sebastientaylor commited on
Commit
82c6636
Β·
verified Β·
1 Parent(s): a04ef5f

Update model card for yolov8-det

Browse files
Files changed (1) hide show
  1. README.md +14 -14
README.md CHANGED
@@ -25,10 +25,10 @@ model-index:
25
  metrics:
26
  - name: "mAP@0.5 (Nano ONNX FP32)"
27
  type: map_50
28
- value: 50.54
29
  - name: "mAP@0.5-0.95 (Nano ONNX FP32)"
30
  type: map
31
- value: 35.86
32
  - name: "mAP@0.5 (Small ONNX FP32)"
33
  type: map_50
34
  value: 59.4
@@ -37,10 +37,10 @@ model-index:
37
  value: 43.22
38
  - name: "mAP@0.5 (Medium ONNX FP32)"
39
  type: map_50
40
- value: 64.7
41
  - name: "mAP@0.5-0.95 (Medium ONNX FP32)"
42
  type: map
43
- value: 48.46
44
  ---
45
 
46
  # YOLOv8 Detection β€” EdgeFirst Model Zoo
@@ -64,9 +64,9 @@ Accuracy ceiling for each size, measured against COCO `val2017` (5,000 images) w
64
 
65
  | Size | Params | GFLOPs | mAP@0.5 | mAP@0.5-0.95 | mAP@0.75 | Source |
66
  |------|--------|--------|---------|--------------|----------|--------|
67
- | Nano | 3.2M | 8.9 | 50.54% | 35.86% | 38.96% | [v-718](https://edgefirst.studio/public/validation/v-718/details?mode=charts) |
68
  | Small | 11.2M | 28.8 | 59.40% | 43.22% | 46.68% | [v-6fd](https://edgefirst.studio/public/validation/v-6fd/details?mode=charts) |
69
- | Medium | 25.9M | 79.3 | 64.70% | 48.46% | 52.67% | [v-71a](https://edgefirst.studio/public/validation/v-71a/details?mode=charts) |
70
  | Large | 43.7M | 165.7 | β€” | β€” | β€” | β€” |
71
  | XLarge | 68.2M | 258.5 | β€” | β€” | β€” | β€” |
72
 
