Update yolonas docs (#17736)

* Adjust pre_nms predictions to increase performance

* Update yolonas export and rocm inference times
This commit is contained in:
Nicolas Mowen 2025-04-16 09:01:15 -06:00 committed by GitHub
parent 28e2f57c76
commit 6a487a5dc7
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 8 additions and 7 deletions

View File

@ -145,7 +145,7 @@ With the [rocm](../configuration/object_detectors.md#amdrocm-gpu-detector) detec
| Name | YoloV9 Inference Time | YOLO-NAS Inference Time | | Name | YoloV9 Inference Time | YOLO-NAS Inference Time |
| --------------- | --------------------- | ------------------------- | | --------------- | --------------------- | ------------------------- |
| AMD 780M | ~ 14 ms | 320: ~ 60 ms 640: ~ 80 ms | | AMD 780M | ~ 14 ms | 320: ~ 30 ms 640: ~ 60 ms |
## Community Supported Detectors ## Community Supported Detectors

View File

@ -13,15 +13,15 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"source": [ "execution_count": null,
"! sed -i 's/sghub.deci.ai/sg-hub-nv.s3.amazonaws.com/' /usr/local/lib/python3.11/dist-packages/super_gradients/training/pretrained_models.py\n",
"! sed -i 's/sghub.deci.ai/sg-hub-nv.s3.amazonaws.com/' /usr/local/lib/python3.11/dist-packages/super_gradients/training/utils/checkpoint_utils.py"
],
"metadata": { "metadata": {
"id": "NiRCt917KKcL" "id": "NiRCt917KKcL"
}, },
"execution_count": null, "outputs": [],
"outputs": [] "source": [
"! sed -i 's/sghub.deci.ai/sg-hub-nv.s3.amazonaws.com/' /usr/local/lib/python3.11/dist-packages/super_gradients/training/pretrained_models.py\n",
"! sed -i 's/sghub.deci.ai/sg-hub-nv.s3.amazonaws.com/' /usr/local/lib/python3.11/dist-packages/super_gradients/training/utils/checkpoint_utils.py"
]
}, },
{ {
"cell_type": "code", "cell_type": "code",
@ -51,6 +51,7 @@
"model.export(\"yolo_nas_s.onnx\",\n", "model.export(\"yolo_nas_s.onnx\",\n",
" output_predictions_format=DetectionOutputFormatMode.FLAT_FORMAT,\n", " output_predictions_format=DetectionOutputFormatMode.FLAT_FORMAT,\n",
" max_predictions_per_image=20,\n", " max_predictions_per_image=20,\n",
" num_pre_nms_predictions=300,\n",
" confidence_threshold=0.4,\n", " confidence_threshold=0.4,\n",
" input_image_shape=(320,320),\n", " input_image_shape=(320,320),\n",
" )" " )"