src/speech_decoding/models/v14_converged_v3/shaft_dataset.py

Selected training excerpt · lines 95–98 · source SHA-256 75d205f093e5

95        t0 = sample_clip_start(96            n_frames=s.n_frames, clip_frames=self.clip_frames, bad_spans_s=s.bad_spans_s,97            fps=self.fps, generator=g, start_align=self.start_align,98        )