software-mansion/react-native-executorch-yolo26-pose

yolo26-pose

This repository hosts the yolo26-pose models exported for the React Native ExecuTorch library as ExecuTorch .pte programs, ready to run on device.

Upstream model: Yolo26

Variants

Path	Backend	Precision
384/vulkan/yolo26n_pose_384_vulkan_fp16.pte	vulkan	fp16
512/vulkan/yolo26n_pose_512_vulkan_fp16.pte	vulkan	fp16
640/vulkan/yolo26n_pose_640_vulkan_fp16.pte	vulkan	fp16
coreml/yolo26n_pose_384_coreml_fp16.pte	coreml	fp16
coreml/yolo26n_pose_512_coreml_fp16.pte	coreml	fp16
coreml/yolo26n_pose_640_coreml_fp16.pte	coreml	fp16
xnnpack/yolo26n_pose_384_xnnpack_fp32.pte	xnnpack	fp32
xnnpack/yolo26n_pose_384_xnnpack_int8.pte	xnnpack	int8
xnnpack/yolo26n_pose_512_xnnpack_fp32.pte	xnnpack	fp32
xnnpack/yolo26n_pose_512_xnnpack_int8.pte	xnnpack	int8
xnnpack/yolo26n_pose_640_xnnpack_fp32.pte	xnnpack	fp32
xnnpack/yolo26n_pose_640_xnnpack_int8.pte	xnnpack	int8

XNNPACK int8

The xnnpack/*_int8.pte files are static int8 (PT2E, per-channel weights) with the first convolution, the neck outputs (layers 16, 19, 22) and the Pose head kept fp32. COCO keypoint (OKS) mAP50-95 on the first 500 val2017 images against the fp32 file:

	384	512	640
fp32	46.08	52.53	56.55
int8	45.32	51.82	56.13

Unshipped backends

A backend directory carrying a NOTES.md and no .pte was exported and not published. Its note records what went wrong.

• mlx (why)

Repository structure

384/vulkan/config.json                       1.1 kB
384/vulkan/yolo26n_pose_384_vulkan_fp16.pte  5.8 MB
512/vulkan/config.json                       1.1 kB
512/vulkan/yolo26n_pose_512_vulkan_fp16.pte  5.8 MB
640/vulkan/config.json                       1.1 kB
640/vulkan/yolo26n_pose_640_vulkan_fp16.pte  5.8 MB
config.json                                  33 B
coreml/config.json                           2.7 kB
coreml/yolo26n_pose_384_coreml_fp16.pte      6.5 MB
coreml/yolo26n_pose_512_coreml_fp16.pte      6.5 MB
coreml/yolo26n_pose_640_coreml_fp16.pte      6.5 MB
mlx/NOTES.md                                 2.7 kB
xnnpack/config.json                          5.2 kB
xnnpack/yolo26n_pose_384_xnnpack_fp32.pte    11.4 MB
xnnpack/yolo26n_pose_384_xnnpack_int8.pte    6.7 MB
xnnpack/yolo26n_pose_512_xnnpack_fp32.pte    11.4 MB
xnnpack/yolo26n_pose_512_xnnpack_int8.pte    6.7 MB
xnnpack/yolo26n_pose_640_xnnpack_fp32.pte    11.4 MB
xnnpack/yolo26n_pose_640_xnnpack_int8.pte    6.7 MB

Compatibility

These files are published for the ExecuTorch v1.4.1 runtime. ExecuTorch gives no forward compatibility guarantee, so an older runtime may fail to load them.

To use them in React Native ExecuTorch, pass the model constant shipped in the library's model registry to the corresponding task pipeline. See the documentation.

To load these files in your own ExecuTorch runtime, read the compatibility note first.