GPU Scheduling

VRAM-aware GPU scheduling for ML model orchestration.

Model Types and VRAM Configuration

use serde::{Deserialize, Serialize};

#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, Serialize, Deserialize)]
pub enum ModelType {
    Whisper = 0,    // 4.5 GB
    VideoMAE = 1,   // 5.0 GB
    CLAP = 2,       // 2.0 GB
    Qwen3 = 3,      // 1.5 GB
    DINOv3 = 4,     // 1.0 GB
}