A keras extension (callback) to record details of a training task
Project description
Keras Experiment Recorder
This is a set of utilities (mostly a keras Callback) that help store various aspects of an experiment (e.g. model information, results of training etc) in firebase backend.
Prerequisites
You will need the service account of a firebase app.
Usage
from keras_experiment_recorder import ExperimentInfo, FirebaseInfo
from keras_experiment_recorder import ExperimentRecorder
# below three lines setup the instance of ExperimentRecorder
fb_info = FirebaseInfo(bucket_name=bucket_name,
credential_path=credential_path)
experiment_info = ExperimentInfo(name='first exp', fb_info=fb_info)
recorder = ExperimentRecorder(experiment_info)
callbacks = [recorder]
history = model.fit(x_train, y_train,
batch_size=batch_size,
epochs=epochs,
verbose=1,
callbacks=callbacks, # pass the callbacks
validation_split=0.1)
Running the example included in this repository
python examples/simple.py --bucket-name keras-experiment-recorder.appspot.com --credential-path ~/Desktop/Dev/keys/keras-experimen
t-recorder-firebase.json
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Close
Hashes for Keras_Experiment_Recorder-0.1.1.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | d7ce4c025357faf5b5ad4e6e747d928c3fde5d5956ccbee53127aa2a64abe6a9 |
|
MD5 | 32ecac663439248c459ef6b600fc6a47 |
|
BLAKE2b-256 | 04b93f795bf5ef68615d45e99a2e44d4cae7253ab4ded4d5d086f5a3f8a96f17 |
Close
Hashes for Keras_Experiment_Recorder-0.1.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | ef67c21243ad5d25eb1a7dbe2f2f273ed378b605b546da9e704750d35b21ff47 |
|
MD5 | 3b2638b280571c12ab5c4ea4d45a8ff6 |
|
BLAKE2b-256 | 5be7e2bcf33e1cd2d38622c3fc97f844a0a5ac1e5281378962a437d50c1c2724 |