RTMO
OpenMMLab's RTMO (t, s, m, l): real-time one-stage multi-person 2D pose; finds people and their 17 COCO keypoints in one pass over the image, no person detector needed. Apache-2.0.
stable Fork of open-mmlab/mmpose@5408bc7 · backends: portable, kernels
- Paper RTMO: Towards High-Performance One-Stage Real-Time Multi-Person Pose Estimation CVPR 2024 · first released 12 Dec 2023
Paper 2023 Added in 0.2.0 Evaluated with 0.2.0
Use it
pip install fastforks Lean install with only what this model needs: pip install fastforks-core
from fastforks import pose
model = pose.load("rtmo-t") # kernels on a supported GPU, else portable PyTorch
result = model(image, boxes=...)
result.keypoints # also: keypoint_scores, boxes, scores
result.backend # what ran: mode, device, dtype, kernels fastforks bench rtmo-t Times the model on this machine. load(…, mode="portable") or mode="kernels" picks a backend; usage guide.
Speed and memory
Every backend on one GPU (RTX 5060 Ti 16 GB) and the same input, the authors’ code included. Kernels runs fp16. Portable runs fp16 on CUDA.
Input: synthetic, 1080x1920 (the task’s standard size)
| Backend | dtype | Time per call (ms) | vs original fp32 | Peak memory allocated / reserved | Load (s) | First call (s) |
|---|---|---|---|---|---|---|
| Original fp32mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed call | fp32 | 11.3 | – | 90 MiB/108 MiB | 5.3 | 0.43 |
| Original fp16torch.autocast(float16) around the call; mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed call | fp16 | 13.3 | 0.85× | 118 MiB/122 MiB | 0.3 | 0.32 |
| Portable fp32 | fp32 | not measured | – | – | – | – |
| Portable | fp16 | 7.62 | 1.5× | 121 MiB/136 MiB | 0.1 | 0.41 |
| Kernels | fp16 | 6.31 | 1.8× | 121 MiB/156 MiB | 1.4 | 0.28 |
| Portable + torch.compile | 3.30 | 3.4× | 111 MiB/134 MiB | 16 (build) | – | |
| TensorRT fp16network on person crops only | fp16 | 0.886 | 13× | 40 MiBenginenot comparable | 37 (build) | – |
Measured on RTX 5060 Ti 16 GB, torch 2.14.1+cu130, 2026-11-11. First call, 3 warm-up, peak reset, median of 3 rounds x median of 5. Ratios divide two
times of this table (same GPU, same input). Original fp16 is the authors’ own precision option, else torch.autocast. torch.compile and TensorRT: build time instead of load time. TensorRT memory (engine, not comparable) is the GPU memory the engine took, read from the driver because TensorRT allocates outside PyTorch, so it can't be compared with the PyTorch peaks of the other rows.
Other devices
| Device | Backend | dtype | ms / call | Peak memory | Output vs reference | Input |
|---|---|---|---|---|---|---|
| Apple M1 Pro 16 GB MPS, torch 2.14.1, 2026-11-12 | Original | fp32 | 25.4 | 1,836 MiB process | – | 1080x1920 |
| Portable | fp16 | 19.9 | 1,506 MiB process | matches | 1080x1920 | |
| Portable | fp32 | 21.0 | 1,523 MiB process | matches | 1080x1920 | |
| Kernels | fp16 | not implemented | – | 1080x1920 | ||
CUDA memory is torch’s peak allocated / reserved; Apple has no such counter, so the Mac rows show the process footprint (weights, Python and torch included). “Output vs reference” compares the device’s output with the stored golden output under the card’s tolerance; – = not checked on that device. Kernels “not implemented” = no kernels for this model on that backend yet: it runs portable there.
Licence and weights
Commercial use allowed
- Code
- Apache-2.0 LICENSE
- Weights
- Apache-2.0 source
- Training data
-
- COCO 2017 (train, instance annotations) CC-BY-4.0
- AI Challenger human keypoints (2017) not stated
- CrowdPose not stated
- MPII Human Pose BSD-2-Clause, non-commercial
- sub-JHMDB (J-HMDB) not stated
- Halpe-FullBody not stated
- PoseTrack 2018 not stated
Our reading of the licences, not legal advice.
Weights
rtmo-t-body7-416x416.pth 27.6 MB· hosted on OpenMMLabfastforks downloads and checks these files on first use (fastforks download rtmo-t).
Accuracy
The test set through the authors’ code and through ours: same checkpoint, same protocol. The distance between the original’s fp32 and fp16 runs is what half precision alone costs; our runs should stay within it.
