Sponsored by
Vmake Labs Video Enhancer
-Vmake Labs Video Enhancer uses AI to improve video quality, restore details, reduce blur, and upscale videos to 8K.
Input
: [[1, 3, 256, 256]] — RGB, ImageNet norm, 256x256 crop around a face (detect first, then crop and resize to this aspect)
Output
: SimCC pair: x [1,106,512] and y [1,106,512] — a 1-D distribution per keypoint per axis. Decode: keypoint k sits at (argmax(x[k]) / 2, argmax(y[k]) / 2) in crop pixels; the max value doubles as the confidence.
Variants
All variants take and return fp32 tensors — swap the
.pte
file, keep your app code.
build
file
size (MB)
parity vs fp32 eager (worst corr)
Mac median (ms)*
fp32
rtmpose_m_face_xnnpack_fp32.pte
67.9
1.000000
10.4
Core ML (fp16, iOS)
rtmpose_m_face_coreml_all.pte
34.4
0.999897
3.3
The Core ML build is the same graph lowered to Apple's Neural Engine instead of
XNNPACK, which is CPU-only. Measured on an iPhone 17 Pro across seven models, it
runs
3.5x to 13.9x faster (median 12x)
at roughly half the file size — for
example Depth-Anything-V2-Small at 500.8 ms against 42.7 ms, and MODNet at 81.7 ms
against 5.9 ms. It computes in fp16 and is iOS-only; the XNNPACK files stay the
portable option and are what runs on Android.
*Mac arm64, single process, median of 10 — a reference point for relative cost
only, not a device number (torch eager fp32 on the same machine: 94.4 ms).
Checked in the task's own units
Correlation is a first filter. These are the numbers that decide:
Core ML (fp16, iOS)
— measured in the units that matter for this model — fraction of keypoints landing within 4 px of fp32: median 0.9811 over 10 real images, worst 0.9151.
Verification (executorch 1.4.0, torch 2.13.0)
Parity is measured against the fp32 eager model on real image input;
corr
is
the correlation over all elements of each output tensor.
output
shape
max_abs_diff
corr
0
[1, 106, 512]
1.147e-05
1.000000
1
[1, 106, 512]
9.179e-06
1.000000
XNNPACK delegate coverage (fp32): 93.9% (306/326 ops); ops left on the portable kernels:
dim_order_ops._to_dim_order_copy.default
x8,
aten.unsqueeze_copy.default
x3,
aten.split_with_sizes_copy.default
x3,
aten.sum.dim_IntList
x2,
aten.pow.Tensor_Scalar
x2,
aten.squeeze_copy.dims
x2
RTMPose-m-Face-ExecuTorch huggingface.co is an AI model on huggingface.co that provides RTMPose-m-Face-ExecuTorch's model effect (), which can be used instantly with this mlboydaisuke RTMPose-m-Face-ExecuTorch model. huggingface.co supports a free trial of the RTMPose-m-Face-ExecuTorch model, and also provides paid use of the RTMPose-m-Face-ExecuTorch. Support call RTMPose-m-Face-ExecuTorch model through api, including Node.js, Python, http.
RTMPose-m-Face-ExecuTorch huggingface.co is an online trial and call api platform, which integrates RTMPose-m-Face-ExecuTorch's modeling effects, including api services, and provides a free online trial of RTMPose-m-Face-ExecuTorch, you can try RTMPose-m-Face-ExecuTorch online for free by clicking the link below.
mlboydaisuke RTMPose-m-Face-ExecuTorch online free url in huggingface.co:
RTMPose-m-Face-ExecuTorch is an open source model from GitHub that offers a free installation service, and any user can find RTMPose-m-Face-ExecuTorch on GitHub to install. At the same time, huggingface.co provides the effect of RTMPose-m-Face-ExecuTorch install, users can directly use RTMPose-m-Face-ExecuTorch installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
RTMPose-m-Face-ExecuTorch install url in huggingface.co: