float-increment-per-latent

Node for single-pass LoRA sweep

Per Sample LoRA Loader (Single Pass)

Inputs:

• model
• lora_name
• mode: range, range_step_by_batch, manual_values
• start, stop
• step (used only in range)
• direction (used only in range)
• manual_values (used only in manual_values)
• fallback_adapter_policy: controls non-fusible adapters such as LoKr

range_step_by_batch formula:

• uses the runtime latent batch from EmptySD3LatentImage
• if runtime batch is 1: value = start
• else: auto_step = (stop - start) / (runtime_batch - 1)

Examples:

• start=-4, stop=5, batch_size=10 -> step=1
• start=-10, stop=10, batch_size=10 -> step=2.222222...

Outputs:

• model (patched)
• weights_used (string preview)

Multiple Per Sample LoRA Loader (Single Pass) nodes can be stacked. Stacked LoRAs are aggregated into shared per-layer bypass hooks, and standard LoRA adapters are fused per layer into one larger rank projection where possible.

fallback_adapter_policy:

• error_on_slow_fallback: stop immediately if any adapter would use the slow runtime path.
• static_base_only: keep non-fusible adapters at the base strength and apply per-sample changes only to fusible LoRA adapters.
• allow_slow_fallback: exact per-sample behavior for non-fusible adapters, but can be very slow with large stacks.

Wiring

1. CheckpointLoaderSimple.model -> Per Sample LoRA Loader (Single Pass).model
2. Output model from this node -> KSampler.model
3. Set EmptySD3LatentImage.batch_size normally; range_step_by_batch reads the runtime batch during sampling.