# this file is only used for continuous evaluation test!
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi
from kpi import DurationKpi
from kpi import AccKpi
each_step_duration_senta_card1 = DurationKpi('each_step_duration_senta_card1', 0.01, 0, actived=False)
train_loss_senta_card1 = CostKpi('train_loss_senta_card1', 0.05, 0, actived=True)
train_acc_senta_card1 = AccKpi('train_acc_senta_card1', 0.02, 0, actived=True)
each_step_duration_senta_card4 = DurationKpi('each_step_duration_senta_card4', 0.01, 0, actived=False)
train_loss_senta_card4 = CostKpi('train_loss_senta_card4', 0.05, 0, actived=True)
train_acc_senta_card4 = AccKpi('train_acc_senta_card4', 0.02, 0, actived=True)
tracking_kpis = [
each_step_duration_senta_card1,
train_loss_senta_card1,
train_acc_senta_card1,
each_step_duration_senta_card4,
train_loss_senta_card4,
train_acc_senta_card4,
]
def parse_log(log):
'''
This method should be implemented by model developers.
The suggestion:
####this file is only used for continuous evaluation test!
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
#### NOTE kpi.py should shared in models in some way!!!!
test_acc = AccKpi('test_acc', 0.001, 0, actived=True, desc="test acc")
test_cost = CostKpi('test_cost', 0.001, 0, actived=True, desc='test cost')
#train_speed_kpi = DurationKpi(
# 'train_speed',
# 0.05,
# 0,
# actived=True,
# unit_repr='seconds/image',
# desc='train speed in one GPU card')
tracking_kpis = [test_acc, test_cost]
def parse_log(log):
'''
This method should be implemented by model developers.
The suggestion:
each line in the log should be key, value, for example:
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import AccKpi
from kpi import DurationKpi
cifar10_128_AllReduce_GPU_4_Cards_train_acc_kpi = AccKpi(
'cifar10_128_AllReduce_GPU_4_Cards_train_acc', 0.03, 0, actived=True)
cifar10_128_AllReduce_GPU_4_Cards_train_speed_kpi = AccKpi(
'cifar10_128_AllReduce_GPU_4_Cards_train_speed', 0.06, 0, actived=False)
cifar10_128_AllReduce_4_Cards_gpu_memory_kpi = DurationKpi(
'cifar10_128_AllReduce_4_Cards_gpu_memory', 0.1, 0, actived=True)
cifar10_128_Reduce_GPU_4_Cards_train_acc_kpi = AccKpi(
'cifar10_128_Reduce_GPU_4_Cards_train_acc', 0.03, 0, actived=True)
cifar10_128_Reduce_GPU_4_Cards_train_speed_kpi = AccKpi(
'cifar10_128_Reduce_GPU_4_Cards_train_speed', 0.06, 0, actived=True)
cifar10_128_Reduce_4_Cards_gpu_memory_kpi = DurationKpi(
'cifar10_128_Reduce_4_Cards_gpu_memory', 0.1, 0, actived=True)
# Single Cards
cifar10_128_GPU_1_Cards_train_acc_kpi = AccKpi(
'cifar10_128_GPU_1_Cards_train_acc', 0.03, 0, actived=True)
cifar10_128_GPU_1_Cards_train_speed_kpi = AccKpi(
'cifar10_128_GPU_1_Cards_train_speed', 0.06, 0, actived=False)
cifar10_128_1_Cards_gpu_memory_kpi = DurationKpi(
'cifar10_128_1_Cards_gpu_memory', 0.1, 0, actived=True)
flowers_64_AllReduce_GPU_4_Cards_train_acc_kpi = AccKpi(
#!/usr/bin/env python
# -*- coding:utf-8 -*-
import os
import sys
import re
