opendatalab/MinerU · error · TypeError

The mixer must be one of [Global, Local, Conv]

Error message

The mixer must be one of [Global, Local, Conv]

What it means

Raised in SVTR recognition blocks (Block/BlockAttn construction) when the mixer string for a stage is not 'Global', 'Local', or 'Conv'. Each block instantiates a GlobalAttention/LocalAttention or ConvMixer depending on this config value; the else branch raises TypeError.

Source

Thrown at mineru/model/utils/pytorchocr/modeling/backbones/rec_svtrnet.py:238

            self.norm1 = eval(norm_layer)(dim, eps=epsilon)
        else:
            self.norm1 = norm_layer(dim)
        if mixer == "Global" or mixer == "Local":
            self.mixer = Attention(
                dim,
                num_heads=num_heads,
                mixer=mixer,
                HW=HW,
                local_k=local_mixer,
                qkv_bias=qkv_bias,
                qk_scale=qk_scale,
                attn_drop=attn_drop,
                proj_drop=drop,
            )
        elif mixer == "Conv":
            self.mixer = ConvMixer(dim, num_heads=num_heads, HW=HW, local_k=local_mixer)
        else:
            raise TypeError("The mixer must be one of [Global, Local, Conv]")

        self.drop_path = DropPath(drop_path) if drop_path > 0.0 else Identity()
        if isinstance(norm_layer, str):
            self.norm2 = eval(norm_layer)(dim, eps=epsilon)
        else:
            self.norm2 = norm_layer(dim)
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp_ratio = mlp_ratio
        self.mlp = Mlp(
            in_features=dim,
            hidden_features=mlp_hidden_dim,
            act_layer=act_layer,
            drop=drop,
        )
        self.prenorm = prenorm

    def forward(self, x):
        if self.prenorm:

View on GitHub (pinned to 4fe4bde114)

Solutions

1. Set each stage's mixer to exactly 'Global', 'Local', or 'Conv' in the backbone config.
2. Check the mixers matrix shape and values against the shipped SVTR config for your model version.
3. Verify capitalization — matching is case-sensitive.

Example fix

# before
mixers: [['Conv','local','Global']]

# after
mixers: [['Conv','Local','Global']]

for stage in mixers:
    for m in stage:
        assert m in {"Global", "Local", "Conv"}, f"bad mixer: {m!r}"

When it happens

Trigger: A rec SVTR config where Backbone.mixers (list like [['Conv','Local','Global'],...]) contains a misspelled entry ('global', 'local_mixer', 'GA'); building rec_svtrnet directly with mixer='Attention'.

Common situations: Hand-edited PP-OCRv3/v4 rec YAMLs mixing mixer vocabulary from different model generations; case-sensitive typos.

Related errors

• mode[{model_name}_model] is not implemented!
• mode[{model_name}_model] is not implemented!
• Unsupported activation: {name}
• PPLCNetV4 {mode} model_size must be one of {list(config_dict
• not support limit type, image

AI-assisted analysis of opendatalab/MinerU@4fe4bde114 (2026-08-14). Data as JSON: /api/errors/aa04e1766a6d4a97. Report an issue: GitHub.