Neo Turbo 加速 (Forge Neo):
NeoTurbo prepares a model for a low-step Turbo workflow by applying a selected Turbo LoRA to the model and packaging sampling settings for a compatible downstream node. This keeps the model change and the intended sampler configuration together, making it easier to use a fast-generation setup consistently. NeoTurbo does not generate an image by itself: a downstream sampling node must use the prepared model and, where supported, the accompanying settings. The node description also mentions a group-level bypass, but the shown apply method only establishes what happens when NeoTurbo runs; bypass behavior depends on the surrounding workflow.
Neo Turbo 加速 (Forge Neo) Input Parameters:
model
The MODEL to which NeoTurbo applies the LoRA. Connect the model you intend to sample with downstream. There is no numeric minimum, maximum, or default; a model connection is required. NeoTurbo modifies the model through LoRA loading, while the sampling settings are returned separately rather than applied to an image here.
lora_name
The LoRA file to load. Its choices come from the files ComfyUI lists in its loras folder, so the exact options depend on your installation; no fixed default is specified in the source. The tooltip identifies the intended choice as the official Anima Turbo LoRA from circlestone-labs, Civitai 2560840. Select the file you mean to use, because this node passes the selected name to its LoRA-loading helper.
strength
Controls how strongly the LoRA is applied to the model. The default is 1.0, the minimum is -10.0, the maximum is 10.0, and the interface changes it in increments of 0.05. A value of 0 requests zero LoRA strength; negative values are allowed by the interface, but the provided source does not establish a useful artistic interpretation for them. The tooltip says a value slightly below 1 can increase variation, making that a reasonable setting to test when results feel too similar.
sampler_name
Selects the sampler name included in the turbo settings passed downstream. The default is Euler; the available options come from SAMPLER_NAMES and are not enumerated in the provided source. This selection does not run a sampler inside NeoTurbo. It matters when a compatible downstream node reads the returned settings.
scheduler
Selects the scheduler name included in the turbo settings. The default is Simple; the available options come from SCHEDULER_NAMES and are not enumerated in the provided source. Like sampler_name, this is a setting for downstream sampling, not an operation performed by NeoTurbo itself.
steps
Sets the step count stored under steps in the turbo settings. The default is 10, the minimum is 1, and the maximum is 100. The tooltip describes this as the text-to-image step count and notes an official recommendation of 8 to 12. Use it as the intended setting for a compatible downstream text-to-image sampling path; NeoTurbo itself does not execute those steps.
hires_steps
Sets the step count stored under hires_steps in the turbo settings. The default is 8, the minimum is 1, and the maximum is 100. The tooltip associates it with high-resolution refinement or image-to-image work. Its effect depends on whether the downstream node reads and uses this setting for that stage.
cfg
Sets the cfg value included in the turbo settings. The default is 1.0, the minimum is 0.0, the maximum is 10.0, and the interface changes it in increments of 0.1. CFG is a sampling guidance setting; NeoTurbo passes its chosen value onward without sampling an image itself. The tooltip recommends 1 and states that, at this value, negative prompts do not take effect in the intended workflow.
Neo Turbo 加速 (Forge Neo) Output Parameters:
MODEL
The input model after the selected LoRA has been loaded at the requested strength. Connect this output to the model input of your downstream sampling workflow. It is a MODEL value, not an image or a finished generation. The LoRA is loaded for the model only; the shown call does not apply it to a text encoder.
turbo
A NEO_TURBO settings value containing the selected sampler_name, scheduler, steps, hires_steps, and cfg. Connect it to a downstream node that accepts NEO_TURBO to pass along the intended fast-sampling configuration. This output is settings data, not a second model or an image, and it does not change a sampler unless a compatible downstream node uses it.
Neo Turbo 加速 (Forge Neo) Usage Tips:
- Start with the tooltip’s intended setup: the specified Anima Turbo LoRA,
strengthat1.0,Euler,Simple,stepsat10,hires_stepsat8, andcfgat1.0. Change one setting at a time so you can tell what affected the result. - For text-to-image testing, try
stepswithin the tooltip’s recommended8to12range before increasing it. For a high-resolution or image-to-image stage, adjusthires_stepsseparately if your downstream node uses that field. - If you want to explore more variation, test a
strengthslightly below1.0, as suggested by the tooltip. Keep the same prompt and other settings while comparing results. - Connect both outputs where your workflow requires them:
MODELsupplies the LoRA-applied model, whileturbosupplies the sampling settings. Connecting only the model does not, by itself, pass the chosen step count or CFG to a sampler. - At the recommended
cfgvalue of1.0, do not rely on a negative prompt to steer results in the intended workflow; the node tooltip says it has no effect at that setting.
Neo Turbo 加速 (Forge Neo) Common Errors and Solutions:
lora_name is not available in the dropdown
- Explanation: The choices come from ComfyUI’s listed
lorasfiles. If your intended file is absent, NeoTurbo cannot select it from this input. - Solution: Check that the LoRA file is in a location recognized as a ComfyUI
lorasfolder, then refresh the file list or restart ComfyUI and select it again.
LoRA file not found or failed to load
- Explanation: NeoTurbo passes
lora_nameto a LoRA-loading helper. A missing, unreadable, or invalid file can prevent the node from producing its outputs. The exact error wording depends on that helper and ComfyUI, not on a message defined in this class. - Solution: Confirm the selected file still exists and can be read. Reselect it from the dropdown; if it continues to fail, verify that you have a valid LoRA file intended for the model you connected.
MODEL input is missing or has an incompatible connection
- Explanation:
modelis required and must be a ComfyUIMODELvalue. Without one, the LoRA-loading call cannot run. - Solution: Connect a model-producing node to
modeland make sure the connection is to itsMODELoutput.
strength, steps, hires_steps, or cfg is outside its allowed range
- Explanation: The interface limits
strengthto-10.0through10.0, both step counts to1through100, andcfgto0.0through10.0. An out-of-range value can fail input validation before NeoTurbo runs. - Solution: Set the value within its listed range. To return to the provided starting settings, use
1.0forstrength,10forsteps,8forhires_steps, and1.0forcfg.
NEO_TURBO cannot connect to the next node
- Explanation: The
turbooutput has its ownNEO_TURBOtype; it is not a standardMODELor image connection. - Solution: Connect
turboonly to a downstream input that acceptsNEO_TURBO. Connect the separateMODELoutput to the downstream model input.
Sampler settings do not change the generated image
- Explanation: NeoTurbo returns sampler settings as data but does not run a sampler or directly change an unrelated sampler node. This can happen when the downstream sampling path does not consume
turbo. - Solution: Check that the compatible downstream node receives the
turbooutput and uses its settings. Also confirm that the sampling path receives NeoTurbo’sMODELoutput.
Negative prompt appears ineffective at cfg 1.0
- Explanation: The node’s tooltip explicitly states that negative prompts do not take effect at its recommended
cfgvalue of1. - Solution: Treat this as an expected limitation of the recommended setup. If negative-prompt control matters to your workflow, test a different supported CFG setting and evaluate the resulting images.
