Fleet Status
GPU fleet health and training activity — live from SSH
Machines Online
5 / 5
Active Training
5
Fleet VRAM
119 GB
67.2 GB used
GPU Efficiency
88%
Cart-01
reflect-cart-01
RTX 5080
16 GB
VRAM11.5G / 16G (72%)
GPU Utilization98%
GPU Temp64°C
CPU Load2.41 1.59 1.61
Uptime1 week, 3 days, 3 hours, 4 minutes
Training Active
Step 23800/60000
Train PSNR: 22.96 dB
LPIPS: 0.5098SSIM: 0.2993
Gaussians: 17,246
Speed: 5.9 it/s
Cart-02
reflect-cart-02
RTX 5080
16 GB
VRAM11.4G / 16G (72%)
GPU Utilization100%
GPU Temp50°C
CPU Load1.18 1.46 1.61
Uptime3 days, 17 hours, 10 minutes
Training Active
Step 24000/60000
Train PSNR: 23.23 dB
LPIPS: 0.5117SSIM: 0.3022
Gaussians: 16,297
Speed: 6.0 it/s
Raider
reflect-gpu
RTX 5090L
25 GB
VRAM21.2G / 24G (89%)
GPU Utilization100%
GPU Temp55°C
CPU Load2.67 3.09 3.23
Uptime1 week, 6 days, 38 minutes
Training Active
Step 21900/60000
Train PSNR: 22.56 dB
LPIPS: 0.4800SSIM: 0.2722
Gaussians: 44,500
Speed: 4.1 it/s
ML-01
ml-01
RTX 5090
32 GB
VRAM11.7G / 32G (37%)
GPU Utilization71%
GPU Temp59°C
CPU Load3.78 2.66 2.28
Uptime1 day, 1 hour, 41 minutes
Training Active
ML-02
ml-02
RTX 5090
32 GB
VRAM11.4G / 32G (36%)
GPU Utilization69%
GPU Temp63°C
CPU Load1.89 1.75 1.85
Uptime1 day, 1 hour, 41 minutes
Training Active