neptune.ai

@neptune_ai

We tweet about best practices and some other cool stuff. Read our blog at Try our free experiment tracking tool!

Warsaw, Poland
neptune.ai
Vrijeme pridruživanja: siječanj 2018.
1.271 fotografija ili videozapis Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @neptune_ai

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @neptune_ai

  1. prije 12 sati

    Monitor ROC curve in : 1. Create a callback that takes model and data 2. in on_epoch_end method: - run prediction on data - create ROC curve figure - save figure somewhere Other tricks in this article:

    1 proslijeđeni tweet
    Poništi
  2. prije 15 sati

    "Exploratory Data Analysis for NLP: A Complete Guide to Python Tools" New article on our blog by . Comes with code snippets and "EDA for NLP template" notebook. Great read!

    3 proslijeđena tweeta 2 korisnika označavaju da im se sviđa
    Poništi
  3. prije 20 sati

    Neptune logging was added to (lightweight wrapper) and it's so easy to use: trainer = Trainer(logger=NeptuneLogger(...)) Check how to use it:

    1 korisnik označava da mu se sviđa
    Poništi
  4. prije 22 sata

    trick 3 Explore text data via Topic Modeling: - run topic modeling (LDA) - visualize topics by showing word frequencies per topic You can use pyLDAvis to do it interactively. More tricks in this post:

    1 proslijeđeni tweet 4 korisnika označavaju da im se sviđa
    Poništi
  5. 4. velj

    TextBlob - good and easy to use lib for sentiment analysis from Check it out here: Read about other useful tools in this post:

    1 proslijeđeni tweet 1 korisnik označava da mu se sviđa
    Poništi
  6. 4. velj

    trick 1 Visualizing text statistics like: - word length frequencies - stopword frequencies - document length - etc. with histograms and bar charts is a simple yet powerful technique. More trick in this post:

    3 proslijeđena tweeta 1 korisnik označava da mu se sviđa
    Poništi
  7. 3. velj

    "Keras metrics: Everything You Need to Know" New article on our blog by ! Talks about in-build and custom metrics for and keras. + callbacks for logging ROC and more.

    1 proslijeđeni tweet 3 korisnika označavaju da im se sviđa
    Poništi
  8. 3. velj

    You wish you could run experiments in notebooks and auto-snapshot .ipynb changes? We made it happen: - install the extension - create an experiment with neptune.create_experiment - check your notebook snapshot!

    1 proslijeđeni tweet 3 korisnika označavaju da im se sviđa
    Poništi
  9. 1. velj

    Neptune integrates with Optuna from ! It is simple. 1. Add a callback: study.optimize(objective, n_trials=100, callbacks=[opt_utils.NeptuneMonitor()]) 2. Monitor your search results. 3. Done.

    1 proslijeđeni tweet 4 korisnika označavaju da im se sviđa
    Poništi
  10. 1. velj

    Neptune logging was added to Ignite (lib that makes writing compact training loops easy)! It's simple: npt_logger = NeptuneLogger() npt_logger.attach(trainer, log_handler=OutputHandler()) trainer. run() That's it! Read more:

    Poništi
  11. 31. sij

    A nicely structured collection of build-it-from-scratch notebooks that really go in-depth into many concepts. If you want to understand a particular concept check this project out!

    4 proslijeđena tweeta 1 korisnik označava da mu se sviđa
    Poništi
  12. 31. sij

    create custom metrics in keras 1. Inherit from tf.keras.metrics.Metric 2. Override methods: - update_state - result - reset_states 3. pass to .compile() Other metrics tricks in this article:

    3 proslijeđena tweeta 1 korisnik označava da mu se sviđa
    Poništi
  13. 31. sij

    trick 8 Use text complexity: - use lib like textstat to extract text readability index - create a histogram of complexity scores - find the most difficult documents More tricks in this post:

    1 proslijeđeni tweet 2 korisnika označavaju da im se sviđa
    Poništi
  14. 31. sij

    Working on notebooks together can be tricky. Being able to track and share your checkpoints can make things a bit easier. Check our docs

    3 proslijeđena tweeta
    Poništi
  15. 30. sij

    textstat - easy to use tool for calculating text complexity (readability indexes) from Check it out here: Read about other useful tools in this post:

    2 proslijeđena tweeta 5 korisnika označava da im se sviđa
    Poništi
  16. 30. sij

    Neptune logging was added to , a tool built on top of that simplifies and model training. Really easy to set up: runner = SupervisedNeptuneRunner() runner.train(...) That's it! Read more:

    8 korisnika označava da im se sviđa
    Poništi
  17. 30. sij

    "Keras metrics: Everything You Need to Know" New article on our blog by Talks about in-build and custom metrics for and keras. + callbacks for logging ROC and more.

    2 proslijeđena tweeta 1 korisnik označava da mu se sviđa
    Poništi
  18. 29. sij

    It is rare to have people among your investors who understand your niche so well. Thank you so much and for believing in us! Also, the reasoning behind your decision is so nicely put. Highly recommended read.

    Here is a longer post on why we invested in . Data science is an explosively expanding market that is still immature in terms of its tool set. That's the problem Neptune attacks in a simple and elegant way.
    1 proslijeđeni tweet 3 korisnika označavaju da im se sviđa
    Poništi
  19. 28. sij

    This is such an awesome idea! Tracking experiments is even easier with this.

    Published fast_neptune, a Python package to make experiments reproducible, using ! By simply adding or to cells, codes and variables can be recorded ! Thanks for nbdev ! Article here or
    1 korisnik označava da mu se sviđa
    Poništi
  20. proslijedio/la je Tweet
    28. sij

    If you're interested in becoming involved in my new project dabl , I tagged some easy first issues. Given that it's pretty early in the development, the barrier to entry should hopefully be much lower than in sklearn for example:

    10 replies 90 proslijeđenih tweetova 364 korisnika označavaju da im se sviđa
    Poništi

@neptune_ai još nije objavio nijedan Tweet.

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·