Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
This is a TensorFlow implementation of the WaveNet generative neural network architecture https://deepmind.com/blog/wavenet-generative-model-raw-audio/ for text generation.
| Date | Stars |
|---|---|
| 2026-07-31 | 344 |
| 2026-08-06 | 344 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# A TensorFlow implementation of DeepMind's WaveNet paper for text generation. This is a TensorFlow implementation of the [WaveNet generative neural network architecture](https://deepmind.com/blog/wavenet-generative-model-raw-audio/) for <b>text</b> generation. # Previous work <table style="border-collapse: collapse"> <tr> <td> <p> Originally, the WaveNet neural network architecture directly generates a raw audio waveform, showing excellent results in text-to-speech and general audio generation (see the DeepMind blog post and paper for details). </p> <p> This network models the conditional probability to generate the next sample in the audio waveform, given all previous samples and possibly additional parameters. </p> <p> After an audio preprocessing step, the input waveform is quantized to a fixed integer range. The integer amplitudes are then one-hot encoded to produce a tensor of shape <code>(num_samples, num_channels)</code>. </p> <p> A convolutional layer that only accesses the current and previous inputs then reduces the channel dimension. </p> <p> The core of the network is constructed as a stack of <em>causal dilated layers</em>, each of which is a dilated convolution (convolution with holes), which only accesses the current and past audio samples. </p> <p> The outputs of all layers are combined and extended back to the original number of channels by a series of dense postprocessing layers, followed by a softmax function to transform the outputs into a categorical distribution. </p> <p> The loss function is the cross-entropy between the output for each timestep and the input at the next timestep. </p> <p> In this repository, the network implementation can be found in <a href="./wavenet/model.py">wavenet.py</a>. </p> </td> <td width="300"> <img src="images/network.png" width="300"></img> </td> </tr> </table> ## New approach This work is based in one implementation of the original WaveNet model ([Wavenet](https://github.com/ibab/tensorflow-wavenet)), but applying some modifications. In summary: we are going to use the WaveNet model as a <b>text generator</b>. We'll use raw text data (characters), instead of raw audio files, and once the network is trained, we'll use the conditional probability found to generate samples (characters) into an self-generating process. Only printable ASCII characters (Dec. 0 up to 255) is supported right now. ## Results Pretty interesting results are reached!! Feeding the network with enough text and training, the model is able to memorize the probability of the characters disposition (in a lenguage), and generate later even a very similar text!! For example, using the <a href="./data/texts/ptb.train.txt">Penn Tree Bank</a> (PTB) dataset, and only after 15000 steps of training (with low set of parameters setting) this was the self-generated output (the final loss was around 1.1): "Prediction is: 300-servenns on the divide mushin <unk> <unk> attore and <unk> operations losers nis called him for investment it was with as pursicularly federal and sotheby <unk> d. reported firsts truckhe of the guarantees as paining at the available ransions i 'm new york for basicane as a facerement of its a set to the u.s. spected on install <unk> <unk> death about in the little there have a $ N million or N N bilot in closing is <unk> <unk> of a trading a congress of society or N cents for policy half feeling the does n't people of general and the <unk> crafted ended yesterday still also arjas trading an effectors that a can singaes about N bound who <unk> that mestituty was below for which unrecontimer 's <unk> have day simple d. frisons already earnings on the annual says had minority four-$ N sance for an advised in reclution by <unk> <unk> from $ N million morris selpiculations the <to will quarter benever on july coming buy-week that the nation tore under new york beyond N million month expected thomas last disappointing to recognition on first had <unk> not year break government thes
Excerpt of 7,806 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:eccb04655d36d22f, llm:description: 'TensorFlow implementation of the WaveNet generative neural network architecture ... for text generation.'
matched fp:eccb04655d36d22f, llm:description: 'TensorFlow implementation of the WaveNet generative neural network architecture ... for text generation.'
matched fp:eccb04655d36d22f, llm:description: 'TensorFlow implementation of the WaveNet generative neural network architecture ... for text generation.'