Python 3.10.12 : My colab test with Gated recurrent unit mechanism - part 037.

This is a simple example for Gated recurrent unit mechanism known as GRUs.

You can find this in my GitHub colab project.

import numpy as np
import tensorflow as tf
import keras
from keras import layers
units = 64
tf.keras.layers.GRU(
    units,
    activation="tanh",
    recurrent_activation="sigmoid",
    use_bias=True,
    kernel_initializer="glorot_uniform",
    recurrent_initializer="orthogonal",
    bias_initializer="zeros",
    kernel_regularizer=None,
    recurrent_regularizer=None,
    bias_regularizer=None,
    activity_regularizer=None,
    kernel_constraint=None,
    recurrent_constraint=None,
    bias_constraint=None,
    dropout=0.0,
    recurrent_dropout=0.0,
    return_sequences=False,
    return_state=False,
    go_backwards=False,
    stateful=False,
    unroll=False,
    time_major=False,
    reset_after=True,
)
inputs = tf.random.normal([32, 10, 8])
gru = tf.keras.layers.GRU(4)
output = gru(inputs)
print(output.shape)

gru = tf.keras.layers.GRU(4, return_sequences=True, return_state=True)
whole_sequence_output, final_state = gru(inputs)
print(whole_sequence_output.shape)
print(final_state.shape)

Python 3.10.12 : My colab tutorials and news from colab - part 036.

Today I recapitulated a bit the artificial intelligence part and a simple example for google drive.

I created two notebooks in collaboration and added them to my github repo.

The most interesting is the one with textgenrnn.

textgenrnn is an modern neural network architecture which utilizes new techniques as attention-weighting and skip-embedding to accelerate training

The last notebook colab is catafest_045.

News : Colab behavior through runtime .

I would like Google to emphasize more on the development side some elements that work like robots by interfacing with the development side.

Today I worked a little on artificial intelligence and I realized that it doesn't create textgenrnn_weights.hdf5 file for training created with the Python textgenrnn mode.

A solution is to reset the runtime with Ctrl+M and resume running.

They specify RESTART RUNTIME when using Python modules, see:

WARNING: The following packages were previously imported in this runtime:
   [numpy]
You must restart the runtime in order to use newly installed versions.

In this case, with the creation of textgenrnn_weights.hdf5 file, it is more difficult to understand and cannot be seen easily.

Python 3.10.12 : My colab tutorials and news from colab - part 035.

In this notebook I will show you how to use python to run a program written in the programming language for CUDA.

This allows you to use NVIDIA CUDA Compiler Driver NVCC, see this official webpage.

NVCC Plugin for Jupyter Notebook by https://github.com/andreinechaev/nvcc4jupyter.

The example I tested is simple:

# This is formatted as CUDA code
__global__ void cuda_hello(){
    printf("Hello World from GPU!\n");
}

int main() {
    cuda_hello<<<1,1>>>();
    return 0;
}

See my GitHub project for colab.

Python 3.8.10 : My colab tutorials and news from colab - part 034.

I add a new colab notebook with a simple source code to list all running VM processes from the colab notebook

You can see more examples on my GitHub colab google repo.

This is the source code:

%%sh
echo "List all running VM processes."
ps -ef
echo "Done"

News : About my work and one of my websites.

I would like to bring to the attention of those who follow my activity on my websites and appreciate the inability to continue with one of the websites: free-tutorials.org. It is currently hosted on a free host, but I cannot import it 100%, which has led me not to complete it with new posts. The continuation of the activities there, considering the limited time, will be carried out on my blogs with the defined theme with which I started: Linux - Fedora, Graphics, Python, Pygame.

In the meantime, because the host is expensive and until now someone has helped me to host it on his server, it is possible to sell the domain: free-tutorials.org - I receive purchase offers at my personal Yahoo email address catafest@yahoo.com.

Minimum starting price 250 euros, because the domain is old from 2018.