Organize machine learning experiments and monitor training progress from mobile.
Screenshots of Web App

Integrate with just 2 lines

                from numpy.random import random
from labml import tracker, experiment as exp

conf = {'batch_size': 20}

def train(n:int):
    loss = 0.999 ** n + random() / 10
    accuracy = 1 - 0.999 ** n + random() / 10
    return loss, accuracy

with exp.record(name='sample', exp_conf=conf):
    for i in range(100000):
        lss, acc = train(i)
        tracker.save(i, loss=lss, accuracy=acc)
                
            
Open In Colab
                from labml import experiment as exp
from labml.utils.lightning import LabMLLightningLogger 

trainer = pl.Trainer(gpus=1,
                     max_epochs=5,
                     progress_bar_refresh_rate=20,
                     logger=LabMLLightningLogger())

with exp.record(name='sample', exp_conf=conf):
        trainer.fit(model, data_loader)
                
            
Open In Colab
                from labml import experiment as exp
from labml.utils.keras import LabMLKerasCallback

with exp.record(name='sample', exp_conf=conf):
    for i in range(50):
        model.fit(x_train,
                  y_train,
                  epochs=conf['epochs'],
                  validation_data=(x_test, y_test),
                  callbacks=[LabMLKerasCallback()],
                  verbose=None)
                
            
Open In Colab

Features

Monitor from Mobile
Easily monitor your experiments from a mobile first web app
Open Source
We are on Github and anyone can contribute
Track commits, configs & hyper-parameters
Keep track of all experiment information, including git commits, configurations and hyper-parameters
Custom Visualizations
Easy to use API to generate your own custom visualizations
Pretty Logs
Logs for the progress and any other outputs
Save & Load Checkpoints
Easily save & load checkpoints to help you with fine tuning your ML Model

Join our slack

Reach us via slack to discuss and with the developers and community