Veritable Tech Blog

Jul 11
[Tip] TorchScript Supports Half Precision
Speeding up inference for models trained with mixed precision
Jul 6
Self-Supervised Domain Adaptation
Preliminary experiments show great promise
Jun 28
Using Julia to Do Whole Word Masking
Syntax almost as friendly as Python, while running up to 100x faster
Jun 16
[Failure Report] Distill Fine-tuned Transformers into Recurrent Neural Networks
Transferring knowledge for downstream tasks requires more efforts
May 4
Deploying EfficientNet Model using TorchServe
A Case Study