Skip to main content

Client library for working with NimbleBox LMAO

Project description

Dainik

Dainik is a client library for NimbleBox.ai LMAO (Logging, Monitoring, Alerts and Observability) service.

Usage

from dainik import Dainik

# initialise the client object
dk = Dainik()

# initialise the run
dk.init(
  "california",
  config = {
    "model_name": "qual",
    "config": {
      "batch_size": 32,
      "n_steps": n_steps,
      "optimizer": "adam",
    }
  },
)

for epoch in range(10):
  log = {
    "loss": 1 / epoch,
    "accuracy": epoch,
  }
  dk.log(log, step = epoch)

  # optionally log metrics directly
  dk.metrics.f1_score(
    x = [[1, 2, 3]],
    y = [[1, 2, 1]],
    step = epoch
  )

  # log files for tracking as well, wildcards work as well!
  dk.save_files([f"./checkpoint-{epoch}/*"])

dk.end() # send signal to sleep

Dev

Due to the current structure the source code for this is only available for NimbleBox.ai engineers. This is spun out of a much larger monorepo which contains the logic for server and client together which helps us work on this faster.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

dainik-0.13.tar.gz (13.5 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page