diff options
| author | Laurent Mazare <laurent.mazare@gmail.com> | 2024-04-28 20:05:05 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-04-28 20:05:05 +0200 |
| commit | eb26e2467eb4cb5ca507324cc3245600c104f219 (patch) | |
| tree | 7aa8fead605a786c38d0b6d2835342240e80c9a2 /candle-nn | |
| parent | c68ed8963fb6fc842f20d84baa07ff97b56aedb4 (diff) | |
| download | candle-eb26e2467eb4cb5ca507324cc3245600c104f219.tar.gz candle-eb26e2467eb4cb5ca507324cc3245600c104f219.tar.bz2 candle-eb26e2467eb4cb5ca507324cc3245600c104f219.zip | |
Add the cuda dequantize f16 kernels. (#2137)
* Add the cuda dequantize f16 kernels.
* Expose the cuda kernels.
* Add some testing + fix.
* Test the other cases too.
* A few more tests.
* Add an environment variable to enable the dequantize f16 + matmul behavior.
Diffstat (limited to 'candle-nn')
0 files changed, 0 insertions, 0 deletions