Open dhruv1710 opened 1 month ago
Dynamic quantization added to the Linear and LSTM layers, converted to 8bit.
Thanks for your contribution! Maybe we can add a -q option to choose whether or not quantize the model.
-q
That's a good feature, I'll add that
You never did
Dynamic quantization added to the Linear and LSTM layers, converted to 8bit.