segmentation_models
segmentation_models copied to clipboard
qubvel
Segmentation models with pretrained backbones. Keras and TensorFlow Keras.
I've trained a Unet model based on the [multiclass segmentation (camvid)](https://github.com/qubvel/segmentation_models/blob/master/examples/multiclass%20segmentation%20(camvid).ipynb) example and I've an issue trying to load it... I've used the following code and callbacks: ``` # create...
I am using the same training and validation data in this model: ``` sm.FPN( backbone, classes=classes, input_shape=input_shape, activation="sigmoid", encoder_freeze=true, ) ``` loss function and metrics are: `from segmentation_models.losses import bce_jaccard_loss`...
I recently compared your EfficientNet-B0 backbone with the state-of-art architecture and detected 4 layers were missing in block 5. Only subblock 5a, 5b and 5c appear when ploting the model...
Hello ! My use case is as follows: I'm working on images from the Cityscapes dataset that I'm trying to segment. I have a tensorflow train dataset with around 60...
I am working on MRI dataset where the input and mask is a greyscale image. How to train the model for this particular case? and also if I don't want...
if I have a massive datasets of images, and unable to fit all of them into memory, how do I load and train them by batches? my folder structure is:...
update to recent tensorflow and keras
This is the error I am getting when I am trying to import segmentation_models
Following along with the multiclass segmentation [example](https://github.com/qubvel/segmentation_models/blob/master/examples/multiclass%20segmentation%20(camvid).ipynb), the model performs fairly well on a custom dataset however when I try to use the mask on the original image using OpenCV's...
https://github.com/qubvel/segmentation_models/blame/e951c6747f75fa9e7240816d1c79dd2e66813123/segmentation_models/losses.py#L54 dice loss returns 1 - f_score so it is working fine but the commentary has an issue at line 59 L(tp, fp, fn) = \frac{(1 + \beta^2) \cdot tp}...
