hacksider/Deep-Live-Cam · critical · FileNotFoundError

Model file not found: {model_path}

Error message

Model file not found: {model_path}

What it means

FileNotFoundError raised in get_enhancer (modules/processors/frame/face_enhancer_gpen512.py) when the GPEN 512px ONNX model is still missing after the automatic download attempt. Identical logic to the 256px variant: check file, conditional_download from MODEL_URL, re-check, raise with the resolved path if absent. The 512px model is a separate, larger file with its own MODEL_FILE/MODEL_URL constants in this module.

Source

Thrown at modules/processors/frame/face_enhancer_gpen512.py:62

def pre_start() -> bool:
    if not is_image(modules.globals.target_path) and not is_video(modules.globals.target_path):
        update_status("Select an image or video for target path.", NAME)
        return False
    return True


def get_enhancer() -> Any:
    global ENHANCER
    with THREAD_LOCK:
        if ENHANCER is None:
            model_path = os.path.join(models_dir, MODEL_FILE)
            if not os.path.exists(model_path):
                from modules.utilities import conditional_download
                conditional_download(models_dir, [MODEL_URL])
            if not os.path.exists(model_path):
                raise FileNotFoundError(f"Model file not found: {model_path}")
            print(f"{NAME}: Loading ONNX model from {model_path}")
            ENHANCER = create_onnx_session(model_path)
            warmup_session(ENHANCER)
            print(f"{NAME}: Model loaded successfully.")
    return ENHANCER


def enhance_face(temp_frame: Frame, face: Face) -> Frame:
    try:
        session = get_enhancer()
    except Exception as e:
        print(f"{NAME}: {e}")
        return temp_frame
    try:
        return enhance_face_onnx(temp_frame, face, session, INPUT_SIZE)
    except Exception as e:
        print(f"{NAME}: Error during face enhancement: {e}")
        return temp_frame

View on GitHub (pinned to 987f6b392b)

Solutions

1. Manually download the 512px model using the MODEL_URL defined in face_enhancer_gpen512.py and place it under the exact MODEL_FILE name in the models dir.
2. If auto-download timed out on the large file, retry on a stable connection or use a resumable downloader; verify final file size.
3. Confirm the models directory is writable and has space for the larger weights.
4. Pre-seed the model in deployment images so runtime download is never needed.

Example fix

# before
# auto-download of the large 512 model fails behind a slow proxy
session = get_enhancer()

# after
# pre-download out-of-band and mount the models dir read-only:
# curl -L --retry 3 -o <models_dir>/<MODEL_FILE> <MODEL_URL>
session = get_enhancer()

import os, urllib.request
model_path = os.path.join(models_dir, MODEL_FILE)  # MODEL_FILE from face_enhancer_gpen512
if not os.path.exists(model_path):
    os.makedirs(models_dir, exist_ok=True)
    urllib.request.urlretrieve(MODEL_URL, model_path)
session = get_enhancer()

try:
    session = get_enhancer()
except FileNotFoundError as e:
    raise SystemExit(
        f"Model download failed: {e}. Fetch {MODEL_URL} into {models_dir} manually."
    )

When it happens

Trigger: Same as the 256 variant but for the 512px weights: offline/blocked network during conditional_download; the 512 model's URL dead or moved; models dir unwritable or full; the 256 model downloaded fine but the 512 download failed partway (larger file, more exposed to timeouts).

Common situations: Environments that allow the smaller 256 model through but time out on the larger 512 file; selectively cached model dirs where only some variants were pre-seeded; proxy size limits; container images pre-built with only the 256 model.

Related errors

• Model file not found: {model_path}
• {NAME}: Model not found at {model_path}
• {NAME}: Failed to initialize GFPGAN ONNX session. Check logs
• embeddings must not be empty
• max_k must be at least 1

AI-assisted analysis of hacksider/Deep-Live-Cam@987f6b392b (2026-08-14). Data as JSON: /api/errors/c9d9deb8ee97236f. Report an issue: GitHub.