3 ms·
Yeah currently experimenting with T5 large in FP16. Will then use the ONNX conversion script to decrease inference time and deploy it!
by FL33TW00D 6y ago
Yeah currently experimenting with T5 large in FP16. Will then use the ONNX conversion script to decrease inference time and deploy it!