sys.path.append(os.environ['ceroot'])
from kpi import AccKpi
test_acc_top1_kpi = AccKpi('test_acc_top1',
0.02,
0,
actived=True,
desc='TOP1 ACC')
test_acc_top5_kpi = AccKpi('test_acc_top5',
0.02,
0,
actived=True,
desc='TOP5 ACC')
tracking_kpis = [test_acc_top1_kpi, test_acc_top5_kpi]
def parse_log(log):
'''
parse log
'''
pattern = r"^.*Final eval result: \['acc_top1', 'acc_top5'\]=\[(?P<test_acc_top1>0\.\d+)\s+(?P<test_acc_top5>0\.\d+)\s*\]"
prog = re.compile(pattern)
for line in log.split('\n'):
result = prog.match(line)
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import AccKpi
from kpi import DurationKpi
cifar10_8_AllReduce_CPU_4_Cards_train_acc_kpi = AccKpi(
'cifar10_8_AllReduce_CPU_4_Cards_train_acc', 0.02, 0, actived=True)
cifar10_8_AllReduce_CPU_4_Cards_train_speed_kpi = AccKpi(
'cifar10_8_AllReduce_CPU_4_Cards_train_speed', 0.06, 0, actived=True)
cifar10_8_Reduce_CPU_4_Cards_train_acc_kpi = AccKpi(
'cifar10_8_Reduce_CPU_4_Cards_train_acc', 0.02, 0, actived=True)
cifar10_8_Reduce_CPU_4_Cards_train_speed_kpi = AccKpi(
'cifar10_8_Reduce_CPU_4_Cards_train_speed', 0.06, 0, actived=True)
cifar10_8_CPU_1_Cards_train_acc_kpi = AccKpi(
'cifar10_8_CPU_1_Cards_train_acc', 0.02, 0, actived=True)
cifar10_8_CPU_1_Cards_train_speed_kpi = AccKpi(
'cifar10_8_CPU_1_Cards_train_speed', 0.06, 0, actived=True)
flowers_8_AllReduce_CPU_4_Cards_train_acc_kpi = AccKpi(
'flowers_8_AllReduce_CPU_4_Cards_train_acc', 0.02, 0, actived=True)
flowers_8_AllReduce_CPU_4_Cards_train_speed_kpi = AccKpi(
'flowers_8_AllReduce_CPU_4_Cards_train_speed', 0.06, 0, actived=True)
flowers_8_Reduce_CPU_4_Cards_train_acc_kpi = AccKpi(
'flowers_8_Reduce_CPU_4_Cards_train_acc', 0.02, 0, actived=True)
flowers_8_Reduce_CPU_4_Cards_train_speed_kpi = AccKpi(
# this file is only used for continuous evaluation test!
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi
from kpi import DurationKpi
from kpi import AccKpi
each_pass_duration_cpu1_thread1_kpi = DurationKpi(
'each_pass_duration_cpu1_thread1', 0.08, 0, actived=True)
train_loss_cpu1_thread1_kpi = CostKpi('train_loss_cpu1_thread1', 0.08, 0)
train_auc_val_cpu1_thread1_kpi = AccKpi('train_auc_val_cpu1_thread1', 0.08, 0)
train_batch_auc_val_cpu1_thread1_kpi = AccKpi(
'train_batch_auc_val_cpu1_thread1', 0.08, 0)
each_pass_duration_cpu1_thread8_kpi = DurationKpi(
'each_pass_duration_cpu1_thread8', 0.08, 0, actived=True)
train_loss_cpu1_thread8_kpi = CostKpi('train_loss_cpu1_thread8', 0.08, 0)
train_auc_val_cpu1_thread8_kpi = AccKpi('train_auc_val_cpu1_thread8', 0.08, 0)
train_batch_auc_val_cpu1_thread8_kpi = AccKpi(
'train_batch_auc_val_cpu1_thread8', 0.08, 0)
each_pass_duration_cpu8_thread8_kpi = DurationKpi(
'each_pass_duration_cpu8_thread8', 0.08, 0, actived=True)
train_loss_cpu8_thread8_kpi = CostKpi('train_loss_cpu8_thread8', 0.08, 0)
train_auc_val_cpu8_thread8_kpi = AccKpi('train_auc_val_cpu8_thread8', 0.08, 0)
train_batch_auc_val_cpu8_thread8_kpi = AccKpi(
'train_batch_auc_val_cpu8_thread8', 0.08, 0)
