine how sharded tensors should load quantized models across multiple GPUs; the issue is complete when the approach is defined and supported or documented.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- distributed-systems, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100