I’m trying to convert an 8-bit quantized TFLITE model to MLIR. Any suggestions on what tool flow to use?
It depends what you mean by “convert to MLIR”: you can always import a TFLite flattbuffer to the MLIR TFLite dialect. Actually to generate the model in the first place the TF to TFlite converter uses MLIR conversions internally.
Something like this should do:
bazelisk run //tensorflow/compiler/mlir/lite:flatbuffer_translate -- --tflite-flatbuffer-to-mlir model.fb