COCO val2017
predicted people, keypoint AP
| Run | dtype | keypoint AP ↑ | Δ keypoint AP vs fp32 original | images | ms / image |
|---|---|---|---|---|---|
| Original fp32 | fp32 | 57.4 | reference | 5000 | – |
| Original fp16 | fp16 | 57.4 | ±0.0 | 5000 | – |
| Portable fp32 | not measured | ||||
| Portable | fp16 | 57.4 | −0.0 | 5000 | – |
| Kernels | fp16 | 57.4 | −0.0 | 5000 | – |
Bottom-up on all 5,000 validation images, no ground-truth boxes; people ranked by their predicted person score, after model-specific decoding; pycocotools OKS AP. Original = the authors’ code, same checkpoint and protocol. Measured on original fp32, original fp16: RTX 5060 Ti, torch 2.14.1+cu130, 2026-11-11; portable, kernels: RTX 5060 Ti, torch 2.14.1+cu130, fastforks be54b8b0, 2026-11-11.
COCO val2017
every 10th image, predicted people, keypoint AP
| Run | dtype | keypoint AP ↑ | Δ keypoint AP vs fp32 original | images | ms / image |
|---|---|---|---|---|---|
| Original fp32 | fp32 | 59.4 | reference | 500 | – |
| Original fp16 | fp16 | 59.4 | ±0.0 | 500 | – |
| Portable fp32 | fp32 | 59.3 | −0.0 | 500 | 9.31 |
| Portable | fp16 | 59.3 | −0.1 | 500 | – |
| Kernels | fp16 | 59.3 | −0.0 | 500 | – |
Bottom-up on the same every-tenth-image subset for all modes, ranked by predicted person score after model-specific decoding; pycocotools OKS AP. Original = the authors’ code, same checkpoint and protocol. Measured on original fp32, original fp16: RTX 5060 Ti, torch 2.14.1+cu130, 2026-11-11; portable fp32: RTX 5060 Ti, torch 2.14.1+cu130, fastforks 44640884, 2026-11-17; portable, kernels: RTX 5060 Ti, torch 2.14.1+cu130, fastforks be54b8b0, 2026-11-11.
Closeness to the original
Beyond the test-set scores: the raw output of ours against the original’s fp32 output on the same input, pixel by pixel. CPU, float32: max |diff| <= 0.014 vs the original. GPU, float16 (the default), portable and kernels: mean |diff| <= 0.47 and p99 <= 9.8 (absolute error of the output keypoints).
| Check | Tolerance | Measured |
|---|---|---|
| CPU fp32, max |diff| | ≤ 0.014 | not recorded |
| GPU default dtypes, mean |diff| | ≤ 0.47 | 0.02725 |
| GPU default dtypes, p99 |diff| | ≤ 9.8 | 0.14195 |
| gpu_min_score | ≤ 0.3 | not recorded |
Recorded for rtmo-t on RTX 5060 Ti, torch 2.14.1+cu130, 2026-11-11, with these kernels: cudnn_haccfvm_convfvm_extrafvm_mbfvm_r3
Details
RTMO (Lu et al. 2024): real-time one-stage multi-person 2D pose from OpenMMLab. A YOLOX-style CSPDarknet backbone, a hybrid encoder neck (one transformer layer on the coarsest map) and a head that predicts, per grid cell, a person score, a box and a pose vector; a dynamic coordinate classifier turns each kept pose vector into the 17 COCO keypoints over 1D bins that span the person's box. Takes the whole image without person boxes or a separate detector: the backbone runs once per image, while decoding and NMS depend on the detections. The image is letterboxed to 640 x 640 (416 x 416 for `rtmo-t`); vendored from mmpose (no mmpose / mmcv dependency). The portable `module()` maps the letterboxed image to the head's per-cell outputs (decoding and NMS stay outside).
- Input / output
- Input: one RGB image and a box per person. Output: keypoints with confidence scores for every person.
- Input size
- any H x W >= 32, padded to a multiple of 32; tested up to 3000x4000 (460 MiB allocated)
- Options
score_threshold=0.10.0 to 1.0· changes the output
- Outputs
keypoints,keypoint_scores,boxes,scores- Checkpoint
deepen=0.33,widen=0.375,size=416- CUDA kernels
cudnn_haccconvolutions with fp16 accumulation (cuDNN)fvm_convfused convolutions for image classifiersfvm_extrafused transformer row operations (rotary split, SwiGLU with LayerNorm)fvm_mbfused MBConv blocks (MobileNet and EfficientNet families)fvm_r3further fused classifier blocks
- Apple (MPS)
- fp32, fp16
Limitations
- module() exports the letterboxed-image network and raw per-cell head outputs; coordinate decoding and NMS remain outside the exported graph.
- Float16 and float32 can retain different people near the score threshold.
