Chirag Jain
|
50896ec574
Make nvcc threads configurable via environment variable (#885)
|
vor 9 Monaten |
Tri Dao
|
dc08ea1c33
Support H100 for other CUDA extensions
|
vor 1 Jahr |
Tri Dao
|
fa6d1ce44f
Add fused_dense and dropout_add_layernorm CUDA extensions
|
vor 2 Jahren |