Chiebot-Mirror
/
ultralytics
9
0
Fork 1
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Update NVIDIA Jetson TensorRT Benchmarks (#16156)

Browse Source 
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
pull/16086/head^2
Lakshantha Dissanayake 8 months ago committed by GitHub
parent be71a43773
commit fbc6bae5af
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 71 additions and 61 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 32
      docs/en/guides/nvidia-jetson.md

32
docs/en/guides/nvidia-jetson.md
Unescape View File

@ -287,7 +287,7 @@ YOLOv8 benchmarks were run by the Ultralytics team on 10 different model formats
Even though all model exports are working with NVIDIA Jetson, we have only included **PyTorch, TorchScript, TensorRT** for the comparison chart below because, they make use of the GPU on the Jetson and are guaranteed to produce the best results. All the other exports only utilize the CPU and the performance is not as good as the above three. You can find benchmarks for all exports in the section after this chart.
<div style="text-align: center;">
<img width="800" src="https://github.com/ultralytics/docs/releases/download/0/nvidia-jetson-ecosystem-1.avif" alt="NVIDIA Jetson Ecosystem">
<img width="800" src="https://github.com/ultralytics/docs/releases/download/0/nvidia-jetson-ecosystem-2.avif" alt="NVIDIA Jetson Ecosystem">
</div>
### Detailed Comparison Table
@ -299,12 +299,14 @@ The below table represents the benchmark results for five different models (YOLO
=== "YOLOv8n"
| Format | Status | Size on disk (MB) | mAP50-95(B) | Inference time (ms/im) |
|---------------|--------|-----------|-------------|------------------------|
|-----------------|--------|-------------------|-------------|------------------------|
| PyTorch | ✅ | 6.2 | 0.6381 | 14.3 |
| TorchScript | ✅ | 12.4 | 0.6117 | 13.3 |
| ONNX | ✅ | 12.2 | 0.6092 | 70.6 |
| OpenVINO | ✅ | 12.3 | 0.6092 | 104.2 |
| TensorRT | ✅ | 13.6 | 0.6117 | 8.9 |
| TensorRT (FP32) | ✅ | 16.1 | 0.6091 | 8.01 |
| TensorRT (FP16) | ✅ | 9.2 | 0.6093 | 4.55 |
| TensorRT (INT8) | ✅ | 5.9 | 0.2759 | 4.09 |
| TF SavedModel | ✅ | 30.6 | 0.6092 | 141.74 |
| TF GraphDef | ✅ | 12.3 | 0.6092 | 199.93 |
| TF Lite | ✅ | 12.3 | 0.6092 | 349.18 |
@ -314,12 +316,14 @@ The below table represents the benchmark results for five different models (YOLO
=== "YOLOv8s"
| Format | Status | Size on disk (MB) | mAP50-95(B) | Inference time (ms/im) |
|---------------|--------|-----------|-------------|------------------------|
|-----------------|--------|-------------------|-------------|------------------------|
| PyTorch | ✅ | 21.5 | 0.6967 | 18 |
| TorchScript | ✅ | 43.0 | 0.7136 | 23.81 |
| ONNX | ✅ | 42.8 | 0.7136 | 185.55 |
| OpenVINO | ✅ | 42.9 | 0.7136 | 243.97 |
| TensorRT | ✅ | 44.0 | 0.7136 | 14.82 |
| TensorRT (FP32) | ✅ | 46.4 | 0.7136 | 14.01 |
| TensorRT (FP16) | ✅ | 24.2 | 0.722 | 7.16 |
| TensorRT (INT8) | ✅ | 13.7 | 0.4233 | 5.49 |
| TF SavedModel | ✅ | 107 | 0.7136 | 260.03 |
| TF GraphDef | ✅ | 42.8 | 0.7136 | 423.4 |
| TF Lite | ✅ | 42.8 | 0.7136 | 1046.64 |
@ -329,12 +333,14 @@ The below table represents the benchmark results for five different models (YOLO
=== "YOLOv8m"
| Format | Status | Size on disk (MB) | mAP50-95(B) | Inference time (ms/im) |
|---------------|--------|-----------|-------------|------------------------|
|-----------------|--------|-------------------|-------------|------------------------|
| PyTorch | ✅ | 49.7 | 0.7370 | 36.4 |
| TorchScript | ✅ | 99.2 | 0.7285 | 53.58 |
| ONNX | ✅ | 99 | 0.7280 | 452.09 |
| OpenVINO | ✅ | 99.1 | 0.7280 | 544.36 |
| TensorRT | ✅ | 100.3 | 0.7285 | 33.21 |
| TensorRT (FP32) | ✅ | 102.4 | 0.7285 | 31.51 |
| TensorRT (FP16) | ✅ | 52.6 | 0.7324 | 14.88 |
| TensorRT (INT8) | ✅ | 28.6 | 0.3283 | 10.89 |
| TF SavedModel | ✅ | 247.5 | 0.7280 | 543.65 |
| TF GraphDef | ✅ | 99 | 0.7280 | 906.63 |
| TF Lite | ✅ | 99 | 0.7280 | 2758.08 |
@ -344,12 +350,14 @@ The below table represents the benchmark results for five different models (YOLO
=== "YOLOv8l"
| Format | Status | Size on disk (MB) | mAP50-95(B) | Inference time (ms/im) |
|---------------|--------|-----------|-------------|------------------------|
|-----------------|--------|-------------------|-------------|------------------------|
| PyTorch | ✅ | 83.7 | 0.7768 | 61.3 |
| TorchScript | ✅ | 167.2 | 0.7554 | 87.9 |
| ONNX | ✅ | 166.8 | 0.7551 | 852.29 |
| OpenVINO | ✅ | 167 | 0.7551 | 1012.6 |
| TensorRT | ✅ | 168.4 | 0.7554 | 51.23 |
| TensorRT (FP32) | ✅ | 170.5 | 0.7554 | 49.79 |
| TensorRT (FP16) | ✅ | 86.1 | 0.7535 | 22.89 |
| TensorRT (INT8) | ✅ | 46.4 | 0.4048 | 14.61 |
| TF SavedModel | ✅ | 417.2 | 0.7551 | 990.45 |
| TF GraphDef | ✅ | 166.9 | 0.7551 | 1649.86 |
| TF Lite | ✅ | 166.9 | 0.7551 | 5652.37 |
@ -359,12 +367,14 @@ The below table represents the benchmark results for five different models (YOLO
=== "YOLOv8x"
| Format | Status | Size on disk (MB) | mAP50-95(B) | Inference time (ms/im) |
|---------------|--------|-----------|-------------|------------------------|
|-----------------|--------|-------------------|-------------|------------------------|
| PyTorch | ✅ | 130.5 | 0.7759 | 93 |
| TorchScript | ✅ | 260.7 | 0.7472 | 135.1 |
| ONNX | ✅ | 260.4 | 0.7479 | 1296.13 |
| OpenVINO | ✅ | 260.6 | 0.7479 | 1502.15 |
| TensorRT | ✅ | 261.8 | 0.7469 | 84.53 |
| TensorRT (FP32) | ✅ | 264.0 | 0.7469 | 80.01 |
| TensorRT (FP16) | ✅ | 133.3 | 0.7513 | 40.76 |
| TensorRT (INT8) | ✅ | 70.2 | 0.4277 | 22.08 |
| TF SavedModel | ✅ | 651.1 | 0.7479 | 1451.76 |
| TF GraphDef | ✅ | 260.5 | 0.7479 | 4029.36 |
| TF Lite | ✅ | 260.4 | 0.7479 | 8772.86 |

Write Preview
Loading…
Cancel
Save