Full reference, generated from the repository
OpenMMLab's RTMO (t, s, m, l): real-time one-stage multi-person 2D pose; finds people and their 17 COCO keypoints in one pass over the image, no person detector needed. Apache-2.0.
Commercial use: yes.
| variant | input | backend | time (ms) | peak memory (MiB, allocated / reserved) |
|---|---|---|---|---|
rtmo-t | synthetic (1080x1920) | original (fp32) original portable kernels portable + torch.compile TensorRT fp16 | 11.3 13.3 7.62 6.31 3.30 0.886 (network on person crops only) | 90 / 108 118 / 122 121 / 136 121 / 156 111 / 134 112 / 150 |
rtmo-s | synthetic (1080x1920) | original (fp32) original portable kernels portable + torch.compile TensorRT fp16 | 11.4 15.5 7.48 6.27 3.35 1.21 (network on person crops only) | 192 / 220 228 / 252 142 / 160 142 / 180 121 / 156 133 / 164 |
rtmo-m | synthetic (1080x1920) | original (fp32) original portable kernels portable + torch.compile TensorRT fp16 | 14.3 17.0 8.71 6.85 4.99 2.03 (network on person crops only) | 332 / 360 422 / 462 166 / 190 166 / 212 146 / 170 194 / 232 |
rtmo-l | synthetic (1080x1920) | original (fp32) original portable kernels portable + torch.compile TensorRT fp16 | 17.1 21.4 10.8 8.30 7.07 2.90 (network on person crops only) | 614 / 652 800 / 824 208 / 226 208 / 228 192 / 224 277 / 306 |
Details
- Original code: https://github.com/open-mmlab/mmpose at
5408bc76f5b8 - Code licence: Apache-2.0 (permissive); files:
LICENSE - Weights licence: Apache-2.0
- Licence evidence: https://github.com/open-mmlab/mmpose/blob/5408bc76f5b848cf925a0d1857899011d8c5b497/README.md#license
- Training data: COCO 2017 (train, instance annotations), AI Challenger human keypoints (2017), CrowdPose, MPII Human Pose, sub-JHMDB (J-HMDB), Halpe-FullBody, PoseTrack 2018 (body7: COCO, AI Challenger, CrowdPose, MPII, sub-JHMDB, Halpe and PoseTrack18 keypoints; the CSPDarknet backbone starts from a YOLOX COCO detection checkpoint)
- Options:
score_threshold= 0.1 (0.0 to 1.0) - Kernel libraries:
cudnn_hacc,fvm_conv,fvm_extra,fvm_mb,fvm_r3 - Variants:
rtmo-t,rtmo-s,rtmo-m,rtmo-l - Outputs: keypoints, keypoint_scores, boxes, scores; label_set:
coco-keypoints-17 - Input size: any H x W >= 32, padded to a multiple of 32
- Tested up to (portable, fp16, RTX 5060 Ti 16 GB):
rtmo-t3000x4000 (0.4 GiB peak allocated, 0.5 reserved);rtmo-s3000x4000 (0.5 GiB peak allocated, 0.5 reserved);rtmo-m3000x4000 (0.5 GiB peak allocated, 0.5 reserved);rtmo-l3000x4000 (0.5 GiB peak allocated, 0.6 reserved) rtmo-t1080x1920, first call: portable + torch.compile 16 s (cold Inductor cache), TensorRT fp16 37 s (ONNX + engine).rtmo-s1080x1920, first call: portable + torch.compile 15 s (cold Inductor cache), TensorRT fp16 37 s (ONNX + engine).rtmo-m1080x1920, first call: portable + torch.compile 18 s (cold Inductor cache), TensorRT fp16 39 s (ONNX + engine).rtmo-l1080x1920, first call: portable + torch.compile 19 s (cold Inductor cache), TensorRT fp16 40 s (ONNX + engine).rtmo-toriginal (fp32) 1080x1920: mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed callrtmo-toriginal 1080x1920: torch.autocast(float16) around the call; mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed callrtmo-soriginal (fp32) 1080x1920: mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed callrtmo-soriginal 1080x1920: torch.autocast(float16) around the call; mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed callrtmo-moriginal (fp32) 1080x1920: mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed callrtmo-moriginal 1080x1920: torch.autocast(float16) around the call; mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed callrtmo-loriginal (fp32) 1080x1920: mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed callrtmo-loriginal 1080x1920: torch.autocast(float16) around the call; mmpose 1.3.2 inference_bottomup (CPU cv2 letterbox, network, NMS, decode), all in the timed callrtmo-tload / first call: original 0.26 s / 0.324 s, portable 0.11 s / 0.411 s, kernels 1.38 s / 0.279 srtmo-sload / first call: original 0.27 s / 0.076 s, portable 0.12 s / 0.412 s, kernels 1.37 s / 0.264 srtmo-mload / first call: original 0.37 s / 0.077 s, portable 0.22 s / 0.428 s, kernels 1.43 s / 0.242 srtmo-lload / first call: original 0.55 s / 0.063 s, portable 0.36 s / 0.243 s, kernels 1.52 s / 0.16 s- Protocol, first call, 3 warm-up, peak reset, median of 3 rounds x median of 5
Measured
Bottom-up on all 5,000 validation images, no ground-truth boxes; people ranked by their predicted person score, after model-specific decoding; pycocotools OKS AP. Bottom-up on the same every-tenth-image subset for all modes, ranked by predicted person score after model-specific decoding; pycocotools OKS AP. Original = the original code named under Details, same checkpoint and protocol.
