Turzo Bose

Turzo Bose

tbose [at] connect.ust.hk

Posts by Tags

Convolutions

Dimension tracking for Conv2D layer

5 minute read

Published:

Keeping track of the dimensions is a very cumbersome process if not properly understood, specially while designing the deep neural network model architectures. The stress adds up swiftly when we have to shift between different machine learning frameworks, ideally PyTorch and Tensorflow. It is crazy how although both these frameworks are essentially doing the same thing, the syntax makes it increasingly difficult to allow for fluid transitions between them. Hence, this post is written to make the dimension tracking process simpler, for both PyTorch and Tensorflow implementations.

DNS

Share WiFi through Ethernet Port with Raspberry Pi

6 minute read

Published:

Imagine your WiFi onboard your laptop broke down and you cannot use the internet even though the router and modem works fine. Or maybe you have a desktop computer without a WiFi adapter or dongle. How can you hack through this situation without having to invest in a dongle in the quickest possible way?

DeepLearning

Dimension tracking for Conv2D layer

5 minute read

Published:

Keeping track of the dimensions is a very cumbersome process if not properly understood, specially while designing the deep neural network model architectures. The stress adds up swiftly when we have to shift between different machine learning frameworks, ideally PyTorch and Tensorflow. It is crazy how although both these frameworks are essentially doing the same thing, the syntax makes it increasingly difficult to allow for fluid transitions between them. Hence, this post is written to make the dimension tracking process simpler, for both PyTorch and Tensorflow implementations.

Embedded Programming

Share WiFi through Ethernet Port with Raspberry Pi

6 minute read

Published:

Imagine your WiFi onboard your laptop broke down and you cannot use the internet even though the router and modem works fine. Or maybe you have a desktop computer without a WiFi adapter or dongle. How can you hack through this situation without having to invest in a dongle in the quickest possible way?

PyTorch

Dimension tracking for Conv2D layer

5 minute read

Published:

Keeping track of the dimensions is a very cumbersome process if not properly understood, specially while designing the deep neural network model architectures. The stress adds up swiftly when we have to shift between different machine learning frameworks, ideally PyTorch and Tensorflow. It is crazy how although both these frameworks are essentially doing the same thing, the syntax makes it increasingly difficult to allow for fluid transitions between them. Hence, this post is written to make the dimension tracking process simpler, for both PyTorch and Tensorflow implementations.

Raspberry Pi

Share WiFi through Ethernet Port with Raspberry Pi

6 minute read

Published:

Imagine your WiFi onboard your laptop broke down and you cannot use the internet even though the router and modem works fine. Or maybe you have a desktop computer without a WiFi adapter or dongle. How can you hack through this situation without having to invest in a dongle in the quickest possible way?

Tensorflow

Dimension tracking for Conv2D layer

5 minute read

Published:

Keeping track of the dimensions is a very cumbersome process if not properly understood, specially while designing the deep neural network model architectures. The stress adds up swiftly when we have to shift between different machine learning frameworks, ideally PyTorch and Tensorflow. It is crazy how although both these frameworks are essentially doing the same thing, the syntax makes it increasingly difficult to allow for fluid transitions between them. Hence, this post is written to make the dimension tracking process simpler, for both PyTorch and Tensorflow implementations.