Downloads · 30 days
0
sam2.1-hiera-large-onnx
latolukasz/sam2.1-hiera-large-onnx
sam2.1-hiera-large-onnx is a machine learning model from latolukasz. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
Server-side export of Meta's SAM 2.1 (sam2.1hieralarge checkpoint, 224.4M params).
Access
Public
Updated Aug 14, 2026
Repo size
869 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.onnx869 MB · 100%
At a glance
- License
- apache-2.0
- Access
- Public
- Created
- Aug 14, 2026
- Updated
- Aug 14, 2026
- SHA
- ccbb4af9
Model card
From the Hugging Face model README
SAM 2.1 Hiera Large — ONNX
Server-side export of Meta's SAM 2.1 (sam2.1_hiera_large checkpoint, 224.4M params).
Files
| File | Size | Notes |
|---|---|---|
encoder.onnx | 852 MB | Image encoder, plain ONNX (fp32). Loads in native onnxruntime (Python/Go/C). NOT converted to .ort — onnxruntime-web's strict shape inference is not a target for this repo. |
decoder.onnx | 16.5 MB | Prompt decoder (points/box → masks), standard ONNX |
Provenance
- Source ONNX: vietanhdev/segment-anything-2.1-onnx-models (
sam2.1_hiera_large_20260221.zip, revision6a3ac868340a3196a349050a6efae22a5acc0330, exported with samexporter) - Files renamed to
encoder.onnx/decoder.onnx, otherwise unmodified.
Interface
Encoder: input image [1,3,1024,1024] float32 (ImageNet-normalized) → high_res_feats_0 [1,32,256,256], high_res_feats_1 [1,64,128,128], image_embed [1,256,64,64].
Decoder: image_embed, high_res_feats_0/1, point_coords [n,p,2], point_labels [n,p] float32, mask_input [n,1,256,256], has_mask_input [n] → masks [n,3,256,256] logits, iou_predictions [n,3].
Same interface as latolukasz/sam2.1-hiera-base-plus-onnx — drop-in backbone swap.
- License
- apache-2.0
- Created
- Aug 14, 2026
- Updated
- Aug 14, 2026