| variant | dataset | original fp32 | original fp16 | portable | portable fp32 | kernels | note |
|---|---|---|---|---|---|---|---|
rtmo-t | COCO val2017, predicted people, keypoint AP | 57.4 | 57.4 | 57.4 | 57.4 | ||
rtmo-t | COCO val2017, every 10th image, predicted people, keypoint AP | 59.4 | 59.4 | 59.3 | 59.3 | 59.3 | |
rtmo-s | COCO val2017, predicted people, keypoint AP | 68.6 | 68.6 | 68.5 | 68.5 | ||
rtmo-s | COCO val2017, every 10th image, predicted people, keypoint AP | 71.1 | 71.2 | 71.2 | 71.2 | 71.2 | |
rtmo-m | COCO val2017, predicted people, keypoint AP | 72.6 | 72.6 | 72.6 | 72.6 | ||
rtmo-m | COCO val2017, every 10th image, predicted people, keypoint AP | 74.2 | 74.2 | 74.2 | 74.2 | 74.2 | |
rtmo-l | COCO val2017, predicted people, keypoint AP | 74.8 | 74.8 | 74.8 | 74.8 | ||
rtmo-l | COCO val2017, every 10th image, predicted people, keypoint AP | 76.1 | 76.0 | 76.0 | 76.1 | 76.0 |
Measured on portable: NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks be54b8b0, 2026-11-11; portable32: NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks 44640884, 2026-11-17 to 2026-11-18; kernels: NVIDIA GeForce RTX 5060 Ti, torch 2.14.1+cu130, fastforks be54b8b0, 2026-11-11.
Parity
CPU, float32: max |diff| <= 0.014 vs the original. GPU, float16 (the default), portable and kernels: mean |diff| <= 0.47 and p99 <= 9.8 (absolute error of the output keypoints).
| variant | kernels error vs original, mean / p99 | GPU, torch | kernels used |
|---|---|---|---|
rtmo-t | 0.02725 / 0.14195 | RTX 5060 Ti, torch 2.14.1+cu130 | cudnn_hacc@cudnn92400, fvm_conv@f1d064e539e1:sm_120, fvm_extra@3208ce525a91:sm_120, fvm_mb@720dd3f29bf7:sm_120, fvm_r3@f54eb890e23c:sm_120 |
rtmo-s | 0.02234 / 0.14143 | RTX 5060 Ti, torch 2.14.1+cu130 | cudnn_hacc@cudnn92400, fvm_conv@f1d064e539e1:sm_120, fvm_extra@3208ce525a91:sm_120, fvm_mb@720dd3f29bf7:sm_120, fvm_r3@f54eb890e23c:sm_120 |
rtmo-m | 0.11064 / 1.21749 | RTX 5060 Ti, torch 2.14.1+cu130 | cudnn_hacc@cudnn92400, fvm_conv@f1d064e539e1:sm_120, fvm_extra@3208ce525a91:sm_120, fvm_mb@720dd3f29bf7:sm_120, fvm_r3@f54eb890e23c:sm_120 |
rtmo-l | 0.30973 / 6.5301 | RTX 5060 Ti, torch 2.14.1+cu130 | cudnn_hacc@cudnn92400, fvm_conv@f1d064e539e1:sm_120, fvm_extra@3208ce525a91:sm_120, fvm_mb@720dd3f29bf7:sm_120, fvm_r3@f54eb890e23c:sm_120 |
Limitations
- module() exports the letterboxed-image network and raw per-cell head outputs; coordinate decoding and NMS remain outside the exported graph.
- Float16 and float32 can retain different people near the score threshold.
Example output
One fixed input through this checkpoint, the same for every model of the task (kernels · float16 · NVIDIA GeForce RTX 5060 Ti).
Input: Astronaut Eileen Collins, NASA (public domain).