lllyasviel/Fooocus · error · NotImplementedError

{model_name} is not implemented.

Error message

{model_name} is not implemented.

What it means

facexlib's init_detection_model only supports two detection model names: 'retinaface_resnet50' and 'retinaface_mobile0.25'. Any other string raises NotImplementedError before any weights are loaded or downloaded.

Source

Thrown at extras/facexlib/detection/__init__.py:16

import torch
from copy import deepcopy

from extras.facexlib.utils import load_file_from_url
from .retinaface import RetinaFace


def init_detection_model(model_name, half=False, device='cuda', model_rootpath=None):
    if model_name == 'retinaface_resnet50':
        model = RetinaFace(network_name='resnet50', half=half, device=device)
        model_url = 'https://github.com/xinntao/facexlib/releases/download/v0.1.0/detection_Resnet50_Final.pth'
    elif model_name == 'retinaface_mobile0.25':
        model = RetinaFace(network_name='mobile0.25', half=half, device=device)
        model_url = 'https://github.com/xinntao/facexlib/releases/download/v0.1.0/detection_mobilenet0.25_Final.pth'
    else:
        raise NotImplementedError(f'{model_name} is not implemented.')

    model_path = load_file_from_url(
        url=model_url, model_dir='facexlib/weights', progress=True, file_name=None, save_dir=model_rootpath)

    # TODO: clean pretrained model
    load_net = torch.load(model_path, map_location=lambda storage, loc: storage, weights_only=True)
    # remove unnecessary 'module.'
    for k, v in deepcopy(load_net).items():
        if k.startswith('module.'):
            load_net[k[7:]] = v
            load_net.pop(k)
    model.load_state_dict(load_net, strict=True)
    model.eval()
    model = model.to(device)
    return model

View on GitHub (pinned to ae05379cc9)

Solutions

1. Use exactly 'retinaface_resnet50' or 'retinaface_mobile0.25'
2. Check the literal strings in extras/facexlib/detection/__init__.py if unsure which names this vendored version supports
3. Upgrade/patch facexlib locally if you genuinely need another detection backbone

Example fix

// before
init_detection_model('resnet50', half=True, device='cuda')

// after
init_detection_model('retinaface_resnet50', half=True, device='cuda')

SUPPORTED_DETECTION_MODELS = {'retinaface_resnet50', 'retinaface_mobile0.25'}
assert model_name in SUPPORTED_DETECTION_MODELS, f'use one of {SUPPORTED_DETECTION_MODELS}'

def is_supported_detection_model(name: str) -> bool:
    return name in {'retinaface_resnet50', 'retinaface_mobile0.25'}

try:
    det = init_detection_model(name)
except NotImplementedError:
    det = init_detection_model('retinaface_resnet50')  # safe default

When it happens

Trigger: Calling init_detection_model('retinaface_mobilenet') (wrong name), 'resnet50' (missing prefix), or a new architecture name that this vendored facexlib version never implemented.

Common situations: Copy-pasting model names from other facexlib/GFPGAN versions or READMEs; name drift between upstream facexlib releases and the vendored copy in extras/; users assuming any timm backbone name works.

Related errors

• network_name={network_name}
• {model_name} is not implemented.
• No paddings to do, output_size must be None or {}
• Not (0 <= inner_padding_factor <= 1.0)
• Not (outer_padding[0] < output_size[0] and outer_padding[1]

AI-assisted analysis of lllyasviel/Fooocus@ae05379cc9 (2026-08-15). Data as JSON: /api/errors/d7390af05c4d9060. Report an issue: GitHub.