Monitor your ML runs live wherever you are

Would you like to know that your model training failed, the loss stopped improving, or the GPU consumption is going crazy? You can. Run experiments anywhere, log anything you want, and see it all in one place.

It really is free and takes 5 min to setup

Start collaborating for FREE

It is a good practice to have control over your model training process

And hundreds of data scientists are using Neptune to see that their experiments are running smoothly. 

“Without the information I have in the Monitoring section I wouldn’t know that my experiments are running 10 times slower than they could.
All of my experiments are being trained on separate machines which I can access only via ssh. If I would need to download and check all of this separately I would be rather discouraged :).”

Michał Kordas

Machine Learning Researcher @TensorCell

Is this honestly the best way to organize my ML experiments?

SSH + console logs

  • You don’t have to ssh to see everything you need
  • Your hardware consumption is logged automatically -> we run nvidia-smi and psutil every second for you
  • You can log things like images, html visualizations, model checkpoints, and more
  • We actually log the stderr and stdout from your console anyway so you can check it out if you want to

Open-source solutions like MLflow, TensorBoard or Sacred

  • You don’t have to set up and maintain the database/filesystem and UI yourself
  • You can monitor things like hardware consumption, html visualizations, image sequences, code snapshots and more
  • You can share your running experiments with your team by sending a link
  • We integrate with all of those tools so you can actually use both at the same time!

It really is free and takes 5 min to setup

Start collaborating for FREE

Monitor your ML models in 3 steps

Add a few lines to your scripts

Connect Neptune to your project by adding literally 3 lines on top of your scripts. Then you just run your training and evaluation and log whatever you care about.

For most machine learning frameworks you don’t even have to write those logging calls -> We created the integrations for you!

import neptune

neptune.init('Me/MyProject')
neptune.create_experiment(params={'lr':0.1, 'dropout':0.4})

# training and evaluation logic
neptune.log_metric('test_accuracy', 0.84)
neptune.log_image('model predictions', image)

Run your experiments the way you usually do

Neptune goes where you work not the other way around.

So if you are running experiments on your laptop, spin up cloud machines, or burn through computational clusters at your university Neptune will keep track of your experiments with no problems.

It works with Colab, Kaggle kernels, and integrates nicely with Jupyter Notebooks.

Before

python main.py

After

python main_with_neptune_lines.py

Monitor your experiments in the app

See your learning curves, hardware consumption, console logs as your model is running.

If you log ROC curves, image predictions, model checkpoints, or other things after every iteration you can scroll through them and see the progress live.

It really is free and takes 5 min to setup

Start collaborating for FREE

Start collaborating on experiments in minutes with our integrations

Are you thinking “Ok but, do I have to write the logging/callback functions myself?”

If you are using Keras, XGBoost, Optuna, or one of the 20+ libraries that we integrate with you don’t need to implement anything to monitor your experiments.

What our users say

Over 5,000 ML people started monitoring their experiments with Neptune this year – read what some of them have to say:

“If you need to monitor and manage your machine learning or any other computational experiments, Neptune.ai is a great choice. It has many features that can make your life easier and your research more organized.”

Boaz Shvartzman

Computer vision researcher and developer @TheWolf

“I’m working with deep learning (music information processing), previously I was using Tensorboard to track losses and metrics in TensorFlow, but now I switched to PyTorch so I was looking for alternatives and I found Neptune a bit easier to use, I like the fact that I don’t need to (re)start my own server all the time and also the logging of GPU memory etc. is nice. So far I didn’t have the need to share the results with anyone, but I may in the future, so that will be nice as well.”

Ondřej Cífka

PhD student in Music Information Processing at Télécom Paris

“Without the information I have in the Monitoring section I wouldn’t know that my experiments are running 10 times slower than they could.
All of my experiments are being trained on separate machines which I can access only via ssh. If I would need to download and check all of this separately I would be rather discouraged :).”

Michał Kordas

Machine Learning Researcher @TensorCell

They already have their ML experimentation in order.
When will you?

✓ Sign up for a free account
✓ Add a few lines to you code
✓ Get back to running your experiments

Start tracking for FREE