sgl-project/sglang · error · ValueError

fl2va requires first_frame, last_frame, or both

Error message

fl2va requires first_frame, last_frame, or both

What it means

Raised by the ComfyUI SGLDiffusion MiniMax-H3 generate node when the task is 'fl2va' (frame-to-video) but neither first_frame nor last_frame is connected. The node pre-validates wiring the server would reject anyway, so the user is told exactly which input to fix.

Source

Thrown at python/sglang/multimodal_gen/apps/ComfyUI_SGLDiffusion/nodes.py:742

                {
                    "type": "video",
                    "uri": self._material_uri(reference_video),
                    "role": "reference",
                }
            )
        if reference_audio:
            conditions.append(
                {
                    "type": "audio",
                    "uri": self._material_uri(reference_audio),
                    "role": "reference",
                }
            )

        # 2. reject wiring the server would reject anyway, but name the input
        #    the user has to change
        if task == "fl2va" and not (first_frame is not None or last_frame is not None):
            raise ValueError("fl2va requires first_frame, last_frame, or both")
        if task == "ref2va" and not conditions:
            raise ValueError(
                "ref2va requires at least one of reference_image, "
                "reference_video, or reference_audio"
            )
        if task == "t2va" and conditions:
            raise ValueError("t2va takes no conditioning inputs; pick another task")

        # 3. `target` resolves the aligned canvas and frame count; the `size`
        #    the server API always sends is unused by H3
        extra_fields = {
            "task": task,
            "conditions": conditions,
            "target": {
                "short_edge": short_edge,
                "aspect_ratio": aspect_ratio,
                "duration_seconds": duration_seconds,
            },

View on GitHub (pinned to 0132848349)

Solutions

1. Connect at least one of first_frame or last_frame to an image node
2. If you only have text, switch task to 't2va'
3. If you have references instead of frames, use task 'ref2va' and wire reference inputs

Example fix

// before
task='fl2va', first_frame=None, last_frame=None
// after
task='fl2va', first_frame=load_image_node_output, last_frame=None

if task == 'fl2va' and first_frame is None and last_frame is None:
    raise UserError('wire first_frame or last_frame for fl2va')

try: node.generate(...) except ValueError as e: show_user(str(e))

When it happens

Trigger: Calling the ComfyUI node generate() with task='fl2va' while first_frame=None and last_frame=None (both IMAGE inputs left unconnected).

Common situations: Selecting the fl2va task in the ComfyUI dropdown but forgetting to wire a start/end frame; switching tasks from t2va to fl2va without updating the graph inputs.

Related errors

• ref2va requires at least one of reference_image, reference_v
• t2va takes no conditioning inputs; pick another task
• Failed to generate MiniMax-H3 video: {str(e)}
• img_position_ids must be [1, S, 3], got {list(img_position_i
• MiniMax H3 model variant must be a non-empty string

AI-assisted analysis of sgl-project/sglang@0132848349 (2026-08-28). Data as JSON: /api/errors/96ff7b1954584c16. Report an issue: GitHub.