huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-15 18:28:24 +00:00
Code Issues Packages Projects Releases Wiki Activity

Improve the quantized whisper setup. (#1018)

Browse Source 
* Improve the quantized whisper setup.

* Fix the config file paths.

* Use the standard matmul where possible.
This commit is contained in:
Laurent Mazare
2023-10-02 17:17:46 +01:00
committed by GitHub
parent e04c789230
commit 089fc3b584
8 changed files with 66 additions and 49 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
candle-pyo3/src/lib.rs
View File

@ -867,7 +867,7 @@ impl PyQTensor {
/// Performs a quantized matrix multiplication, with the quantized tensor as the right hand side.
/// &RETURNS&: Tensor
fn matmul_t(&self, lhs: &PyTensor) -> PyResult<PyTensor> {
let qmatmul = ::candle::quantized::QMatMul::from_arc(self.0.clone());
let qmatmul = ::candle::quantized::QMatMul::from_arc(self.0.clone()).map_err(wrap_err)?;
let res = qmatmul.forward(lhs).map_err(wrap_err)?;
Ok(PyTensor(res))
}