huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-16 10:38:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix the quantized example. (#564)

Browse Source
This commit is contained in:
Laurent Mazare
2023-08-23 11:09:55 +01:00
committed by GitHub
parent 89a00b56cc
commit 0f4ff8a739
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
candle-examples/examples/quantized/main.rs
View File

@ -605,8 +605,8 @@ fn main() -> anyhow::Result<()> {
);
println!("params: {:?}", model.hparams);
let default_gqa = match args.which {
Which::L7b | Which::L13b => 1,
Which::L70b => 8,
Which::L7b | Which::L13b | Which::L7bChat | Which::L13bChat => 1,
Which::L70b | Which::L70bChat => 8,
};
ModelWeights::from_ggml(model, args.gqa.unwrap_or(default_gqa))?
}