CoTracker / CoTracker3
Source: https://github.com/facebookresearch/co-tracker
Structured fields
{
"archived": false,
"checked_at": "2026-10-17T00:00:00.000Z",
"usefulness": "Transformer point tracker that follows any pixel or quasi-dense grids through video (CoTracker3 pseudo-labelling). Basis for 2D track data used in stabilization, patch pinning and motion transfer.",
"source_type": "repository",
"availability": "available",
"source_owner": "Meta AI / Oxford VGG",
"next_check_at": "2026-10-31T00:00:00.000Z",
"hardware_notes": "GPU strongly recommended per README.",
"upstream_updated_at": "2026-03-03T21:51:22.000Z"
}Evidence claims
- [upstream_claim] README: CoTracker is a fast transformer-based model that can track any pixel, or a quasi-dense set, selected manually or sampled on a grid; GPU strongly recommended, CPU possible for small tasks. · locator: README, intro and installation source
- [sourced_fact] README License: majority of CoTracker is CC-BY-NC; portions under separate terms (Particle Video Revisited MIT; TAP-Vid and LocoTrack Apache 2.0). · locator: README, License source
- [sourced_fact] LICENSE.md is the Creative Commons Attribution-NonCommercial 4.0 International license text. · locator: LICENSE.md, header source
- [upstream_claim] Research review verdict at export: reviewed (as of 2026-10-17). Recorded by the research pipeline, not a hub review outcome; evidence.review stays unreviewed until an independent reviewer records one.
Rights
| component | license | hosting | redistribution | modification | training | Evidence and notes |
|---|---|---|---|---|---|---|
| code | CC-BY-NC-4.0 | restricted | restricted | restricted | unknown | Non-commercial; commercial integration requires separate permission. Mixed components noted in README. |
| model_weights | unknown | unknown | unknown | unknown | unknown | Checkpoint terms not separately read; README says majority of project is CC-BY-NC. |