Liuyi Jin comments

Results 14 comments of


                                            Liuyi Jin

Error: Trying to access flag --preserve_unused_tokens before flags were parsed

I got same error, and it just happens today.

Roadmap

Hi, I'm wondering if you would share your modification code to TFLite, which implements the GPU operator-level profiling?

sre_constants.error: unterminated character set at position 5

I am in the same issue, have anyone got any update on this. I am using python 3.7.3. I am wondering if this is because of the incompatible python versions.

Pre-trained RNN-T .h5 to tflite Conversion Issue

Any updated solution here? I encountered the same issue here with tf2.9.

Pre-trained RNN-T .h5 to tflite Conversion Issue

This bug has been resolved [here](https://github.com/tensorflow/tensorflow/issues/42410#issuecomment-1201150119). Basically, we just need to load model while are are loading the concrete function, i.e., change [this line](https://github.com/TensorSpeech/TensorFlowASR/blob/main/tensorflow_asr/helpers/exec_helpers.py#L41) to `converter = tf.lite.TFLiteConverter.from_concrete_functions([concrete_func], model)`.

RNN-Transducer TFlite conversion Broken

Any updated solution here? I encountered the same issue here with tf2.9.

RNN-Transducer TFlite conversion Broken

This bug has been resolved https://github.com/tensorflow/tensorflow/issues/42410#issuecomment-1201150119. Basically, we just need to load model while are are loading the concrete function, i.e., change [this line](https://github.com/TensorSpeech/TensorFlowASR/blob/main/tensorflow_asr/helpers/exec_helpers.py#L41) to converter = tf.lite.TFLiteConverter.from_concrete_functions([concrete_func], model).

ConverterError:input resource[0] expected type resource !=float, the type of contextnet_greedy_while_contextnet_decoder_contextnet_prediction_embedding_embedding_lookup_24797_0[0]

Any updated solution here? I encountered the same issue here with tf2.9.

ConverterError:input resource[0] expected type resource !=float, the type of contextnet_greedy_while_contextnet_decoder_contextnet_prediction_embedding_embedding_lookup_24797_0[0]

Revert the mobile support

> @kolaente @techknowlogick thanks for the quick response! > > > Just out of curiosity, what is your use case for mobile? > > A bit of context: We need...