tracking_kpis = [
each_pass_duration_cpu1_thread1_kpi,
train_loss_cpu1_thread1_kpi,
# this file is only used for continuous evaluation test!
import os
import sys
from kpi import CostKpi, AccKpi
sys.path.append(os.environ['ceroot'])
# NOTE kpi.py should shared in models in some way!!!!
train_cost_kpi = CostKpi('train_cost', 0.02, 0, actived=True)
test_recall_kpi = AccKpi('test_recall', 0.02, 0, actived=True)
tracking_kpis = [
train_cost_kpi,
test_recall_kpi,
]
def parse_log(log):
'''
This method should be implemented by model developers.
The suggestion:
each line in the log should be key, value, for example:
"
train_cost\t1.0
test_cost\t1.0
train_cost\t1.0
train_cost\t1.0
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, AccKpi, DurationKpi
train_cost_kpi = CostKpi('train_cost', 0.05, 0, actived=True)
train_acc_kpi = AccKpi('train_acc', 0.02, 0, actived=True)
test_acc_kpi = AccKpi('test_acc', 0.05, 0, actived=True)
train_speed_kpi = AccKpi('train_speed', 0.01, 0, actived=True)
train_duration_kpi = DurationKpi('train_duration', 0.02, 0, actived=True)
tracking_kpis = [
train_cost_kpi,
train_acc_kpi,
test_acc_kpi,
train_speed_kpi,
train_duration_kpi,
]
####this file is only used for continuous evaluation test!
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
#### NOTE kpi.py should shared in models in some way!!!!
train_cost_kpi = CostKpi('train_cost', 0.02, 0, actived=True)
test_acc_kpi = AccKpi('test_acc', 0.01, 0, actived=False)
train_speed_kpi = DurationKpi('train_speed', 0.1, 0, actived=True, unit_repr="s/epoch")
train_cost_card4_kpi = CostKpi('train_cost_card4', 0.02, 0, actived=True)
test_acc_card4_kpi = AccKpi('test_acc_card4', 0.01, 0, actived=False)
train_speed_card4_kpi = DurationKpi('train_speed_card4', 0.1, 0, actived=True, unit_repr="s/epoch")
tracking_kpis = [
train_cost_kpi,
test_acc_kpi,
train_speed_kpi,
train_cost_card4_kpi,
test_acc_card4_kpi,
train_speed_card4_kpi,
]
def parse_log(log):
'''
This method should be implemented by model developers.
The suggestion:
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
p99_kpi = DurationKpi('99', 0.05, actived=True)
avg_kpi = DurationKpi('avg', 0.05, actived=True)
qps_kpi = AccKpi('qps', 0.05, actived=True)
#ok_kpi = AccKpi('ok', 0.05, actived=True)
tracking_kpis = [
p99_kpi,
avg_kpi,
qps_kpi,
]
####this file is only used for continuous evaluation test!
import os
import sys
sys.path.insert(0, os.environ['ceroot'])
#sys.path.append('.')
