Part Number: TDA4VM
I find yolov3-tiny config file in the path ti_dl/test/testvecs/config/import/public/onnx/tidl_import_tiny_yolo3.txt. If I want use yolov3-tiny model in TDA4, is the darknet version supported?
Now what I have done is that I convert Darknet weight file of yolov3-tiny to onnx file, and I used tidl_model_import.out to complie onnx file to BIN file. But when I run tidl_model_import.out, it shows some error logs.
SDK version: 7.3
prototxt:
name: "tiny_yolo_v3"
tidl_yolo {
name: "tiny_yolo_v3"
in_width: 416
in_height: 416
#10,14, 23,27, 37,58, 81,82, 135,169, 344,319
yolo_param {
input: "016_convolutional"
anchor_width: 10
anchor_width: 23
anchor_width: 37
anchor_height: 14
anchor_height: 27
anchor_height: 58
}
yolo_param {
input: "023_convolutional"
anchor_width: 81
anchor_width: 135
anchor_width: 344
anchor_height: 82
anchor_height: 169
anchor_height: 319
}
detection_output_param {
num_classes: 80
share_location: true
background_label_id: -1
nms_param {
nms_threshold: 0.45
top_k: 200
}
code_type: CENTER_SIZE
keep_top_k: 100
confidence_threshold: 0.2
}
}
Here is the model file. You need to modify the suffix to onnx.
So if I want to use yolov3-tiny, is the darknet yolov3-tiny that exported to onnx supported? Or is there TI version of yolov3-tiny can be used, like yolov5?
