AuraFlow模型配置器:
AuraFlowModelNode prepares an AuraFlow model configuration for use elsewhere in a workflow. It collects the model locations and sequence-length setting into one configuration value and converts Windows-style absolute paths to paths suitable for a WSL environment. It does not load model files or generate images itself, so you can use it to organize your model setup before passing the configuration to a compatible downstream node.
AuraFlow模型配置器 Input Parameters:
transformer_path
The full path to the AuraFlow transformer model file. This required STRING input defaults to an empty string; no minimum or maximum length is specified. It is the one path the node requires to be nonblank: if you leave it empty or enter only spaces, the node returns an error instead of a usable configuration. The node removes surrounding whitespace and normalizes supported Windows-style absolute paths, but it does not check whether the file exists.
text_encoder_path
The full path to the text encoder model file. This required STRING input defaults to an empty string; no minimum or maximum length is specified. Although the interface requires the input to be present, the node accepts a blank value and includes it as an empty path in the configuration. Enter the path that matches your intended model setup when a downstream component needs a separate text encoder. The node removes surrounding whitespace and normalizes the path, but does not check the file.
vae_path
The full path to the VAE model file. This required STRING input defaults to an empty string; no minimum or maximum length is specified. A VAE is used in image workflows to convert between the model’s internal image representation and viewable images; this node only records its location for downstream use. A blank value is accepted and stored as an empty path. The node removes surrounding whitespace and normalizes the path, but does not verify that the file exists.
max_sequence_length
The sequence-length setting included in the AuraFlow configuration. This required INT input defaults to 768, with a minimum of 128, a maximum of 8192, and adjustments in steps of 128. The node’s tooltip identifies 768 for Pony-V7 and 256 for Base AuraFlow. Choose the value appropriate to your model: the node passes it through in the configuration, rather than checking model compatibility or processing text itself.
AuraFlow模型配置器 Output Parameters:
model_path
A single model_path output containing a configuration dictionary, not a file path string. On success, it contains type set to auraflow, the configured transformer_path, text_encoder_path, and vae_path, and max_sequence_length. The paths are trimmed and normalized where the supplied format is supported. If configuration fails, this same output contains an error entry instead, so an error value should not be treated as a ready-to-use model configuration.
AuraFlow模型配置器 Usage Tips:
- Enter the full transformer file path and check it carefully before running the workflow. This node requires a nonblank path but does not confirm that the file is present.
- Use
max_sequence_length768 for Pony-V7 or 256 for Base AuraFlow, as indicated by the node’s tooltip. Match this setting to the model you intend to use. - If your workflow runs in WSL, Windows-style absolute paths such as
C:\models\model.safetensorsare converted to paths such as/mnt/c/models/model.safetensors. Check that the resulting location is accessible from your runtime. - Supply text encoder and VAE paths when your downstream setup requires them. Leaving either blank is accepted by this node, but the resulting configuration will contain an empty path.
AuraFlow模型配置器 Common Errors and Solutions:
aura_flow_path不能为空
- Explanation: The transformer path is empty or contains only spaces. The message means the AuraFlow path cannot be blank.
- Solution: Enter the full transformer model file path in
transformer_path.
An error entry containing an exception message
- Explanation: The node catches an unexpected failure while preparing the configuration and returns the exception text in an
errorentry. The exact message depends on what went wrong. - Solution: Check that the path inputs contain text and that
max_sequence_lengthis a valid integer within the interface’s range. If the error persists, use the returned message to identify the failing value.
Downstream file-not-found or model-loading error
- Explanation: This node records and normalizes paths but does not verify files or load models. A path that is misspelled, inaccessible from WSL, or blank for a file required downstream can therefore fail later in the workflow.
- Solution: Confirm each needed file exists at the path visible to your runtime, and fill in
text_encoder_pathandvae_pathif your downstream setup requires them.