@@ -86,15 +86,15 @@ Row conventions in the table below:
86
 
87
  | Size | Platform | mAP@0.5 | Ξ” vs FP32 (pp) | mAP@0.5-0.95 | Inference (ms) | End-to-end (ms) | Throughput (FPS) | Source |
88
  |------|----------|---------|----------------|--------------|----------------|-----------------|------------------|--------|
89
- | Nano | ONNX FP32 (AWS Graviton Β· 48-core) | 50.54% | ref | 35.86% | 29.97 | 38.58 | 125.0 | [v-718](https://edgefirst.studio/public/validation/v-718/details?mode=charts) |
90
- | Nano | ONNX FP32 (CUDA) | 50.49% | -0.05 | 35.83% | 8.15 | 13.45 | 344.0 | [v-6fb](https://edgefirst.studio/public/validation/v-6fb/details?mode=charts) |
91
- | Nano | ONNX FP16 (CUDA) | 50.49% | -0.05 | 35.79% | 6.31 | 11.88 | 448.0 | [v-6fc](https://edgefirst.studio/public/validation/v-6fc/details?mode=charts) |
92
  | Nano | macOS CoreML β€” Neural Engine (FP16) | 50.53% | -0.03 | 35.83% | 1.75 | 5.53 | 749.6 | [v-514](https://edgefirst.studio/public/validation/v-514/details?mode=charts) |
93
  | Nano | macOS CoreML β€” Metal GPU (FP16) | 50.52% | -0.04 | 35.83% | 3.77 | 7.73 | 431.0 | [v-51a](https://edgefirst.studio/public/validation/v-51a/details?mode=charts) |
94
  | Nano | macOS CoreML β€” CPU (FP16) | 50.56% | +0.00 | 35.83% | 17.00 | 21.78 | 116.5 | [v-51b](https://edgefirst.studio/public/validation/v-51b/details?mode=charts) |
95
  | Nano | NXP i.MX 8M Plus + VeriSilicon NPU | 48.34% | -2.22 | 33.98% | 60.76 | 105.27 | 14.0 | [v-529](https://edgefirst.studio/public/validation/v-529/details?mode=charts) |
96
- | Nano | NXP i.MX 95 + eIQ Neutron NPU | 48.77% | -1.77 | 34.34% | 46.14 | 73.35 | 77.0 | [v-6b9](https://edgefirst.studio/public/validation/v-6b9/details?mode=charts) |
97
- | Nano | NXP ARA240 (Kinara DVM) | 45.71% | -4.83 | 31.48% | 7.42 | 47.14 | 225.5 | [v-6bb](https://edgefirst.studio/public/validation/v-6bb/details?mode=charts) |
98
  | Nano | Raspberry Pi 5 + Hailo-8L NPU | 48.55% | -2.01 | 34.36% | 13.20 | 23.79 | 65.0 | [v-4d9](https://edgefirst.studio/public/validation/v-4d9/details?mode=charts) |
99
  | Nano | NVIDIA Jetson Orin Nano (TensorRT FP16) | 50.50% | -0.06 | 35.80% | 7.68 | 18.41 | 179.2 | [v-4d8](https://edgefirst.studio/public/validation/v-4d8/details?mode=charts) |
100
  | Small | ONNX FP32 (AWS Graviton Β· 48-core) | 59.37% | -0.03 | 43.18% | 74.56 | 83.28 | 52.0 | [v-719](https://edgefirst.studio/public/validation/v-719/details?mode=charts) |
@@ -106,9 +106,9 @@ Row conventions in the table below:
106
  | Small | NXP ARA240 (Kinara DVM) | 54.79% | -4.61 | 38.18% | 11.59 | 25.06 | 119.0 | [v-6bc](https://edgefirst.studio/public/validation/v-6bc/details?mode=charts) |
107
  | Small | Raspberry Pi 5 + Hailo-8L NPU | 57.80% | -1.57 | 41.77% | 30.37 | 40.86 | 28.8 | [v-4da](https://edgefirst.studio/public/validation/v-4da/details?mode=charts) |
108
  | Small | NVIDIA Jetson Orin Nano (TensorRT FP16) | 59.40% | +0.03 | 43.16% | 14.96 | 24.20 | 113.2 | [v-4db](https://edgefirst.studio/public/validation/v-4db/details?mode=charts) |
109
- | Medium | ONNX FP32 (AWS Graviton Β· 48-core) | 64.70% | ref | 48.46% | 181.16 | 189.63 | 20.0 | [v-71a](https://edgefirst.studio/public/validation/v-71a/details?mode=charts) |
110
- | Medium | ONNX FP32 (CUDA) | 64.68% | -0.02 | 48.45% | 43.58 | 49.61 | 79.0 | [v-701](https://edgefirst.studio/public/validation/v-701/details?mode=charts) |
111
- | Medium | ONNX FP16 (CUDA) | 64.68% | -0.02 | 48.40% | 23.10 | 29.83 | 148.0 | [v-702](https://edgefirst.studio/public/validation/v-702/details?mode=charts) |
112
 
113
 
114
 
 
25
  metrics:
26
  - name: "mAP@0.5 (Nano ONNX FP32)"
27
  type: map_50
28
+ value: 50.49
29
  - name: "mAP@0.5-0.95 (Nano ONNX FP32)"
30
  type: map
31
+ value: 35.83
32
  - name: "mAP@0.5 (Small ONNX FP32)"
33
  type: map_50
34
  value: 59.4
 
37
  value: 43.22
38
  - name: "mAP@0.5 (Medium ONNX FP32)"
39
  type: map_50
40
+ value: 64.68
41
  - name: "mAP@0.5-0.95 (Medium ONNX FP32)"
42
  type: map
43
+ value: 48.45
44
  ---
45
 
46
  # YOLOv8 Detection β€” EdgeFirst Model Zoo
 
64
 
65
  | Size | Params | GFLOPs | mAP@0.5 | mAP@0.5-0.95 | mAP@0.75 | Source |
66
  |------|--------|--------|---------|--------------|----------|--------|
67
+ | Nano | 3.2M | 8.9 | 50.49% | 35.83% | 38.84% | [v-6fb](https://edgefirst.studio/public/validation/v-6fb/details?mode=charts) |
68
  | Small | 11.2M | 28.8 | 59.40% | 43.22% | 46.68% | [v-6fd](https://edgefirst.studio/public/validation/v-6fd/details?mode=charts) |
69
+ | Medium | 25.9M | 79.3 | 64.68% | 48.45% | 52.62% | [v-701](https://edgefirst.studio/public/validation/v-701/details?mode=charts) |
70
  | Large | 43.7M | 165.7 | β€” | β€” | β€” | β€” |
71
  | XLarge | 68.2M | 258.5 | β€” | β€” | β€” | β€” |
72
 
