How to use _format_logs method in rester

Best Python code snippet using rester_python

train.py

Source:train.py

...17        self.model.to(self.device)18        self.loss.to(self.device)19        for metric in self.metrics:20            metric.to(self.device)21    def _format_logs(self, logs):22        str_logs = ['{} - {:.4}'.format(k, v) for k, v in logs.items()]23        s = ', '.join(str_logs)24        return s25    def batch_update(self, x, y):26        raise NotImplementedError27    def on_epoch_start(self):28        pass29    def run(self, dataloader):30        self.on_epoch_start()31        logs = {}32        loss_meter = AverageValueMeter()33        metrics_meters = {metric.__name__: AverageValueMeter() for metric in self.metrics}34        with tqdm(dataloader, desc=self.stage_name, file=sys.stdout, disable=not (self.verbose)) as iterator:35            for x, y in iterator:36                x, y = x.to(self.device), y.to(self.device)37                loss, y_pred = self.batch_update(x, y)38                # update loss logs39                loss_value = loss.cpu().detach().numpy()40                loss_meter.add(loss_value)41                loss_logs = {self.loss.__name__: loss_meter.mean}42                logs.update(loss_logs)43                # update metrics logs44                for metric_fn in self.metrics:45                    metric_value = metric_fn(y_pred, y).cpu().detach().numpy()46                    metrics_meters[metric_fn.__name__].add(metric_value)47                metrics_logs = {k: v.mean for k, v in metrics_meters.items()}48                logs.update(metrics_logs)49                if self.verbose:50                    s = self._format_logs(logs)51                    iterator.set_postfix_str(s)52        return logs53class TrainEpoch(Epoch):54    def __init__(self, model, loss, metrics, optimizer, device='cpu', verbose=True):55        super().__init__(56            model=model,57            loss=loss,58            metrics=metrics,59            stage_name='train',60            device=device,61            verbose=verbose,62        )63        self.optimizer = optimizer64    def on_epoch_start(self):65        self.model.train()66    def batch_update(self, x, y):67        self.optimizer.zero_grad()68        prediction = self.model.forward(x)69        loss = self.loss(prediction, y)70        loss.backward()71        self.optimizer.step()72        return loss, prediction73    74class TrainEpochCustom(Epoch):75    def __init__(self, model, loss, metrics, optimizer, device='cpu', verbose=True):76        super().__init__(77            model=model,78            loss=loss,79            metrics=metrics,80            stage_name='train',81            device=device,82            verbose=verbose,83        )84        self.optimizer = optimizer85        self.loss_clf = nn.BCEWithLogitsLoss()86    def run(self, dataloader):87        self.on_epoch_start()88        logs = {}89        loss_meter = AverageValueMeter()90        metrics_meters = {metric.__name__: AverageValueMeter() for metric in self.metrics}91        with tqdm(dataloader, desc=self.stage_name, file=sys.stdout, disable=not (self.verbose)) as iterator:92            for x, y in iterator:93                x, y = x.to(self.device), (y[0].to(self.device), y[1].to(self.device))94                loss, y_pred = self.batch_update(x, y)95                # update loss logs96                loss_value = loss.cpu().detach().numpy()97                loss_meter.add(loss_value)98                loss_logs = {self.loss.__name__: loss_meter.mean}99                logs.update(loss_logs)100                # update metrics logs101                for metric_fn in self.metrics:102                    metric_value = metric_fn(y_pred, y[0]).cpu().detach().numpy()103                    metrics_meters[metric_fn.__name__].add(metric_value)104                metrics_logs = {k: v.mean for k, v in metrics_meters.items()}105                logs.update(metrics_logs)106                if self.verbose:107                    s = self._format_logs(logs)108                    iterator.set_postfix_str(s)109        return logs110    def on_epoch_start(self):111        self.model.train()112    def batch_update(self, x, y):113        self.optimizer.zero_grad()114        prediction_mask, prediction_clf = self.model.forward(x)115        y_mask, y_clf = y116        loss = self.loss(prediction_mask, y_mask) + self.loss_clf(prediction_clf, y_clf) * 0.5117        loss.backward()118        self.optimizer.step()119        return loss, prediction_mask120class ValidEpochCustom(Epoch):121    def __init__(self, model, loss, metrics, device='cpu', verbose=True):122        super().__init__(123            model=model,124            loss=loss,125            metrics=metrics,126            stage_name='valid',127            device=device,128            verbose=verbose,129        )130        self.loss_clf = nn.BCEWithLogitsLoss()131    def on_epoch_start(self):132        self.model.eval()133    def batch_update(self, x, y):134        y_mask, y_clf = y135        with torch.no_grad():136            prediction_mask, prediction_clf = self.model.forward(x)137            loss = self.loss(prediction_mask, y_mask) + self.loss_clf(prediction_clf, y_clf) * 0.5138        return loss, prediction_mask139    def run(self, dataloader):140        self.on_epoch_start()141        logs = {}142        loss_meter = AverageValueMeter()143        metrics_meters = {metric.__name__: AverageValueMeter() for metric in self.metrics}144        with tqdm(dataloader, desc=self.stage_name, file=sys.stdout, disable=not (self.verbose)) as iterator:145            for x, y in iterator:146                x, y = x.to(self.device), (y[0].to(self.device), y[1].to(self.device))147                loss, y_pred = self.batch_update(x, y)148                # update loss logs149                loss_value = loss.cpu().detach().numpy()150                loss_meter.add(loss_value)151                loss_logs = {self.loss.__name__: loss_meter.mean}152                logs.update(loss_logs)153                # update metrics logs154                for metric_fn in self.metrics:155                    metric_value = metric_fn(y_pred, y[0]).cpu().detach().numpy()156                    metrics_meters[metric_fn.__name__].add(metric_value)157                metrics_logs = {k: v.mean for k, v in metrics_meters.items()}158                logs.update(metrics_logs)159                if self.verbose:160                    s = self._format_logs(logs)161                    iterator.set_postfix_str(s)162        return logs163class ValidEpoch(Epoch):164    def __init__(self, model, loss, metrics, device='cpu', verbose=True):165        super().__init__(166            model=model,167            loss=loss,168            metrics=metrics,169            stage_name='valid',170            device=device,171            verbose=verbose,172        )173    def on_epoch_start(self):174        self.model.eval()...

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.