WZMIAOMIAO/deep-learning-for-image-processing · error · AttributeError

'{}' object has no attribute '{}'

Error message

'{}' object has no attribute '{}'

What it means

SmoothedValue meter container implements __getattr__ that raises AttributeError when the requested attribute is neither in self.meters nor self.__dict__. Python's default attribute resolution failure surfaces through this custom path, so accessing a misspelled or never-logged metric name raises this error.

Source

Thrown at pytorch_object_detection/retinaNet/train_utils/distributed_utils.py:161

class MetricLogger(object):
    def __init__(self, delimiter="\t"):
        self.meters = defaultdict(SmoothedValue)
        self.delimiter = delimiter

    def update(self, **kwargs):
        for k, v in kwargs.items():
            if isinstance(v, torch.Tensor):
                v = v.item()
            assert isinstance(v, (float, int))
            self.meters[k].update(v)

    def __getattr__(self, attr):
        if attr in self.meters:
            return self.meters[attr]
        if attr in self.__dict__:
            return self.__dict__[attr]
        raise AttributeError("'{}' object has no attribute '{}'".format(
            type(self).__name__, attr))

    def __str__(self):
        loss_str = []
        for name, meter in self.meters.items():
            loss_str.append(
                "{}: {}".format(name, str(meter))
            )
        return self.delimiter.join(loss_str)

    def synchronize_between_processes(self):
        for meter in self.meters.values():
            meter.synchronize_between_processes()

    def add_meter(self, name, meter):
        self.meters[name] = meter

    def log_every(self, iterable, print_freq, header=None):

View on GitHub (pinned to 1ec3fe6f37)

Solutions

1. Print type(self).meters keys to see registered metric names and fix the typo
2. Ensure the metric was added before access: metric_logger.meters['name'] = SmoothedValue()
3. Use getattr(metric, name, default) when metric presence is optional

Example fix

// before
print(metric_logger.lerning_rate)  # typo
// after
print(metric_logger.learning_rate)

available = list(metric_logger.meters.keys())
assert metric_name in available, f'{metric_name} not logged; have {available}'

def has_metric(logger, name: str) -> bool:
    return name in getattr(logger, 'meters', {}) or name in getattr(logger, '__dict__', {})

try:
    value = getattr(metric_logger, metric_name)
except AttributeError as e:
    value = None
    print(f'Metric missing: {e}')

When it happens

Trigger: Accessing metric.something on a MetricLogger where 'something' was never registered via meters, e.g. a typo like metric.losss or reading a metric before it was added with meters[key] = SmoothedValue().

Common situations: Typos in metric names during train loop logging; expecting a metric that only exists in another repo version; accessing attributes before first training step populated them.

Related errors

• '{}' object has no attribute '{}'
• '{}' object has no attribute '{}'
• not support data format '{self.data_format}'
• image: {} isn't RGB mode.
• Transformer input dimension should be divisible by head dime

AI-assisted analysis of WZMIAOMIAO/deep-learning-for-image-processing@1ec3fe6f37 (2026-08-30). Data as JSON: /api/errors/2f043e5e92d4da64. Report an issue: GitHub.