 
86
 
87
  | Size | Platform | mAP@0.5 | Ξ” vs FP32 (pp) | mAP@0.5-0.95 | Inference (ms) | End-to-end (ms) | Throughput (FPS) | Source |
88
  |------|----------|---------|----------------|--------------|----------------|-----------------|------------------|--------|
89
+ | Nano | ONNX FP32 (AWS Graviton Β· 48-core) | 50.54% | +0.05 | 35.86% | 29.97 | 38.58 | 125.0 | [v-718](https://edgefirst.studio/public/validation/v-718/details?mode=charts) |
90
+ | Nano | ONNX FP32 (CUDA) | 50.49% | ref | 35.83% | 8.15 | 13.45 | 344.0 | [v-6fb](https://edgefirst.studio/public/validation/v-6fb/details?mode=charts) |
91
+ | Nano | ONNX FP16 (CUDA) | 50.49% | +0.00 | 35.79% | 6.31 | 11.88 | 448.0 | [v-6fc](https://edgefirst.studio/public/validation/v-6fc/details?mode=charts) |
92
  | Nano | macOS CoreML β€” Neural Engine (FP16) | 50.53% | -0.03 | 35.83% | 1.75 | 5.53 | 749.6 | [v-514](https://edgefirst.studio/public/validation/v-514/details?mode=charts) |
93
  | Nano | macOS CoreML β€” Metal GPU (FP16) | 50.52% | -0.04 | 35.83% | 3.77 | 7.73 | 431.0 | [v-51a](https://edgefirst.studio/public/validation/v-51a/details?mode=charts) |
94
  | Nano | macOS CoreML β€” CPU (FP16) | 50.56% | +0.00 | 35.83% | 17.00 | 21.78 | 116.5 | [v-51b](https://edgefirst.studio/public/validation/v-51b/details?mode=charts) |
95
  | Nano | NXP i.MX 8M Plus + VeriSilicon NPU | 48.34% | -2.22 | 33.98% | 60.76 | 105.27 | 14.0 | [v-529](https://edgefirst.studio/public/validation/v-529/details?mode=charts) |
96
+ | Nano | NXP i.MX 95 + eIQ Neutron NPU | 48.77% | -1.72 | 34.34% | 46.14 | 73.35 | 77.0 | [v-6b9](https://edgefirst.studio/public/validation/v-6b9/details?mode=charts) |
97
+ | Nano | NXP ARA240 (Kinara DVM) | 45.71% | -4.78 | 31.48% | 7.42 | 47.14 | 225.5 | [v-6bb](https://edgefirst.studio/public/validation/v-6bb/details?mode=charts) |
98
  | Nano | Raspberry Pi 5 + Hailo-8L NPU | 48.55% | -2.01 | 34.36% | 13.20 | 23.79 | 65.0 | [v-4d9](https://edgefirst.studio/public/validation/v-4d9/details?mode=charts) |
99
  | Nano | NVIDIA Jetson Orin Nano (TensorRT FP16) | 50.50% | -0.06 | 35.80% | 7.68 | 18.41 | 179.2 | [v-4d8](https://edgefirst.studio/public/validation/v-4d8/details?mode=charts) |
100
  | Small | ONNX FP32 (AWS Graviton Β· 48-core) | 59.37% | -0.03 | 43.18% | 74.56 | 83.28 | 52.0 | [v-719](https://edgefirst.studio/public/validation/v-719/details?mode=charts) |
 
106
  | Small | NXP ARA240 (Kinara DVM) | 54.79% | -4.61 | 38.18% | 11.59 | 25.06 | 119.0 | [v-6bc](https://edgefirst.studio/public/validation/v-6bc/details?mode=charts) |
107
  | Small | Raspberry Pi 5 + Hailo-8L NPU | 57.80% | -1.57 | 41.77% | 30.37 | 40.86 | 28.8 | [v-4da](https://edgefirst.studio/public/validation/v-4da/details?mode=charts) |
108
  | Small | NVIDIA Jetson Orin Nano (TensorRT FP16) | 59.40% | +0.03 | 43.16% | 14.96 | 24.20 | 113.2 | [v-4db](https://edgefirst.studio/public/validation/v-4db/details?mode=charts) |
109
+ | Medium | ONNX FP32 (AWS Graviton Β· 48-core) | 64.70% | +0.02 | 48.46% | 181.16 | 189.63 | 20.0 | [v-71a](https://edgefirst.studio/public/validation/v-71a/details?mode=charts) |
110
+ | Medium | ONNX FP32 (CUDA) | 64.68% | ref | 48.45% | 43.58 | 49.61 | 79.0 | [v-701](https://edgefirst.studio/public/validation/v-701/details?mode=charts) |
111
+ | Medium | ONNX FP16 (CUDA) | 64.68% | +0.00 | 48.40% | 23.10 | 29.83 | 148.0 | [v-702](https://edgefirst.studio/public/validation/v-702/details?mode=charts) |
112
 
113
 
114