from kpi import CostKpi, DurationKpi, AccKpi
#### NOTE kpi.py should shared in models in some way!!!!
train_cost_xnli_card1_kpi = CostKpi('train_cost_xnli_card1',
0.002,
0,
actived=True)
train_acc_xnli_card1_kpi = AccKpi('train_acc_xnli_card1',
0.002,
0,
actived=True)
train_duration_xnli_card1_kpi = DurationKpi('train_duration_xnli_card1',
0.01,
0,
actived=True)
train_cost_xnli_card4_kpi = CostKpi('train_cost_xnli_card4',
0.002,
0,
actived=True)
train_acc_xnli_card4_kpi = AccKpi('train_acc_xnli_card4',
0.02,
0,
actived=True)
train_duration_xnli_card4_kpi = DurationKpi('train_duration_xnli_card4',
0.03,
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
cifar10_128_train_acc_kpi = AccKpi('cifar10_128_train_acc', 0.05, 0)
cifar10_128_train_speed_kpi = AccKpi('cifar10_128_train_speed', 0.05, 0)
cifar10_128_gpu_memory_kpi = DurationKpi('cifar10_128_gpu_memory', 0.01, 0)
flowers_64_train_speed_kpi = AccKpi('flowers_64_train_speed', 0.05, 0)
flowers_64_gpu_memory_kpi = DurationKpi('flowers_64_gpu_memory', 0.01, 0)
tracking_kpis = [
cifar10_128_train_acc_kpi,
cifar10_128_train_speed_kpi,
cifar10_128_gpu_memory_kpi,
flowers_64_train_speed_kpi,
flowers_64_gpu_memory_kpi,
]
"""
continuous_evaluation.py
"""
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import AccKpi
from kpi import DurationKpi
train_acc_kpi = AccKpi('train_acc', 0.2, 0)
pass_duration_kpi = DurationKpi('pass_duration', 0.02, 0, actived=True)
tracking_kpis = [
train_acc_kpi,
pass_duration_kpi,
]
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
train_acc_kpi = AccKpi('train_acc', 0.05)
test_acc_kpi = AccKpi('test_acc', 0.05)
train_duration_kpi = DurationKpi('train_duration', 0.1)
tracking_kpis = [
train_acc_kpi,
test_acc_kpi,
train_duration_kpi,
]
# this file is only used for continuous evaluation test!
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi
from kpi import DurationKpi
from kpi import AccKpi
each_pass_duration_cpu1_thread1_kpi = DurationKpi(
'each_pass_duration_cpu1_thread1', 0.08, 0, actived=True)
train_recall_cpu1_thread1_kpi = AccKpi('train_recall_cpu1_thread1', 0.08, 0)
each_pass_duration_gpu1_kpi = DurationKpi('each_pass_duration_gpu1',
0.08,
0,
actived=True)
train_recall_gpu1_kpi = AccKpi('train_recall_gpu1', 0.08, 0)
each_pass_duration_gpu4_kpi = DurationKpi('each_pass_duration_gpu4',
0.08,
0,
actived=True)
train_recall_gpu4_kpi = AccKpi('train_recall_gpu4', 0.08, 0)
tracking_kpis = [
each_pass_duration_cpu1_thread1_kpi,
train_recall_cpu1_thread1_kpi,
each_pass_duration_gpu1_kpi,
train_recall_gpu1_kpi,
each_pass_duration_gpu4_kpi,
train_recall_gpu4_kpi,
]
####this file is only used for continuous evaluation test!
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
#### NOTE kpi.py should shared in models in some way!!!!
train_acc_kpi = AccKpi('train_precision', 0.005, 0, actived=True)
test_acc_kpi = CostKpi('test_precision', 0.005, 0, actived=True)
train_duration_kpi = DurationKpi('train_duration', 0.05, 0, actived=True)
tracking_kpis = [
train_acc_kpi,
test_acc_kpi,
train_duration_kpi,
]
def parse_log(log):
for line in log.split('\n'):
fs = line.strip().split('\t')
print(fs)
if len(fs) == 3 and fs[0] == 'kpis':
print("-----%s" % fs)
kpi_name = fs[1]
kpi_value = float(fs[2])
yield kpi_name, kpi_value
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
p99_kpi = DurationKpi('99', 0.2, actived=True)
avg_kpi = DurationKpi('avg', 0.2, actived=True)
qps_kpi = AccKpi('qps', 0.2, actived=True)
err_kpi = AccKpi('err', 0.2, actived=True)
tracking_kpis = [
p99_kpi,
avg_kpi,
qps_kpi,
err_kpi
]
"""
continuous_evaluation.py
"""
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import AccKpi
