Has any one run detect.py ?

[infrontofme]

Hi, I've trained the model following the instructions, after finetuning on novel classes, the AP results are as follows:

AP for aeroplane = 0.6602
AP for bicycle = 0.4766
AP for bird = 0.3573
AP for boat = 0.4778
AP for bottle = 0.3153
AP for bus = 0.2155
AP for car = 0.6892
AP for cat = 0.8144
AP for chair = 0.3652
AP for cow = 0.3928
AP for diningtable = 0.5538
AP for dog = 0.6887
AP for horse = 0.6958
AP for motorbike = 0.4364
AP for person = 0.6416
AP for pottedplant = 0.2839
AP for sheep = 0.5383
AP for sofa = 0.3712
AP for train = 0.7394
AP for tvmonitor = 0.6614
~~~~~~~~
Mean AP = 0.5187
Mean Base AP = 0.5734
Mean Novel AP = 0.3546

Then, I want to make predictions using pre-trained weights, but I found it in darknet_dynamic.cfg, classes=1. I modify classes=20 during inference in order to make predictions on VOC. but the result is worse.

how can I use pre-trained weight to make the right predictions?

thanks:)

[christegho]

Which script did you use to get the predictions for the image you posted?

The following command should work (you need to change the weights path and the image path):

python2 detect.py cfg/darknet_dynamic.cfg cfg/reweighting_net.cfg  backup/metatunetest15_novel0_neg0/000010.weights image_test.jpg

[infrontofme]

@christegho

python2 detect.py cfg/darknet_dynamic.cfg cfg/reweighting_net.cfg backup/metatunetest15_novel0_neg0/000010.weights image_test.jpg

This script doesn't work, the model in detect.py can't take cfg/reweightng_net as an input parameter.

I use the following script, but the result is worse.

python2 detect.py cfg/darknet_dynamic.cfg backup/metatunetest15_novel0_neg0/000010.weights image_test.jpg

before tested on images, do I need to generate reweighting vectors using training images ?

[kshitijagrwl]

detect.py will not work as you rightly mentioned @infrontofme

You will have to follow some methodology similar to valid_ensemble.py:

1. Generate dynamic_weights from model.meta_forward()
2. Generate output from model.detect_forward()
3. Generate bbox from get_region_boxes()

[infrontofme]

@kshitijagrwl Thanks, I get it now.