Add GPU (NVENC AV1) webm format and save_latent toggle to FastAbsoluteSaver

- nvenc_av1-webm: GPU-encoded webm via av1_nvenc (NVENC has no VP9
  encoder); uses libopus since webm rejects AAC audio.
- save_latent boolean gates the latent sidecar write while keeping the
  latent passthrough intact.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

tests/test_fast_saver_latent.py

@@ -4,7 +4,7 @@ import torch
 from fast_saver import FastAbsoluteSaver
 
 
-def _save_args(tmp_path, *, save_format="png", latent=None):
+def _save_args(tmp_path, *, save_format="png", latent=None, save_latent=True):
     return {
         "images": torch.zeros((1, 2, 2, 3), dtype=torch.float32),
         "output_path": str(tmp_path),
@@ -18,6 +18,7 @@ def _save_args(tmp_path, *, save_format="png", latent=None):
         "metadata_key": "sharpness_score",
         "save_workflow_metadata": False,
         "save_metadata_png": False,
+        "save_latent": save_latent,
         "webp_lossless": True,
         "webp_quality": 100,
         "webp_method": 4,
@@ -55,6 +56,17 @@ def test_png_save_returns_latent_passthrough(tmp_path):
     assert result["result"][0] is latent
 
 
+def test_png_save_latent_false_skips_sidecar_but_keeps_passthrough(tmp_path):
+    saver = FastAbsoluteSaver()
+    latent = {"samples": torch.ones((1, 1, 2, 2))}
+
+    result = saver.save_images_fast(**_save_args(tmp_path, latent=latent, save_latent=False))
+
+    assert not (tmp_path / "frame_0000.latent").exists()
+    assert result["result"] == (latent,)
+    assert result["result"][0] is latent
+
+
 def test_video_save_writes_latent_sidecar_next_to_video(tmp_path):
     saver = FastAbsoluteSaver()
     latent = {"samples": torch.arange(8, dtype=torch.float32).reshape(2, 1, 2, 2)}