from kpi import CostKpi
from kpi import DurationKpi
wmb_128_train_speed_kpi = AccKpi('wmb_128_train_speed', 0.2, 0)
wmb_128_gpu_memory_kpi = DurationKpi('wmb_128_gpu_memory', 0.2, 0)
tracking_kpis = [
wmb_128_train_speed_kpi,
wmb_128_gpu_memory_kpi,
]
"""
continuous_evaluation.py
"""
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import AccKpi
from kpi import CostKpi
from kpi import DurationKpi
imdb_32_train_speed_kpi = AccKpi('imdb_32_train_speed', 0.08, 0, actived=True)
imdb_32_gpu_memory_kpi = DurationKpi('imdb_32_gpu_memory', 0.05, 0, actived=True)
tracking_kpis = [
imdb_32_train_speed_kpi,
imdb_32_gpu_memory_kpi,
]
from __future__ import division
from __future__ import absolute_import
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi
from kpi import AccKpi
conv_train_cost_kpi = CostKpi('conv_train_cost',
0.02,
0,
actived=True,
desc='train cost')
conv_train_acc_kpi = AccKpi('conv_train_acc',
0.02,
0,
actived=True,
desc='train acc')
conv_test_cost_kpi = CostKpi('conv_test_cost',
0.02,
0,
actived=True,
desc='test cost')
conv_test_acc_kpi = AccKpi('conv_test_acc',
0.02,
0,
actived=True,
desc='test acc')
rnn_train_cost_kpi = CostKpi('rnn_train_cost',
0.02,
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
train_acc_top1_kpi = AccKpi('train_acc_top1_kpi',
0.05,
0,
actived=True,
desc='TOP1 ACC')
train_acc_top5_kpi = AccKpi('train_acc_top5_kpi',
0.05,
0,
actived=True,
desc='TOP5 ACC')
train_cost_kpi = CostKpi('train_cost_kpi',
0.05,
0,
actived=True,
desc='train cost')
train_speed_kpi = AccKpi('train_speed_kpi',
0.05,
0,
actived=True,
unit_repr='images/s',
desc='train speed in one GPU card')
four_card_train_speed_kpi = AccKpi('four_card_train_speed_kpi',
0.05,
0,
actived=True,
unit_repr='images/s',
####this file is only used for continuous evaluation test!
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
#### NOTE kpi.py should shared in models in some way!!!!
train_acc = AccKpi('train_acc', 0.1, 0, actived=True, desc="train acc")
train_loss = CostKpi('train_loss', 0.1, 0, actived=True, desc="train loss")
tracking_kpis = [train_acc, train_loss]
def parse_log(log):
'''
This method should be implemented by model developers.
The suggestion:
each line in the log should be key, value, for example:
"
train_cost\t1.0
test_cost\t1.0
train_cost\t1.0
train_cost\t1.0
train_acc\t1.2
"
"""
continuous_evaluation.py
"""
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import AccKpi
from kpi import DurationKpi
train_acc_kpi = AccKpi('train_acc', 0.2, 0)
pass_duration_kpi = DurationKpi('pass_duration', 0.02, 0, actived=True)
train_acc_kpi_card4 = AccKpi('train_acc_card4', 0.2, 0)
pass_duration_kpi_card4 = DurationKpi('pass_duration_card4',
0.02,
0,
actived=True)
tracking_kpis = [
train_acc_kpi,
pass_duration_kpi,
train_acc_kpi_card4,
pass_duration_kpi_card4,
]
from __future__ import print_function
from __future__ import division
from __future__ import absolute_import
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi
from kpi import AccKpi
train_cost_kpi = CostKpi('train_cost',
0.02,
0,
actived=True,
desc='train cost')
test_cost_kpi = CostKpi('test_cost', 0.02, 0, actived=True, desc='test cost')
test_acc_kpi = AccKpi('test_acc', 0.02, 0, actived=True, desc='test acc')
tracking_kpis = [train_cost_kpi, test_cost_kpi, test_acc_kpi]
def parse_log(log):
for line in log.split('\n'):
fs = line.strip().split('\t')
print(fs)
if len(fs) == 3 and fs[0] == 'kpis':
kpi_name = fs[1]
kpi_value = float(fs[2])
yield kpi_name, kpi_value
def log_to_ce(log):
kpi_tracker = {}
####this file is only used for continuous evaluation test!
