ComfyUI Node
Flux2TTRController
Inference node: load TTR + controller checkpoints and patch the model for dynamic layer routing.
Flux2TTRController
- model
- MODEL
◄ttr_checkpoint_path►
◄controller_checkpoint_path►
◄quality_speed0.500►
◄policy_modestochastic►
◄policy_temperature1.000►
Categoryadvanced/attention
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| model | MODEL | — | |
| ttr_checkpoint_path | STRING | Path to the Phase-1 TTR checkpoint (distilled attention layers). | |
| controller_checkpoint_path | STRING | Path to the Phase-2 controller checkpoint used for per-step routing decisions. | |
| quality_speed | FLOAT | 0.5000–1 | Quality/speed tradeoff. Internally mapped to controller threshold (0.1 + 0.8 * quality_speed). Lower values keep more layers on full attention (usually higher quality, lower speed). Higher values route more layers to TTR (usually faster, potentially lower quality). |
| policy_mode | COMBO | stochastic | Inference routing policy. 'stochastic' samples one controller mask per diffusion step (matching sigma-aware training behavior). 'threshold' uses deterministic thresholding from controller probabilities. |
| policy_temperature | FLOAT | 1.0000.001–10 | Sampling temperature used only when policy_mode='stochastic'. Lower values make decisions harder/more deterministic near the threshold; higher values make decisions softer/more random. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| MODEL | MODEL | — |