TL;DR

Segment ‘Anything’ using FAIR’s SAM with an interactive run on VESSL.

Description

The Segment Anything (SA) project introduces a task, model, and dataset for image segmentation, including over 1 billion masks on 11 million images. Their promptable model demonstrates impressive zero-shot performance, rivaling or surpassing prior fully supervised methods, and they are releasing the Segment Anything Model (SAM) and dataset (SA-1B) to foster research in computer vision.

YAML

name : segment-anything
description: "Segment ‘Anything’ using FAIR’s SAM with an interactive run on VESSL."
resources:
  cluster: aws-apne2
  preset: v1.v100-1.mem-52
image: nvcr.io/nvidia/pytorch:21.05-py3
run:
  - workdir: /root/examples/segment-anything/
    command: |
      bash ./setup.sh
import:
  /root/segment-anything: git://github.com/vessl-ai/examples
interactive:
  max_runtime: 24h
  jupyter:
    idle_timeout: 120m
ports:
  - name: streamlit
    type: http
    port: 8501

Input

Output