Skip to content

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.

Default the media encoding base to the model root when --base is not given

Bowen Xuee4e7ed339e75
diff --git a/freevideo_engine/generate.py b/freevideo_engine/generate.py
index 243ddce..d5815e6 100644
--- a/freevideo_engine/generate.py
+++ b/freevideo_engine/generate.py
@@ -772,7 +772,8 @@ def _run(args):
'metrics': str(destination.with_suffix('.encoding.json')),
'diagnostic_output': str(destination)}
if media:
- encoding.update(media=media, geometry=canvas, base=profile['engine'].get('base'),
+ from .paths import base_path
+ encoding.update(media=media, geometry=canvas, base=profile['engine'].get('base') or str(base_path()),
reference_latent_cache=str(input_cache.root.parent.parent) if input_cache else None)
# Lets the encoder child read the video model while it computes, on a slow drive.
from .model_read_ahead import request_fields