It became larger after converting to tflite model

[weinixuehao]

The ckpt file contains two model states so a single one is half the size.

I found that if the input resolution is larger the converted tflite model is larger

[peterjc123]

https://github.com/alibaba/TinyNeuralNetwork/blob/main/docs/FAQ.md#what-if-duplicate-tensors-is-generated-in-the-tflite-model-eg-when-performing-static-quantization-for-lstms

[peterjc123]

Q: What if duplicate tensors is generated in the TFLite model (e.g. when performing static quantization for LSTMs)?

A: You may try out group_tensors=True to remove those duplicates.

[weinixuehao]

@peterjc123
After converting with group_tensors, it is still much larger and the inference result is wrong.

[peterjc123]

It was caused by the pattern. a = torch.cat((b, torch.zeros(...)), dim=...). We may write a optimizer pass for that.