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

Add to the cuda example a reproduction of the issue. (#579)

Browse Source 
* Add to the cuda example a reproduction of the issue.

* Tweak.

* Add a test using non-square matrixes.

* Fix the conv2d kernel.

* Display the error.

* And tweak the comment.
This commit is contained in:
Laurent Mazare
2023-08-24 12:07:31 +01:00
committed by GitHub
parent dd64465899
commit ca318a6ec7
3 changed files with 58 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
candle-core/examples/cuda_basics.rs
View File

@ -9,8 +9,17 @@ use candle_core::{Device, Tensor};
fn main() -> Result<()> {
let device = Device::new_cuda(0)?;
let t = Tensor::rand(-1f32, 1f32, 96, &device)?;
println!("{t}");
let in_t = Tensor::rand(-1f32, 1f32, (1, 3, 12, 7), &device)?;
let k_t = Tensor::rand(-1f32, 1f32, (6, 3, 1, 1), &device)?;
let out_t = in_t.conv2d(&k_t, 0, 1, 1)?;
println!("{out_t}");
let in_t = in_t.to_device(&Device::Cpu)?;
let k_t = k_t.to_device(&Device::Cpu)?;
let out_t2 = in_t.conv2d(&k_t, 0, 1, 1)?;
let diff = (out_t.to_device(&Device::Cpu)? - out_t2)?
.sqr()?
.sum_all()?;
println!("{diff}");
let t = Tensor::randn(0f32, 1f32, (2, 4, 96, 96), &device)?;
let w = Tensor::randn(0f32, 1f32, (320, 4, 3, 3), &device)?;