Change Log

0.1.9 (draft)

  • New Features:
    • segmind ls for List files/folders in Datastore.
    • segmind delete for Delete files/folders in Datastore.


  • Improvements & Bug fixes
    • Persisting User Tokens.
    • Added an alternative way for setting segmind.config_nb on any external (non-Segmind) Notebook
    • segmind upload improved for folder uploads
    • segmind upload fix progress bar for nested-folders upload
    • Bump protobuf from 3.13.0 to 3.15.0
    • Bump psutil >= 5.7
  • New Features:
    • segmind sync improved way to sync BIG folders - skips already synced.


  • Token Based Authentication, where the api-token (which is generated from Segmind UI) is replaced by the password. (while doing segmind config)
  • MMDetection Integration:
    • Importing this alone will do the job
      • segmind.mmdetection.init_segmind_hook function
  • Add Custom Tags to a Run. You can add below Tags to the Run: (These will show in the UI)
    • Run Name
    • Algo Name
  • Multi-User Runs can be now recognized via UI


  • segming.data.upload functionality: Uploads a file/folder to your Segmind datastore.
    • Via CLI (Local Computer or Segmind Jupyter Terminal)
      • segmind upload --path <file/folder path> --destination_path <path in datastore> --datastore_name <your datastore name>
    • Via Segmind Notebook
      • from segmind.data import uplaod
      • upload(path="", destination_path="", datastore_name="")
  • Security Updates + Code Refactoring


  • Added Autologging for PyTorch & PyTorch-Lightning. Callback lets user autolog pytorch and Lightning runs inlcuding metrics, params and model summary artifact.
  • Logging only primary metrics. Removed Temp. RAM and other secondry metrics, and GPU metrics. The metrics might have the the best of all the project data.
  • Udpated API endpoints with a more consistent pattern.


28 Sept, 2021

  1. Added Autologging for Keras. Callback lets user autolog keras runs inlcuding metrics, params and model summary artifact.
  2. Removed requirements.txt. This helps install TF, PyTorch and other frameworks over current installs.

Updated 07 Mar 2022
Did this page help?