Notebooks
W
Weights and Biases Semantic Segmentation Demo With W&B

Semantic Segmentation Demo With W&B

wandb-examplescolabsfastai
alph-notebooks/wandb-examples / Semantic_Segmentation_Demo_with_W&B.ipynb

Open In Colab

❌ This Notebooks is using Fastai V1 that is no longer available on colab

Semantic Segmentation of Driving Scenes

This demo shows how to log image masks for a semantic segmentation model. This model version is simpler and smaller, and the training is shorter for demo purposes. Please see this W&B report for the more detailed model and this GithHub repository by Boris Dayma for the best reference code. demoshot. Note that everyone's runs from this demo are logged to a shared Weights & Biases project page by default.

Resources

About

Weights & Biases helps you log experiments, visualize and analyze them faster, collaborate with others, and share your findings. Here we use Google Colab as a convenient hosted environment, but you can run your own training scripts from any local or cloud setup with W&B.

Setup

[ ]
[ ]

W&B – Login to your wandb account so you can log all your metrics

[ ]
[ ]

Segmentation labels extracted from dataset source code

[ ]

Data paths

[ ]

Set low for faster iteration

[ ]

Associate a label to an input

[ ]
[ ]
[ ]
[ ]
[ ]

How to See Live Results in Shared Project

  1. Check out the project page to see your results in the shared project.
  2. Press 'option+space' to expand the runs table, comparing all the results from everyone who has tried this script.
  3. Click on the name of a run to dive deeper into that specific run on its own run page.

project page

Visualize Relationships

Use a parallel coordinates chart to see the relationship between hyperparameters and output metrics. Here, we can see how the learning rate and other metrics saved in "config" affect loss and accuracy.

parallel coordinates plot

More about Weights & Biases

We're always free for academics and open source projects. Email carey@wandb.com with any questions or feature suggestions. Here are some more resources:

  1. Documentation - Python docs
  2. Gallery - example reports in W&B
  3. Articles - blog posts and tutorials
  4. Community - join our Slack community forum