model should predict some sort of result
enter image description here
I tried to see the prediction of YOLOv5 but at but at this stage it is not working
Related
I am trying to train a FCN model for segmentation. I have an array of train images of size (256,256,3) and another array of ground truth(masks) of size (256,256,3). My model summary is as below :
Model Structure
I am not sure how to train the model so i am using model.fit() to train it.
Model Compile and fit
But i am getting incompatible input image size error
Error
Can anyone, please help me to rectify the code and train the model. Thank you
I'm trying to object detection with yolov3 and yolov4 and i just want to train 1 class= person.
I'm using COCO datasets.
After training how i can tested my weight file? I want to test all "test dataset" not only 1 image.
I have train loss graphic but this is not enough.
I want to calculate accuracy, precision, recall etc.
I found this examples : https://darknet.gong.im/
./darknet detector test cfg/coco.data cfg/yolov3.cfg weights/yolov3.weights batch ./in_images/ ./out_images/ >./results.txt
Can you tell me how i can compare with success metric this two algorithm?
How can i test yolov3 and yolov4 weight files ?
There is map function for testing the model.
./darknet detector map cfg/coco.data cfg/yolov3.cfg weights/yolov3.weights
If you label your test dataset and give the path of it to the 'valid' field inside the data file, you can use map function over your dataset.
Also you can use -map flag while training to see the map results on the graph.
If you use -map flag while training darknet also saves best weight file so you don't need to find the best one.
I'm a beginner with some programming experince. I'm trying to train darkflow with my own dataset. I'm following these instructions.
https://github.com/thtrieu/darkflow
So far I have done the following steps.
installed darkflow and the relevant modules
created test images and made annotations (Pascal VOC).
https://ibb.co/y4HmtGz
https://ibb.co/GkxLshK
If I have understood correctly the darkflow training requires Pascal VOC?
My problem is that I don't know how to start the training. How can I start the training process and how can I test if the neuralnet is working? Am I supposed to get weights as a result of training?
You can choose to use pre-trained weights from here. Download cfg and weights.
Assuming you have darkflow installed, you can train your network like this:
flow --model cfg/<your-config-filename>.cfg --load bin/<filename>.weights --train --annotation train/Annotations --dataset train/Images --epoch 100 --gpu 1.0
If you want to train your network from scratch w/o using any pre-trained weights,
you can do this:
flow --model cfg/<your-config-filename>.cfg --train --annotation train/Annotations --dataset train/Images --epoch 100 --gpu 1.0
After the training starts, model checkpoints are saved inside ckpt directory. You can load latest checkpoint and test on sample images.
I have created text classification model(.pb) using tensorflow. Prediction is good.
Is it possible to check the sentence using for prediction is already used to train the model or not. I need to retrain the model when new sentence is given to model to predict.
I did some research and couldn't find a way to get the train data only with the pb file because that file only stores the features and not the actual train data(obviously),but if you have the dataset,then you can easily verify duh....
I don't think you can ever find the exact train data with only the trained model,cause the model only contains the features and not the actual train data
Hello tensorflow Community.
i am new in tesnsorflow , i use tensorflow to classify images now i work with cats_dogs dataset.
i want to save my model after training,and load it in an other program to predict other input
Is there a way to do that ?