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
#### NOTE kpi.py should shared in models in some way!!!!
train_acc1 = AccKpi('train_acc1', 0.01, 0, actived=True, desc="train acc1")
train_acc5 = AccKpi('train_acc5', 0.01, 0, actived=True, desc="train acc5")
train_loss = CostKpi('train_loss', 0.01, 0, actived=True, desc="train loss")
test_acc1 = AccKpi('test_acc1', 0.01, 0, actived=True, desc='test acc1')
test_acc5 = AccKpi('test_acc5', 0.01, 0, actived=True, desc='test acc5')
test_loss = CostKpi('test_loss', 0.01, 0, actived=True, desc='test loss')
#train_speed_kpi = DurationKpi(
# 'train_speed',
# 0.05,
# 0,
# actived=True,
# unit_repr='seconds/image',
# desc='train speed in one GPU card')
tracking_kpis = [
train_acc1, train_acc5, train_loss, test_acc1, test_acc5, test_loss
]
def parse_log(log):
'''
"""
continuous_evaluation.py
"""
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import AccKpi
from kpi import CostKpi
from kpi import DurationKpi
cifar10_128_train_speed_kpi = AccKpi('cifar10_128_train_speed',
0.03,
0,
actived=True)
cifar10_128_gpu_memory_kpi = DurationKpi('cifar10_128_gpu_memory',
0.1,
0,
actived=True)
flowers_32_train_speed_kpi = AccKpi('flowers_32_train_speed',
0.03,
0,
actived=True)
flowers_32_gpu_memory_kpi = DurationKpi('flowers_32_gpu_memory',
0.1,
0,
actived=True)
tracking_kpis = [
cifar10_128_train_speed_kpi,
cifar10_128_gpu_memory_kpi,
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
train_cost_kpi = CostKpi('train_cost', 0.02, actived=True)
test_acc_kpi = AccKpi('test_acc', 0.005, actived=True)
train_duration_kpi = DurationKpi('train_duration', 0.02, actived=True)
train_acc_kpi = AccKpi('train_acc', 0.005, actived=True)
tracking_kpis = [
train_acc_kpi,
train_cost_kpi,
test_acc_kpi,
train_duration_kpi,
]
# this file is only used for continuous evaluation test!
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi
from kpi import DurationKpi
from kpi import AccKpi
each_pass_duration_cpu1_thread1_kpi = DurationKpi(
'each_pass_duration_cpu1_thread1', 0.08, 0, actived=True)
train_acc_cpu1_thread1_kpi = AccKpi('train_acc_cpu1_thread1', 0.08, 0)
each_pass_duration_gpu1_kpi = DurationKpi('each_pass_duration_gpu1',
0.08,
0,
actived=True)
train_acc_gpu1_kpi = AccKpi('train_acc_gpu1', 0.08, 0)
each_pass_duration_gpu4_kpi = DurationKpi('each_pass_duration_gpu4',
0.08,
0,
actived=True)
train_acc_gpu4_kpi = AccKpi('train_acc_gpu4', 0.08, 0)
tracking_kpis = [
each_pass_duration_cpu1_thread1_kpi,
train_acc_cpu1_thread1_kpi,
each_pass_duration_gpu1_kpi,
train_acc_gpu1_kpi,
each_pass_duration_gpu4_kpi,
train_acc_gpu4_kpi,
]
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import AccKpi
flowers_resnet50_dist_train_acc_kpi = AccKpi('flowers_resnet50_dist_train_acc',
0.08,
0,
actived=True)
flowers_resnet50_dist_train_speed_kpi = AccKpi(
'flowers_resnet50_dist_train_speed', 0.1, 0, actived=True)
tracking_kpis = [
flowers_resnet50_dist_train_speed_kpi,
flowers_resnet50_dist_train_acc_kpi,
]
import os
import sys
sys.path.append(os.environ['ceroot'])
from kpi import CostKpi, DurationKpi, AccKpi
train_cost_kpi = CostKpi('train_cost', 0.02, 0, actived=True)
train_speed_kpi = AccKpi('train_speed', 0.02, 0, actived=True)
four_card_speed_kpi = AccKpi('four_card_train_speed', 0.02, 0, actived=True)
tracking_kpis = [train_cost_kpi, train_speed_kpi, four_card_speed_kpi]