https://github.com/martinthoma/msthesis-experiments

Experiments for my masters thesis
https://github.com/martinthoma/msthesis-experiments

tex thesis

Last synced: 2 days ago
JSON representation

Experiments for my masters thesis

• Host: GitHub
• URL: https://github.com/martinthoma/msthesis-experiments
• Owner: MartinThoma
• Created: 2017-03-03T11:40:26.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2020-06-27T02:29:42.000Z (over 4 years ago)
• Last Synced: 2024-12-26T07:41:52.619Z (8 days ago)
• Topics: tex, thesis
• Language: Python
• Size: 645 KB
• Stars: 16
• Watchers: 5
• Forks: 4
• Open Issues: 3
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# msthesis-experiments

Experiments for the masters thesis of Martin Thoma

Every experiment is a YAML file. See the experiments folder for some examples.

## Requirements

* matplotlib

* Tensorflow 1.0

* Keras 2.0 (adjusted [preprocessing/image.py](https://github.com/fchollet/keras/pull/6003), see misc directory)

* seaborn

If you get `TypeError: __init__() got an unexpected keyword argument

'hsv_augmentation'` you didn't adjust the `image.py`. Just copy the one in

the misc folder to `python -c "import keras.preprocessing.image as k;print(k.__file__)"`

## Scripts

* `./run_training.py -f experiments/cifar100_baseline.yaml`: Train a model. Downloads everything by its own

* `./analyze_training.py -d artifacts/cifar100_baseline`: Show some training statistics

* `./inference_timing.py -f experiments/cifar100_baseline.yaml`: Run inference on a given trained model and measure the time

* `./eval_ensemble.py -f ensemble/cifar100_baseline.yaml`: Evaluate an ensemble

* `./visualize.py --cm artifacts/cifar100_root/cm-test.json`: Confusion matrix optimization

* `./create_cm.py --indices cm.indices.pickle -f experiments/cifar100_root-g5.yaml`

## Run timining experiments

Measuring inference time needs about 2 minutes:

```

$ ./inference_timing.py -f experiments/cifar10_baseline.yaml

```

Measuring training time takes about 70 minutes:

```

$ ./run_training.py -f experiments/cifar10_baseline.yaml

```