Promptable segmentation models
8 promptable segmentation models. Each section says what the model is, whether it may be used commercially and how fast it is; licences and checks are under Details.
Input: one image plus a point, box or text prompt. Output: masks with quality scores.
Measured on RTX 5060 Ti, torch 2.14.1+cu130; fastforks bench MODEL --size HxW measures yours. The summary uses one input, 1024x1024: kernels time with its speed-up over the faster original run, and kernels peak reserved memory with its ratio to the smaller original one. Where measured (one variant per family), a variant's table also lists portable under torch.compile and the fp16 TensorRT engine of portable (export; memory: the engine's device memory). – = not measured.
Other devices: Apple M1 Pro 16 GB, NVIDIA L4, NVIDIA L40S, NVIDIA A10, NVIDIA A100-SXM4-40GB, NVIDIA RTX PRO 6000; across devices.
| model | commercial | kernels time (ms), speed-up vs original | kernels peak memory (MiB, reserved), ratio to original |
|---|---|---|---|
sam-vit-b | yes | 36.95, 2.7x | 420, 0.23x |
sam-vit-h | yes | 143.13, 2.1x | 1524, 0.28x |
sam-vit-l | yes | 81.53, 2.4x | 862, 0.25x |
sam2.1-hiera-base-plus | yes | 25.09, 1.9x | 524, 0.6x |
sam2.1-hiera-large | yes | 48.81, 1.8x | 828, 0.55x |
sam2.1-hiera-small | yes | 15.56, 2.7x | 386, 0.58x |
sam2.1-hiera-tiny | yes | 13.91, 2.7x | 362, 0.57x |
sam3 | yes | 116.47, 2.4x | 2264, 0.58x |
All models
Modes, licences, training data and outputs per model
Commercial status per variant: yes if the code and the weights licence both allow commercial use (the training data is listed but does not change it; weights without a licence of their own are assumed to carry the code licence); no if either forbids it; unclear for custom licence terms (treat as no).
| model | modes | code licence | weights licence | training data | commercial | outputs |
|---|---|---|---|---|---|---|
sam-vit-b | portable, kernels | Apache-2.0 | Apache-2.0 | unstated | yes | pred_masks, iou_scores |
sam-vit-h | portable, kernels | Apache-2.0 | Apache-2.0 | unstated | yes | pred_masks, iou_scores |
sam-vit-l | portable, kernels | Apache-2.0 | Apache-2.0 | unstated | yes | pred_masks, iou_scores |
sam2.1-hiera-base-plus | portable, kernels | Apache-2.0 | Apache-2.0 | unstated | yes | pred_masks, iou_scores |
sam2.1-hiera-large | portable, kernels | Apache-2.0 | Apache-2.0 | unstated | yes | pred_masks, iou_scores |
sam2.1-hiera-small | portable, kernels | Apache-2.0 | Apache-2.0 | unstated | yes | pred_masks, iou_scores |
sam2.1-hiera-tiny | portable, kernels | Apache-2.0 | Apache-2.0 | unstated | yes | pred_masks, iou_scores |
sam3 | portable, kernels | LicenseRef-SAM-License AND Apache-2.0 | LicenseRef-SAM-License | unstated | yes | boxes, scores, masks |
Training datasets
| id | dataset | licence | commercial use |
|---|