85 lines
2.6 KiB
Python
85 lines
2.6 KiB
Python
|
import functions as aux
|
||
|
|
import statistics as st
|
||
|
|
# import pandas as pd
|
||
|
|
|
||
|
|
data_types = ['mni', 'cif']
|
||
|
|
model_types = ['fre', 'bay']
|
||
|
|
|
||
|
|
o_data_types = ['MNIST', 'CIFAR']
|
||
|
|
o_model_types = ['LeNet', 'BCNN']
|
||
|
|
|
||
|
|
gpu_ene_data = aux.load_pickle("energy_gpu.pkl")
|
||
|
|
cpu_ene_data = aux.load_pickle("energy_cpu.pkl")
|
||
|
|
|
||
|
|
gpu_exp_data = aux.load_pickle("result_gpu.pkl")
|
||
|
|
|
||
|
|
for data in data_types:
|
||
|
|
for model in model_types:
|
||
|
|
for size in range(1, 8):
|
||
|
|
gpu_ene_data[data][model][size] = \
|
||
|
|
list(
|
||
|
|
aux.split(gpu_ene_data[data][model][size], 100)
|
||
|
|
)
|
||
|
|
|
||
|
|
for data in data_types:
|
||
|
|
for model in model_types:
|
||
|
|
for size in range(1, 8):
|
||
|
|
cpu_ene_data[data][model][size] = \
|
||
|
|
list(
|
||
|
|
aux.split(cpu_ene_data[data][model][size], 100)
|
||
|
|
)
|
||
|
|
|
||
|
|
spl_ene_data = dict(gpu_ene_data)
|
||
|
|
for data in data_types:
|
||
|
|
for model in model_types:
|
||
|
|
for size in range(1, 8):
|
||
|
|
for i in range(0, 100):
|
||
|
|
spl_ene_data[data][model][size][i] = \
|
||
|
|
gpu_ene_data[data][model][size][i] +\
|
||
|
|
cpu_ene_data[data][model][size][i]
|
||
|
|
|
||
|
|
for data in data_types:
|
||
|
|
for model in model_types:
|
||
|
|
for size in range(1, 8):
|
||
|
|
for i in range(0, 100):
|
||
|
|
spl_ene_data[data][model][size][i] = \
|
||
|
|
sum(spl_ene_data[data][model][size][i])
|
||
|
|
|
||
|
|
for data in data_types:
|
||
|
|
for model in model_types:
|
||
|
|
for size in range(1, 8):
|
||
|
|
temp = []
|
||
|
|
for i in range(0, 100):
|
||
|
|
temp.append(
|
||
|
|
# st.mean(spl_ene_data[data][model][size][0:i+1])
|
||
|
|
sum(spl_ene_data[data][model][size][0:i+1])
|
||
|
|
)
|
||
|
|
spl_ene_data[data][model][size] = temp
|
||
|
|
|
||
|
|
eff_data = dict(gpu_ene_data)
|
||
|
|
for data in data_types:
|
||
|
|
for model in model_types:
|
||
|
|
for size in range(1, 8):
|
||
|
|
for i in range(0, 100):
|
||
|
|
eff_data[data][model][size][i] = \
|
||
|
|
(gpu_exp_data[data][model][size]['acc'][i] /
|
||
|
|
spl_ene_data[data][model][size][i]) * 100
|
||
|
|
|
||
|
|
|
||
|
|
for data, o_data in zip(data_types, o_data_types):
|
||
|
|
eff_data[o_data] = \
|
||
|
|
eff_data.pop(data)
|
||
|
|
|
||
|
|
|
||
|
|
for o_data in o_data_types:
|
||
|
|
for model, o_model in zip(model_types, o_model_types):
|
||
|
|
eff_data[o_data][o_model] = \
|
||
|
|
eff_data[o_data].pop(model)
|
||
|
|
|
||
|
|
# mul = pd.MultiIndex.from_product([[1, 2, 3, 4, 5, 6, 7],
|
||
|
|
# ['bay', 'fre'], ['cif', 'mni']])
|
||
|
|
|
||
|
|
# eff_data = pd.DataFrame(eff_data)
|
||
|
|
|
||
|
|
aux.save_pickle("efficiency_data.pkl", eff_data)
|