DeepSpeed
f7ef4b5e
- fixing the softmax masking when using triangular masking
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
4 years ago
fixing the softmax masking when using triangular masking
References
#1491 - Fixing the transformer APIs to return tuple as the output (if needed)
#2451 - Inference support for encoder-decoder architecture
Author
Reza Yazdani
Parents
fcb3ca5e
Loading