Hugging Face's logo

Ambarella
/
SuperPoint

PyTorch
llama
Model card Files
xet
 Community
cooper_robot commited on
Commit
b2210dd
·
1 Parent(s): 4a19780

Add release note for v1.3.1

Browse files
Files changed (4) hide show
  1. README.md +33 -32
  2. config.json +3 -0
  3. x7_superpoint_act16.bin +3 -0
  4. x7_superpoint_amba_optimized.bin +3 -0
README.md CHANGED
@@ -1,32 +1,33 @@
1
- ---
2
- library_name: pytorch
3
- ---
4
-
5
- ![superpoint_logo](resource/Superpoint.png)
6
-
7
- SuperPoint is a self-supervised feature detection and description network that jointly learns interest point locations and corresponding descriptors for robust image matching.
8
-
9
- Original paper: [SuperPoint: Self-Supervised Interest Point Detection and Description](https://arxiv.org/abs/1712.07629)
10
-
11
- # SuperPoint
12
-
13
- This model uses the SuperPoint feature extractor, which detects repeatable keypoints and generates compact descriptors that are robust to viewpoint, scale, and illumination changes. It is well suited for applications such as visual localization, SLAM, structure-from-motion, image stitching, and feature-based tracking in robotics and augmented reality systems.
14
-
15
- Model Configuration:
16
- - Reference implementation: [Official SuperPoint source code](https://github.com/rpautrat/SuperPoint)
17
- - Original Weight: [pytorch-superpoint pretrained checkpoints](https://github.com/eric-yyjau/pytorch-superpoint#pretrained-models) (Model Garden training/export flow)
18
- - Resolution: 1x240x320 (grayscale uint8 input)
19
- - Support Cooper version:
20
- - Cooper SDK: [2.5.4]
21
- - Cooper Foundry: [2.3]
22
-
23
- | Model | Device | Compression | Model Link |
24
- | :-----: | :-----: | :-----: | ------- |
25
- | SuperPoint | N1-655 | Amba_optimized | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/n1-655_superpoint_amba_optimized.bin) |
26
- | SuperPoint | N1-655 | Activation_fp16 | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/n1-655_superpoint_act16.bin) |
27
- | SuperPoint | CV7 | Amba_optimized | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv7_superpoint_amba_optimized.bin) |
28
- | SuperPoint | CV7 | Activation_fp16 | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv7_superpoint_act16.bin) |
29
- | SuperPoint | CV72 | Amba_optimized | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv72_superpoint_amba_optimized.bin) |
30
- | SuperPoint | CV72 | Activation_fp16 | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv72_superpoint_act16.bin) |
31
- | SuperPoint | CV75 | Amba_optimized | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv75_superpoint_amba_optimized.bin) |
32
- | SuperPoint | CV75 | Activation_fp16 | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv75_superpoint_act16.bin) |
 
 
1
+ ---
2
+ library_name: pytorch
3
+ ---
4
+
5
+ ![superpoint_logo](resource/Superpoint.png)
6
+
7
+ SuperPoint is a self-supervised feature detection and description network that jointly learns interest point locations and corresponding descriptors for robust image matching.
8
+
9
+ Original paper: [SuperPoint: Self-Supervised Interest Point Detection and Description](https://arxiv.org/abs/1712.07629)
10
+
11
+ # SuperPoint
12
+
13
+ This model uses the SuperPoint feature extractor, which detects repeatable keypoints and generates compact descriptors that are robust to viewpoint, scale, and illumination changes. It is well suited for applications such as visual localization, SLAM, structure-from-motion, image stitching, and feature-based tracking in robotics and augmented reality systems.
14
+
15
+ Model Configuration:
16
+ - Reference implementation: [Official SuperPoint source code](https://github.com/rpautrat/SuperPoint)
17
+ - Original Weight: [pytorch-superpoint pretrained checkpoints](https://github.com/eric-yyjau/pytorch-superpoint#pretrained-models) (Model Garden training/export flow)
18
+ - Dataset: [HPatches](https://github.com/hpatches/hpatches-dataset)
19
+ - Resolution: 1x240x320 (grayscale uint8 input)
20
+ - Support Cooper version:
21
+ - Cooper SDK: [2.5.4]
22
+ - Cooper Foundry: [2.3]
23
+
24
+ | Model | Device | Compression | Model Link |
25
+ | :-----: | :-----: | :-----: | ------- |
26
+ | SuperPoint | N1-655 | Amba_optimized | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/n1-655_superpoint_amba_optimized.bin) |
27
+ | SuperPoint | N1-655 | Activation_fp16 | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/n1-655_superpoint_act16.bin) |
28
+ | SuperPoint | CV7 | Amba_optimized | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv7_superpoint_amba_optimized.bin) |
29
+ | SuperPoint | CV7 | Activation_fp16 | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv7_superpoint_act16.bin) |
30
+ | SuperPoint | CV72 | Amba_optimized | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv72_superpoint_amba_optimized.bin) |
31
+ | SuperPoint | CV72 | Activation_fp16 | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv72_superpoint_act16.bin) |
32
+ | SuperPoint | CV75 | Amba_optimized | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv75_superpoint_amba_optimized.bin) |
33
+ | SuperPoint | CV75 | Activation_fp16 | [Model_Link](https://huggingface.co/Ambarella/SuperPoint/blob/main/cv75_superpoint_act16.bin) |
config.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:779942ebb16708036af3626823520ad296c3b78c42d5a1dababbdfc2dc496a09
3
+ size 28
x7_superpoint_act16.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3007d44225608709d194a3f0a1041c0f7a29eaaed3c091629634e810b3fb179
3
+ size 687367
x7_superpoint_amba_optimized.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9619c81d6eaefe4db82c6bb165c8e44b1bafeb38ec225f9ca651a8e8b228422f
3
+ size 683319