MiniMax H3 First Last Frame LoRA Trainer

other

Train a MiniMax H3 LoRA on first/last/both keyframe signatures, teaching it to generate video with audio that starts on one image and lands on another.

Capabilities

No supported capabilities recorded.

Details

Model IDminimax/h3/flf2v/trainer
Provider fal
Creator MiniMax
Family—
License—
Statusactive
Typeother
Input modalitiesimage
Output modalities—
Open weight—
Parameters—
Architecture—
Context window—
Max input—
Max output—
Reasoning tokens—
Knowledge cutoff—
Release date
Sourceofficial

API

GET/v1/models/fal/minimax/h3/flf2v/trainer

MiniMax H3 First Last Frame LoRA Trainer

minimax/h3/flf2v/trainer

Train a MiniMax H3 LoRA on first/last/both keyframe signatures, teaching it to generate video with audio that starts on one image and lands on another.

Changes · 4 entries
minimax/h3/flf2v/trainercreateca7a400Nov 2, 2026, 03:02 AM
minimax/h3/flf2v/trainerupdate77b8547Sep 2, 2026, 06:42 AM
last_seen_at2026-08-25→2026-09-01
minimax/h3/flf2v/trainerupdatee3d7dedAug 31, 2026, 04:13 AM
last_seen_at2026-08-10→2026-08-25
minimax/h3/flf2v/trainercreate3cddff5Aug 13, 2026, 02:54 AM