Published December 31, 2023 | Version 1.0.0
Model Open

Omnidata Pretrained Models

Authors/Creators

  • 1. ROR icon Meta (United States)
  • 2. ROR icon École Polytechnique Fédérale de Lausanne
  • 3. University of California Berkeley

Description

These contain pretrained models for monocular depth and surface normal estimation. The models were trained on the Omnidata dataset -- information here: https://github.com/EPFL-VILAB/omnidata

Files

Files (6.6 GB)

Name Size
md5:2d0a2a4507ff1ca74eab809dd2d32b2d
1.5 GB Download
md5:1071266b8bfdb58cdcb490d6fcddedb0
1.9 GB Download
md5:f2e0962ea6e248a1648f20d74b1b034d
1.9 GB Download
md5:2c0546bc68dff5c6dd013da98e54c0e6
1.2 GB Download