Make videos on the computer you already own. FreeVideo runs MiniMax H3 in as little as 8 GB of VRAM and 16 GB of RAM, and adapts its acceleration path to your hardware.
Frame preview tolerates a canvas without latent frames; editor route count includes request assets
Bowen Xue6d7f85ed5cfb
diff --git a/freevideo_engine/runtime.py b/freevideo_engine/runtime.pyindex b92bee5..6bc3298 100644--- a/freevideo_engine/runtime.py+++ b/freevideo_engine/runtime.py@@ -830,7 +830,7 @@ class Engine:**({'video_offset': continuation_offset} if continuation_offset else {}),**({'audio_offset': continuation_audio_offset} if continuation_audio_offset else {})) as prefix, \(watch.layers(self.transformer.transformer_blocks) if watch is not None else nullcontext()), \- frame_preview.observing(latent_frames=canvas['latent_frames'], frame_h=height // 32,+ frame_preview.observing(latent_frames=canvas.get('latent_frames'), frame_h=height // 32,frame_w=width // 32, step_offset=progress_offset,total=progress_total if progress_total is not None else active_steps):latents, audio = generate_latents(self.transformer, prompt, tags, frames,