XingangPan/DragGAN · error · ValueError

cannot parse 2-vector {s}

Error message

cannot parse 2-vector {s}

What it means

Raised by legacy.convert_tf_generator when converting an official TensorFlow StyleGAN/StyleGAN2 generator pickle whose internal version attribute is below 4. Only TF pickles of version >= 4 (StyleGAN2 or later configs) are supported by the PyTorch conversion path; older v1-v3 network classes have incompatible static_kwargs and params layout.

Source

Thrown at gen_images.py:53

        if m:
            ranges.extend(range(int(m.group(1)), int(m.group(2))+1))
        else:
            ranges.append(int(p))
    return ranges

#----------------------------------------------------------------------------

def parse_vec2(s: Union[str, Tuple[float, float]]) -> Tuple[float, float]:
    '''Parse a floating point 2-vector of syntax 'a,b'.

    Example:
        '0,1' returns (0,1)
    '''
    if isinstance(s, tuple): return s
    parts = s.split(',')
    if len(parts) == 2:
        return (float(parts[0]), float(parts[1]))
    raise ValueError(f'cannot parse 2-vector {s}')

#----------------------------------------------------------------------------

def make_transform(translate: Tuple[float,float], angle: float):
    m = np.eye(3)
    s = np.sin(angle/360.0*np.pi*2)
    c = np.cos(angle/360.0*np.pi*2)
    m[0][0] = c
    m[0][1] = s
    m[0][2] = translate[0]
    m[1][0] = -s
    m[1][1] = c
    m[1][2] = translate[1]
    return m

#----------------------------------------------------------------------------

@click.command()

View on GitHub (pinned to 336f120ce1)

Solutions

1. Use a StyleGAN2/StyleGAN2-ADA TensorFlow pickle with version >= 4 (the official stylegan2-ada-pytorch ones)
2. Re-export the TF network from the original stylegan2 repo ensuring version is set
3. If you must convert an old StyleGAN v1 network, write a custom converter instead of convert_tf_generator
4. Load pre-converted PyTorch .pkl checkpoints directly (already TorchScript/native) so no TF conversion runs

Example fix

# before
G = legacy.load_network_pickle('stylegan1-1024x1024.pkl')  # version < 4 -> ValueError

# after
G = legacy.load_network_pickle('stylegan2-ffhq-1024x1024.pkl')  # version >= 4 / native

import pickle
with open(tf_pkl, 'rb') as f: net = pickle.load(f)
if net.version < 4:
    raise ValueError('Needs a StyleGAN2-era (version>=4) TF pickle')

def is_convertible_tf_generator(tf_G) -> bool:
    return getattr(tf_G, 'version', 0) >= 4

try:
    G = legacy.convert_tf_generator(tf_G)
except ValueError as e:
    if 'version too low' in str(e):
        # source a v4 pickle or write a custom converter
        ...
    raise

When it happens

Trigger: Calling legacy.load_network_pickle on a TensorFlow .pkl from the original StyleGAN (v1/v2/v3) repo or an early experimental checkpoint instead of a StyleGAN2-era pickle; converting TF networks exported before the version field was standardized to 4.

Common situations: Trying to port weights from the original stylegun/stylegan repo (2019) or stylegan2 TensorFlow pickles with nonstandard versions; pickles re-saved by third-party tools that dropped/renamed the version attribute.

Related errors

• Unknown TensorFlow kwarg
• TensorFlow pickle version too low
• Cannot infer type name from input
• No data received
• Google Drive virus checker nag

AI-assisted analysis of XingangPan/DragGAN@336f120ce1 (2026-08-27). Data as JSON: /api/errors/bff3054846d66cbd. Report an issue: GitHub.