ALsats is a project that aims to reduce the investment needed to create minimum viable datasets for supervised machine learning. The aim is to allow data scientists to intelligently label their datasets without subscriptions, lock-ins, proprietary downloads, log-ins/passwords or any proprietary data or model storage. The data scientist should be allowed to focus on what they need - labels for data - and pay strictly for the compute that is consumed. No more and no less.
The current proposal is a follow-up from rounds 4 and 6.
The funding in round six was tied to two main deliverables:
- Label Checkpointing
- Model Checkpointing
Both deliverables were met, as outlined in the linked video.
(NB: If you are a Algovera associated data scientist looking forward use ALsats to label image data, please contact me directly on Discord or DM me on Twitter. Iām presently beta testing out the infrastructure myself.)
Round 9 proposal:
The round 9 proposal focuses on issues identified in feedback received from data scientists who used ALsats:
- Ability to configure hyperparameters of the learning algorithm - Giving users the ability to configure model parameters while training.
- Label Studio / Labeling App Interface - Explore creation of a back-end interface to standard/popular annotation app front-ends such as Label Studio/ CVAT/ LabelImg etc.
- Exploring payment interfaces to Strike (USD payments), USDC on Eth/Polygon/Algorand.
How round 9 allotments will be used if funded:
Round 6 deliverables allowed upgrading of the ALsats back-end to S3 usage, greater RAM allocation. Round 9 deliverables will be used to fund application development, subscriptions to annotation front-ends and any ongoing expenses that may arise during development.
I look forward to receiving your support in round 9.
Squad Members:
antaraxia/antaraxia.eth: Algovera member.
Twitter Handle: https://twitter.com/antaraxia_kk