Deyu Huang

Hi @doomooo , sorry for the late reply. Could you please share your `INT8` case or any simple reproduction script code? We will try to find a way to solve...

The reason why the [`FakeQuantWithMinMaxArgs`](https://tensorflow.google.cn/versions/r1.15/api_docs/python/tf/quantization/fake_quant_with_min_max_vars) only supports unit8 is the quantization range belonged. ``` inputs values are quantized into the quantization range ([0; 2^num_bits - 1] when narrow_range is false...

Hi @TomWildenhain-Microsoft, based on the result of my debugging, this repro code is crashed by this line. https://github.com/onnx/tensorflow-onnx/blob/a1a9343cef193ef45b7f91d511e9c7bd1dffb3e3/tf2onnx/tflite_utils.py#L200 Cause the `_get_tensor_details()` function will catch the `TensorSparsityParameters` of tflite interpreter. While...

Hi @r-remus @igalkr, thanks for your comments! `TensorScatterAdd` has added with this PR [Add TensorScatterAdd op for opset 16](https://github.com/onnx/tensorflow-onnx/pull/1949). You can try to convert it using latest main branch code...

The root cause is mainly from `select` op implementation. The select will choose `Mul/Add` ops implementation firstly which is faster than `Where` op. After debugging, the model gets wrong `nan`...

The fusion logic of `Conv` and `BatchNormalization` do in back-to-back optimizer. Could you check if your model conversion process is passed through the below code? https://github.com/onnx/tensorflow-onnx/blob/c67bcfb580be741ece8d9978a9b57bd2ce7367ee/tf2onnx/optimizer/back_to_back_optimizer.py#L191

Hi @Mypathissional , I think it's a expected behavior for tensorflow-onnx. Cause when I do the convert script, there is no `BatchNormalization` op, even before running the optimizer conversion. ```...

> @hwangdeyu Deyu Huang FTE Can you tell just for my understanding what happens when the operation that is present in the saved model but not present in the onnx...

Related issue: Keras LSTM converted to loop instead of ONNX LSTM op https://github.com/onnx/tensorflow-onnx/issues/1851

Thank you so much for so detailed issue, I am not very familiar with this part. I guess the keras2onnx is a good example to fix it. Need to do...