Upload NOTICE with huggingface_hub
Browse files
NOTICE
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
mlx-swift-AnyCalib
|
| 2 |
+
Copyright 2026 Hiroaki Yamane
|
| 3 |
+
|
| 4 |
+
This product includes software developed as a port of AnyCalib.
|
| 5 |
+
|
| 6 |
+
--------------------------------------------------------------------------------
|
| 7 |
+
AnyCalib
|
| 8 |
+
https://github.com/javrtg/AnyCalib
|
| 9 |
+
Copyright (c) Javier Tirado-Garín and Javier Civera, I3A, University of Zaragoza
|
| 10 |
+
Licensed under the Apache License, Version 2.0.
|
| 11 |
+
|
| 12 |
+
"AnyCalib: On-Manifold Learning for Model-Agnostic Single-View Camera
|
| 13 |
+
Calibration", ICCV 2025.
|
| 14 |
+
|
| 15 |
+
The model architecture, camera models, on-manifold optimization, and the
|
| 16 |
+
pretrained weights redistributed by this project all originate there. The
|
| 17 |
+
Swift/MLX sources in Sources/MLXAnyCalib are a translation of that PyTorch
|
| 18 |
+
implementation.
|
| 19 |
+
|
| 20 |
+
--------------------------------------------------------------------------------
|
| 21 |
+
DINOv2
|
| 22 |
+
https://github.com/facebookresearch/dinov2
|
| 23 |
+
Copyright (c) Meta Platforms, Inc. and affiliates.
|
| 24 |
+
Licensed under the Apache License, Version 2.0.
|
| 25 |
+
|
| 26 |
+
The backbone (Sources/MLXAnyCalib/DINOv2.swift) is a translation of DINOv2's
|
| 27 |
+
vision transformer, vendored into AnyCalib and fine-tuned by its authors.
|
| 28 |
+
|
| 29 |
+
--------------------------------------------------------------------------------
|
| 30 |
+
Modifications made by this project
|
| 31 |
+
* Translated from PyTorch to Swift/MLX.
|
| 32 |
+
* Checkpoint weights are converted from PyTorch .pt to safetensors, with
|
| 33 |
+
convolution tensors transposed from NCHW to NHWC for MLX's native layout.
|
| 34 |
+
No weight values are altered beyond an optional float32 -> float16 cast.
|
| 35 |
+
* Deviations from the reference implementation's numerics are enumerated in
|
| 36 |
+
the "Deliberate deviations" section of README.md.
|