pytorch-fanatics 0.2.7

A new library for Computer Vision

Pytorch_Fanatics

Pytorch_Fanatics is a Python library for ease of Computer Vision tasks.This contains a bunch of various tools which will help to create customized codes to train CV models.

###This library includes:

• Dataset class
• LRFinder
• EarlyStopping
• Trainer
• Logger

Installation

Use the package manager pip to install pytorch_fanatics.

pip install pytorch_fanatics

Usage

1)Dataset Class

from pytorch_fanatics.dataloader import Cloader

Cloader(image_path,targets,resize,transforms) # returns {'image':tensor_image,'target':tensor_labels}
#Use it to load images and labels

2)LRFinder

from pytorch_fanatics.utils import LRFinder

lr_finder=LRFinder(model,train_dataloader,optimizer,device,initial_lr,final_lr,beta) #Create a object
lr_finder.find()       #To find the lr
lr_finder.plot()       #To plot the graph (Loss V/S lr)
#NOTE:
#Use this only after training the model with all layers (except the last) freezed.

3)EarlyStopping

from pytorch_fanatics.utils import EarlyStop

es=EarlyStop(patience=7, mode="max", delta=0.0001) #Create a object
es(epoch_score, model, path)
if es.early_stop=True:
	break
es.reset() # to reset

4)Trainer

from pytorch_fanatics.trainer import Trainer

Trainer.train(model,data_loader,optimizer,device) # trains the model

score=Trainer.evaluate(model,data_loader,device,scheduler=None,metric=metrics.accuracy_score,plot=True)
#Use the score to feed for earlystop if used
#plot=True specifies live plot b/w (training and validation) vs num_epochs

Trainer.predict(model,data_loader,device) # returns probability of classes

Trainer.get_log() #returns a DataFrame object of logs 
#Currently the metrics supported are f1_score,accuracy,precision,recall,roc_auc_score and log_loss
#We are working on other too ..

5)Logger

from pytorch_fanatics.logger import Logger

Logger.save(model,optimizer,scheduler,path)  # saves model,optimizer and schedulers

#To load:
checkpoint=Logger.load(path)  #returns a dictionary
model,optimizer,scheduler=checkpoint['model'],checkpoint['optimizer'],checkpoint['scheduler']

#Helps keep track of training.It will restart from where it had stopped.

---

NOTE(Regarding model)

class Resnet18(nn.Module):
    def __init__(self):
        super(Resnet18, self).__init__()
        self.base_model =timm.create_model('resnet18',pretrained=True,num_classes=1)
    def forward(self, image, targets):
        batch_size, _, _, _ = image.shape
        out = self.base_model(image)
        loss = nn.BCEWithLogitsLoss()(out.view(batch_size,), targets.type_as(out))
        return out, loss
#Since every loss function has its own format of inputs,To generalise I have created this model.Use this model(edit if required) if you are #using Trainer/LRFinder.For others your simple model will also work fine..

---

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Please make sure to update tests as appropriate.

License

MIT

References

• FastAi Documentations for LRFinder.
• Pytorch Documentations.
• Abhishek Thakur Sir's Github Project wtfml(Awesome work Have a look at it too).