comfyui-spawner-nodes

AllOnesMaskGenerator

Creates an int32 all-ones mask using the input tensor's batch size and the selected sequence dimension.

Spawner/Tensor · 1 input · 1 parameter · 1 output

ConditioningConcatenation

Concatenates two conditioning embeddings along dimension 1 and combines their pooled outputs and attention masks when available.

Spawner/Conditioning · 2 inputs · 0 parameters · 1 output

ConditioningCrossAttention

Applies multihead attention using one conditioning input as query and another as key and value, with an optional residual connection.

Spawner/Conditioning · 2 inputs · 2 parameters · 1 output

ConditioningInspector

Extracts the conditioning tensor and details, reporting nested detail keys, pooled output, and an attention mask when present.

Spawner/Conditioning · 1 input · 0 parameters · 5 outputs

ConditioningPacker

Packages an embedding tensor with optional pooled embedding, attention mask, and details into ComfyUI CONDITIONING format.

Spawner/Conditioning · 4 inputs · 0 parameters · 1 output

ConditioningPooledMerge

Pools two conditioning embeddings, using attention masks when available, then merges them with concat, add, mean, or max.

Spawner/Conditioning · 2 inputs · 1 parameter · 1 output

ImageFromBase64

Decodes a base64 image string, including a data URI prefix, and returns an RGB image tensor and alpha mask.

Spawner/Utils · 0 inputs · 1 parameter · 2 outputs

ImageMaskToBase64

Encodes an image as PNG base64, optionally adding a supplied mask as alpha and a data URI prefix.

Spawner/Utils · 2 inputs · 1 parameter · 1 output

ImageMetadataReader

Reads an input image and returns JSON with its filename, dimensions, available EXIF data, and PNG text chunks.

Spawner/Utils · 0 inputs · 1 parameter · 1 output

json_process

Parses a JSON string and returns up to four selected values using key and list-index paths.

Spawner/Utils · 1 input · 4 parameters · 4 outputs

SaveTrainingDataPair

Saves an image and conditioning data under a root path and style name, including pooled output and attention mask files when present.

Spawner/Dataset · 2 inputs · 2 parameters · 1 output

TensorAttentionFusion

Concatenates two tensors along the sequence dimension, applies multihead self-attention, and scales the output by temperature.

Spawner/Tensor · 2 inputs · 2 parameters · 1 output

TensorConcatenation

Concatenates two tensors along a selected dimension and can return debug information about the concatenation.

Spawner/Tensor · 2 inputs · 2 parameters · 2 outputs

TensorCrossAttention

Applies multihead attention to query, key, and value tensors with an optional key mask and residual connection.

Spawner/Tensor · 5 inputs · 2 parameters · 1 output

TensorInspector

Returns a tensor's shape, dtype, and device along with minimum, maximum, and mean values.

Spawner/Tensor · 1 input · 0 parameters · 6 outputs

TensorPooledMerge

Pools one or two tensors along a selected dimension, using optional masks, and merges the pooled results with concat, add, mean, or max.

Spawner/Tensor · 4 inputs · 2 parameters · 1 output

TensorShapeAdapter

Compact description unavailable for this node.

Spawner/Tensor · 2 inputs · 0 parameters · 1 output

TensorUnsqueeze

Compact description unavailable for this node.

Spawner/Tensor · 1 input · 1 parameter · 1 output

TextEncoderDecoder

Compact description unavailable for this node.

Spawner/Utils · 0 inputs · 6 parameters · 1 output

XMLJSONConverter

Compact description unavailable for this node.

Spawner/Utils · 0 inputs · 3 parameters · 1 output