rino20

Hi @felicitywang1 This answer might not be correct since I don't know your model exactly. If your model is available to use tf.keras.models.clone_model(your_model), you can try applying quantization by manually...

MOT officially only suupport keras model. But you can try the manual loading weights if you want to apply MOT techniques. We have wrapper class for applying MOT to the...

Hi, could you share more info about the OD model? I am confused whether the model is keras model or not. You said that it is not a keras model,...

HI @teijeong could you take a look?

Hi @somum This has been resolved in https://github.com/tensorflow/model-optimization/pull/754 Could you sync your tfmot with HEAD or nightly and re-test the code?

@wwwind Could you take a look? Thanks.

Sorry, the example code is not available at this moment. Btw, I cannot imagine any specific issue on applying Pruning API to GNMT model, so what about trying it similar...

Hi @YannPourcenoux Since you haven't set the pruning parameters, the default option is applied - ConstantSparsity, with pruning frequency 100. https://github.com/tensorflow/model-optimization/blob/master/tensorflow_model_optimization/python/core/sparsity/keras/pruning_schedule.py#L141 That means, your model will be pruned at every...

Hi @jhelsas, Thanks for reporting this. The Pruning policy you are using (PruneForLatencyOnXNNPack) is customized for a specific subgraph structure as shown in the guide's example. It searches for a...

Hi @vvolhejn , Since you haven't set the pruning parameters, the default option is applied - ConstantSparsity, with pruning frequency 100. https://github.com/tensorflow/model-optimization/blob/master/tensorflow_model_optimization/python/core/sparsity/keras/pruning_schedule.py#L141 That means, your model will be pruned at...