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

Unsupported fields:{} in cfg

Error message

Unsupported fields:{} in cfg

What it means

After parsing the .cfg into a list of dicts, parse_model_cfg checks every key in every module definition against a whitelist of supported fields. An unknown key (often a typo or a field from a newer/older YOLO cfg dialect) raises ValueError('Unsupported fields:{} in cfg').

Source

Thrown at pytorch_object_detection/yolov3_spp/build_utils/parse_config.py:56

            else:
                # TODO: .isnumeric() actually fails to get the float case
                if val.isnumeric():  # return int or float 如果是数值的情况
                    mdefs[-1][key] = int(val) if (int(val) - float(val)) == 0 else float(val)
                else:
                    mdefs[-1][key] = val  # return string  是字符的情况

    # check all fields are supported
    supported = ['type', 'batch_normalize', 'filters', 'size', 'stride', 'pad', 'activation', 'layers', 'groups',
                 'from', 'mask', 'anchors', 'classes', 'num', 'jitter', 'ignore_thresh', 'truth_thresh', 'random',
                 'stride_x', 'stride_y', 'weights_type', 'weights_normalization', 'scale_x_y', 'beta_nms', 'nms_kind',
                 'iou_loss', 'iou_normalizer', 'cls_normalizer', 'iou_thresh', 'probability']

    # 遍历检查每个模型的配置
    for x in mdefs[1:]:  # 0对应net配置
        # 遍历每个配置字典中的key值
        for k in x:
            if k not in supported:
                raise ValueError("Unsupported fields:{} in cfg".format(k))

    return mdefs


def parse_data_cfg(path):
    # Parses the data configuration file
    if not os.path.exists(path) and os.path.exists('data' + os.sep + path):  # add data/ prefix if omitted
        path = 'data' + os.sep + path

    with open(path, 'r') as f:
        lines = f.readlines()

    options = dict()
    for line in lines:
        line = line.strip()
        if line == '' or line.startswith('#'):
            continue
        key, val = line.split('=')

View on GitHub (pinned to 1ec3fe6f37)

Solutions

1. Open the cfg at the reported key's block and remove or rename the unsupported field to one the parser supports
2. Compare against the repo's bundled cfg files (cfg/yolov3-spp.cfg) and align your cfg with that dialect
3. Update parse_config.py's supported set if you intentionally need the new field

Example fix

// before (in .cfg)
[convolutional]
batch_normalize=1
size=3
channels_out=128
// after (in .cfg)
[convolutional]
batch_normalize=1
size=3
filters=128

supported = {'net','convolutional','maxpool','route','upsample','shortcut','yolo'}
import re
for block in re.findall(r'\[([^\]]+)\]', open(cfg).read()):
    assert block.strip() in supported, f'unsupported block [{block}] in {cfg}'

try:
    mdefs = parse_model_cfg(cfg)
except ValueError as e:
    if 'Unsupported fields' in str(e):
        bad_key = e.args[0].split(':')[1].strip()
        raise SystemExit(f'{cfg}: remove/rename unsupported key {bad_key}')
    raise

When it happens

Trigger: A [convolutional], [net], [yolo], [route], etc. block in the .cfg contains a key not in the supported set, e.g. misspelled 'batch_normalize' instead of 'batch_normalizing' variant, 'channels_out', or fields like 'groups'/'width_multiple' from other YOLO versions.

Common situations: Using a cfg copied from another YOLO repo (e.g. AlexeyAB/darknet cfgs with extra fields); hand-editing the cfg and typo-ing a key; upgrading the repo but keeping old cfgs.

Related errors

• the cfg file not exist...
• conv2d filter size must be int type.
• VOCdevkit dose not in path:'{}'.
• return_layers are not present in model
• VOCdevkit dose not in path:'{}'.

AI-assisted analysis of WZMIAOMIAO/deep-learning-for-image-processing@1ec3fe6f37 (2026-08-30). Data as JSON: /api/errors/ce04083be356d055. Report an issue: GitHub.