Report realized vs theoretical FPS; fix duplicate-row dedup
Browse files
README.md
CHANGED
|
@@ -24,18 +24,18 @@ Each model family lives in its own HuggingFace repo containing all size variants
|
|
| 24 |
|
| 25 |
| Repo | Model | Sizes | Nano mAP@0.5 |
|
| 26 |
|------|-------|-------|-------------|
|
| 27 |
-
| [EdgeFirst/yolo26-det](https://huggingface.co/EdgeFirst/yolo26-det) | YOLO26 | n/s/m
|
| 28 |
-
| [EdgeFirst/yolo11-det](https://huggingface.co/EdgeFirst/yolo11-det) | YOLO11 | n/s/m
|
| 29 |
-
| [EdgeFirst/yolov8-det](https://huggingface.co/EdgeFirst/yolov8-det) | YOLOv8 | n/s/m
|
| 30 |
-
| [EdgeFirst/yolov5-det](https://huggingface.co/EdgeFirst/yolov5-det) | YOLOv5 | n/s/m
|
| 31 |
|
| 32 |
### Segmentation
|
| 33 |
|
| 34 |
| Repo | Model | Sizes | Nano Mask mAP |
|
| 35 |
|------|-------|-------|--------------|
|
| 36 |
-
| [EdgeFirst/yolo26-seg](https://huggingface.co/EdgeFirst/yolo26-seg) | YOLO26 | n/s/m
|
| 37 |
-
| [EdgeFirst/yolo11-seg](https://huggingface.co/EdgeFirst/yolo11-seg) | YOLO11 | n/s
|
| 38 |
-
| [EdgeFirst/yolov8-seg](https://huggingface.co/EdgeFirst/yolov8-seg) | YOLOv8 | n/s/m
|
| 39 |
|
| 40 |
---
|
| 41 |
|
|
@@ -93,16 +93,15 @@ EdgeFirst/yolov8-det/
|
|
| 93 |
|
| 94 |
## Supported Hardware
|
| 95 |
|
| 96 |
-
     | YOLO26 | n/s/m | 55.0% |
|
| 28 |
+
| [EdgeFirst/yolo11-det](https://huggingface.co/EdgeFirst/yolo11-det) | YOLO11 | n/s/m | 53.1% |
|
| 29 |
+
| [EdgeFirst/yolov8-det](https://huggingface.co/EdgeFirst/yolov8-det) | YOLOv8 | n/s/m | 50.5% |
|
| 30 |
+
| [EdgeFirst/yolov5-det](https://huggingface.co/EdgeFirst/yolov5-det) | YOLOv5 | n/s/m | 47.8% |
|
| 31 |
|
| 32 |
### Segmentation
|
| 33 |
|
| 34 |
| Repo | Model | Sizes | Nano Mask mAP |
|
| 35 |
|------|-------|-------|--------------|
|
| 36 |
+
| [EdgeFirst/yolo26-seg](https://huggingface.co/EdgeFirst/yolo26-seg) | YOLO26 | n/s/m | 32.7% |
|
| 37 |
+
| [EdgeFirst/yolo11-seg](https://huggingface.co/EdgeFirst/yolo11-seg) | YOLO11 | n/s | 30.2% |
|
| 38 |
+
| [EdgeFirst/yolov8-seg](https://huggingface.co/EdgeFirst/yolov8-seg) | YOLOv8 | n/s/m | 28.7% |
|
| 39 |
|
| 40 |
---
|
| 41 |
|
|
|
|
| 93 |
|
| 94 |
## Supported Hardware
|
| 95 |
|
| 96 |
+
       
|
| 97 |
|
| 98 |
- **Linux x86_64** β ONNX Runtime CUDA / CPU (FP32 reference)
|
| 99 |
- **Linux aarch64** β ONNX Runtime / TFLite (ARM64 generic Linux)
|
| 100 |
- **Apple macOS** β ONNX Runtime + CoreML ANE / GPU / CPU (FP16)
|
| 101 |
- **NXP i.MX 8M Plus** β 2.3 TOPS, TFLite INT8
|
| 102 |
+
- **NXP i.MX 95** β 2.0 TOPS, eIQ Neutron TFLite *(YOLOv5 / YOLOv8 only; YOLO11 / YOLO26 not yet supported on eIQ Neutron)*
|
| 103 |
+
- **NXP Ara240** β 40 eTOPS, .DVM
|
| 104 |
+
- **RPi5 + Hailo-8L** β 13 TOPS, HailoRT HEF
|
|
|
|
| 105 |
- **NVIDIA Jetson Orin** β 67β157 TOPS, TensorRT
|
| 106 |
|
| 107 |
## Validation Pipeline
|