Created
September 19, 2024 13:43
-
-
Save antonioanerao/39a72a200c2a23e31542bcd3e3cbfb95 to your computer and use it in GitHub Desktop.
Calcular VRAM necessária para Model
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
def calcular_vram(num_parametros, quantization_bits, overhead_factor=1.2): | |
memoria_modelo = (num_parametros * quantization_bits) / (8 * 10**9) | |
memoria_total = memoria_modelo * overhead_factor | |
return memoria_total | |
def main(): | |
try: | |
num_parametros = float(input("Número de parâmetros do modelo (em bilhões, ex: 8 para 8B): ")) * 10**9 | |
quantization_bits = int(input("Tamanho da quantização (quantization size em bits, ex: 4, 6, 8): ")) | |
overhead_input = input("Informe o fator de overhead (Padrão 1.2): ") | |
if overhead_input == "": | |
overhead_factor = 1.2 | |
else: | |
overhead_factor = float(overhead_input) | |
if overhead_factor == 0: | |
overhead_factor = 1 | |
vram_utilizada = calcular_vram(num_parametros, quantization_bits, overhead_factor) | |
print(f"\nVRAM necessária é: {vram_utilizada:.2f} GB") | |
except ValueError: | |
print("Insira números válidos.") | |
if __name__ == "__main__": | |
main() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment