Pablo Alonso

Thanks for the fast reply. I tried installing from the master, but I got the same result. This is the TensorFlow log I got with the `prepare` call: ``` 2021-06-04...

I found that the model initializes in a reasonable time when doing the inference from Tensorflow's Python bindings: ```Python import tensorflow as tf import numpy as np model_file = 'effnet5'...

Hello, lately I've been revisiting this problem, and I've found a way to reproduce it directly in TensorFlow by using the CLI tflite converter: ```bash tflite_convert --saved_model_dir effnet --output_file effnet.tflite...

Could you provide a link to the Colab and the CUDA/CuDNN versions you are using?

Essentia [was built with CUDA 11.2](https://github.com/MTG/essentia-builds/blob/2c63617dd7b04bee5b21fc7b1dab91f6e4cb1492/build_scripts/install_cuda.sh#L4) and it is incompatible with CUDA 11.1. While I couldn't find an easy way to update the CUDA version in colab, by [coping you...

I think that neither all the `*.py` files are examples nor all the `*.ipynb` are tutorials. But maybe the extension is already good enough for the division. What about `src/examples/python`...

@dbogdanov should I go on and create a PR with this organization?

Mentioned in #1180

Hi, have you checked the parameters `lowlevelSilentFrames` and `tonalSilentFrames` from [MusicExtractor](https://essentia.upf.edu/reference/std_MusicExtractor.html)? By default, they add noise to the silent frames making it non-deterministic.

You can deactivate Essentia logging as in #1213 The proposed solution for the TensorFlow logging does work for me. It is important that the `TF_CPP_MIN_LOG_LEVEL` variable is set before importing...