Prashant Chaudhari
Prashant Chaudhari
My quantization, ``` converter = tf.lite.TFLiteConverter.from_keras_model(model) input_shape = model.input.shape inference_dtype = tf.uint8 converter.optimizations = [tf.lite.Optimize.DEFAULT] np.random.seed(20) def representative_dataset_gen(): for _ in range(100): # Get sample input data as a numpy...
@Namburger this problem was solved when I use a smaller input size. For example, `[1, 4000, 1]`. However, I incorporated some custom normalization layers for a new model. This is...
> @prashant45 did you added the line to make the batch size static as I suggested? @Namburger no, I just used a smaller input-size
> @prashant45 lol can you try it? that was my suggestion. I tired it and it did not work. Otherwise I would had mentioned it. :P
> I see, can you possibly attach the code to reproduce your model? it is rather weird that I've not seen this issue before ``` import os os.environ["CUDA_VISIBLE_DEVICES"] = "-1"...
Even I am having problems connecting 2 bluetooth devices? Any alternatives for this issue ?
How can I read continuously from the bluetooth device?
@joshuapinter Thanks that worked for now 👍
@joshuapinter did you manage to set up the event listener?
perhaps this could be of help [react-native-event-listeners](https://github.com/meinto/react-native-event-listeners).