trained-ternary-quantization

trained-ternary-quantization copied to clipboard

Hi, it seems that `HYPERPARAMETER_T` is set to a constant value of `0.15` and is never updated during backpropagation. If that's true, I just want to mention it would be...

vinsis

Size mismatch error on Resnet at quantization stage

2

Hi, Trying TTQ on RESNET18 but getting a runtime error. Can't seem to find what the issue is: /home/user2/Desktop/pttq/resnet_caltech/trained-ternary-quantization-master/utils/training.pyc in train(model, loss, optimization_step_fn, train_iterator, val_iterator, n_epochs, patience, threshold, lr_scheduler) 53...

pgadosey

Gradient w.r.t the negative scaling factor might be wrong?

1

In trained-ternary-quantization/utils/quantization.py, line42 The last returned value of function "get_grads" is the gradient w.r.t the negative scaling factor. I think the code might be wrong...... (not sure) Consider a simple...

magicwyzh

Possible to use pretrained models in quantization?

9

Hi, Thanks for this. Is it possible to just use my own pretrained models for quantization?

pgadosey

Are full precision weights normalized before quantization as in the original paper? The original paper says: "First, we normalize the full-precision weights to the range [-1, +1] by dividing each...

sangmain