CUDA VRAM Usage Estimator
100% Client-Side
Model Configuration
Cargar Ejemplo
Model Parameters (Billions, e.g. 7B)
Precision Quantization
16-bit (Half Precision / FP16)
8-bit Quantized (INT8)
4-bit Quantized (INT4 / NF4)
Estimate Memory Requirements
VRAM Budget Forecast
Press calculate to view estimation.