`fastvideo.eval.metrics.common.ssim.metric.BaseMetric`

BaseMetric()

Abstract base class for all eval metrics.

Two execution shapes:

Per-sample (is_set_metric=False, default) — implement :meth:compute. The Evaluator calls it once per input sample and returns one :class:MetricResult per sample.
Set-vs-set (is_set_metric=True) — implement :meth:accumulate (called once per sample to buffer features) and :meth:finalize (called once after all samples to compute the corpus-level result). Use :meth:reset to clear buffers and :meth:merge_from to fold multi-GPU per-worker state together.

Optionally override :meth:setup to eagerly load models. Metrics that chunk along the time dim for memory hardcode their own chunk size in __init__ (see optical_flow for the canonical example). Eval always processes one video per :meth:Evaluator.evaluate call; compute / accumulate receive a single sample, not a batch.

`fastvideo.eval.metrics.common.ssim.metric.BaseMetric.to`

to(device: str | torch.device) -> BaseMetric

Move metric (and its internal models) to device.

`fastvideo.eval.metrics.common.ssim.metric.BaseMetric.setup`

setup() -> None

Eagerly load models. Called once by :class:EvalWorker.

Default is a no-op; metrics with no eager state (pixel math, closed-form ops) inherit this. Override only if your metric needs to load weights.

`fastvideo.eval.metrics.common.ssim.metric.BaseMetric.compute`

compute(sample: dict) -> MetricResult

Per-sample metrics: compute the score for one sample.

sample["video"] is (T, C, H, W) float in [0, 1]. sample["reference"] (if used) has the same shape. Return self._skip(sample, reason) for missing inputs.

`fastvideo.eval.metrics.common.ssim.metric.BaseMetric.reset`

reset() -> None

Clear accumulator state at the start of each evaluate() call.

`fastvideo.eval.metrics.common.ssim.metric.BaseMetric.accumulate`

accumulate(sample: dict) -> None

Buffer per-sample features for a corpus-level metric.

`fastvideo.eval.metrics.common.ssim.metric.BaseMetric.finalize`

finalize() -> MetricResult

Compute the corpus-level result from buffered state.

`fastvideo.eval.metrics.common.ssim.metric.BaseMetric.merge_from`

merge_from(other: BaseMetric) -> None

Multi-GPU: fold another worker's accumulator state into this one.

`fastvideo.eval.metrics.common.ssim.metric.register`

register(name: str)

Decorator to register a metric class.

Usage::

@register("ssim")
class SSIMMetric(BaseMetric):
    ...

`fastvideo.eval.metrics.common.ssim.metric.MetricResult`

MetricResult(name: str, score: float | None, details: dict[str, Any] = dict())

Standard result container returned by all metrics.

score is None when the metric was skipped (e.g. missing required input). Check details["skipped"] for the reason.

`fastvideo.eval.metrics.common.ssim.metric.SSIMMetric`

SSIMMetric(window_size: int = 11, chunk_size: int = 16)

Source

@register("common.ssim")
class SSIMMetric(BaseMetric):
    name = "common.ssim"
    requires_reference = True
    higher_is_better = True
    # Five depthwise conv2d's per chunk — GPU is ~100× the CPU path at
    # 1080p, and it frees the host bus for the loader thread.
    needs_gpu = True

    def __init__(self, window_size: int = 11, chunk_size: int = 16) -> None:
        super().__init__()
        self.window_size = window_size
        # Caps the ~5-7 conv intermediates at ~4 GB on a 1080p clip.
        self._chunk_size = chunk_size

    def compute(self, sample: dict) -> MetricResult:
        gen = sample["video"].float().to(self.device)  # (T, C, H, W)
        ref = sample["reference"].float().to(self.device)
        n = min(gen.shape[0], ref.shape[0])
        gen, ref = gen[:n], ref[:n]

        chunk = self._chunk_size or n
        parts = []
        for i in range(0, n, chunk):
            parts.append(_ssim_per_frame(gen[i:i + chunk], ref[i:i + chunk], self.window_size))
        per_frame = torch.cat(parts)  # (n,)

        return MetricResult(
            name=self.name,
            score=per_frame.mean().item(),
            details={"per_frame": per_frame.tolist()},
        )

`fastvideo.eval.metrics.common.ssim.metric.SSIMMetric.compute`

compute(sample: dict) -> MetricResult

Source

def compute(self, sample: dict) -> MetricResult:
    gen = sample["video"].float().to(self.device)  # (T, C, H, W)
    ref = sample["reference"].float().to(self.device)
    n = min(gen.shape[0], ref.shape[0])
    gen, ref = gen[:n], ref[:n]

    chunk = self._chunk_size or n
    parts = []
    for i in range(0, n, chunk):
        parts.append(_ssim_per_frame(gen[i:i + chunk], ref[i:i + chunk], self.window_size))
    per_frame = torch.cat(parts)  # (n,)

    return MetricResult(
        name=self.name,
        score=per_frame.mean().item(),
        details={"per_frame": per_frame.tolist()},
    )