public
/
ottertune
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

rename hyperparameters and use them as defaults

This commit is contained in:
yangdsh 2020-01-18 16:40:09 +00:00 committed by Dana Van Aken
parent cb8c597818
commit c2d5cbadd6
10 changed files with 112 additions and 207 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
script/formatting/config/pylintrc
View File

@ -13,7 +13,7 @@ profile=no
# Add files or directories to the blacklist. They should be base names, not # Add files or directories to the blacklist. They should be base names, not
# paths. # paths.
ignore=CVS,.git,manage.py,0001_initial.py,0002_enable_compression.py,0003_load_initial_data.py,0004_add_lhs.py,0005_add_workload_field.py,0006_session_hyper_parameters.py,credentials.py,create_knob_settings.py ignore=CVS,.git,manage.py,0001_initial.py,0002_enable_compression.py,0003_load_initial_data.py,0004_add_lhs.py,0005_add_workload_field.py,0006_session_hyperparameters.py,credentials.py,create_knob_settings.py
# ignore-patterns=**/migrations/*.py # ignore-patterns=**/migrations/*.py

2
server/website/tests/test_views.py
View File

@ -150,7 +150,7 @@ class SessionViewsTests(TestCase):
'storage': '32', 'storage': '32',
'storage_type': 5, 'storage_type': 5,
'dbms': 1, 'dbms': 1,
'hyper_parameters': '{}' 'hyperparameters': '{}'
} }
def setUp(self): def setUp(self):

4
server/website/website/forms.py
View File

@ -134,12 +134,12 @@ class SessionForm(forms.ModelForm):
model = Session model = Session
fields = ('name', 'description', 'tuning_session', 'dbms', 'cpu', 'memory', 'storage', fields = ('name', 'description', 'tuning_session', 'dbms', 'cpu', 'memory', 'storage',
'algorithm', 'target_objective', 'hyper_parameters') 'algorithm', 'target_objective', 'hyperparameters')
widgets = { widgets = {
'name': forms.TextInput(attrs={'required': True}), 'name': forms.TextInput(attrs={'required': True}),
'description': forms.Textarea(attrs={'maxlength': 500, 'rows': 5}), 'description': forms.Textarea(attrs={'maxlength': 500, 'rows': 5}),
'hyper_parameters': forms.Textarea(attrs={'maxlength': 2000, 'rows': 10}), 'hyperparameters': forms.Textarea(attrs={'maxlength': 5000, 'rows': 15}),
} }
labels = { labels = {
'dbms': 'DBMS', 'dbms': 'DBMS',

20
server/website/website/migrations/0006_session_hyper_parameters.py
View File

@ -1,20 +0,0 @@
# -*- coding: utf-8 -*-
# Generated by Django 1.11.23 on 2020-01-12 07:29
from __future__ import unicode_literals
from django.db import migrations, models
class Migration(migrations.Migration):
dependencies = [
('website', '0005_add_workload_field'),
]
operations = [
migrations.AddField(
model_name='session',
name='hyper_parameters',
field=models.TextField(default='{}'),
),
]

20
server/website/website/migrations/0006_session_hyperparameters.py Normal file
View File

@ -0,0 +1,20 @@
# -*- coding: utf-8 -*-
# Generated by Django 1.11.23 on 2020-01-18 16:22
from __future__ import unicode_literals
from django.db import migrations, models
class Migration(migrations.Migration):
dependencies = [
('website', '0005_add_workload_field'),
]
operations = [
migrations.AddField(
model_name='session',
name='hyperparameters',
field=models.TextField(default='{\n "DDPG_ACTOR_HIDDEN_SIZES": [128, 128, 64],\n "DDPG_ACTOR_LEARNING_RATE": 0.02,\n "DDPG_CRITIC_HIDDEN_SIZES": [64, 128, 64],\n "DDPG_CRITIC_LEARNING_RATE": 0.001,\n "DDPG_BATCH_SIZE": 32,\n "DDPG_GAMMA": 0.0,\n "DDPG_SIMPLE_REWARD": true,\n "DDPG_UPDATE_EPOCHS": 30,\n "DDPG_USE_DEFAULT": false,\n "DNN_DEBUG": true,\n "DNN_DEBUG_INTERVAL": 100,\n "DNN_EXPLORE": false,\n "DNN_EXPLORE_ITER": 500,\n "DNN_GD_ITER": 100,\n "DNN_NOISE_SCALE_BEGIN": 0.1,\n "DNN_NOISE_SCALE_END": 0.0,\n "DNN_TRAIN_ITER": 100,\n "FLIP_PROB_DECAY": 0.5,\n "GPR_BATCH_SIZE": 3000,\n "GPR_DEBUG": true,\n "GPR_EPS": 0.001,\n "GPR_EPSILON": 1e-06,\n "GPR_LEARNING_RATE": 0.01,\n "GPR_LENGTH_SCALE": 2.0,\n "GPR_MAGNITUDE": 1.0,\n "GPR_MAX_ITER": 500,\n "GPR_MAX_TRAIN_SIZE": 7000,\n "GPR_MU_MULTIPLIER": 1.0,\n "GPR_MODEL_NAME": "BasicGP",\n "GPR_HP_LEARNING_RATE": 0.001,\n "GPR_HP_MAX_ITER": 5000,\n "GPR_RIDGE": 1.0,\n "GPR_SIGMA_MULTIPLIER": 1.0,\n "GPR_UCB_SCALE": 0.2,\n "GPR_USE_GPFLOW": true,\n "GPR_UCB_BETA": "get_beta_td",\n "IMPORTANT_KNOB_NUMBER": 10000,\n "INIT_FLIP_PROB": 0.3,\n "NUM_SAMPLES": 30,\n "TF_NUM_THREADS": 4,\n "TOP_NUM_CONFIG": 10}'),
),
]

43
server/website/website/models.py
View File

@ -162,7 +162,48 @@ class Session(BaseModel):
target_objective = models.CharField( target_objective = models.CharField(
max_length=64, default=target_objectives.default()) max_length=64, default=target_objectives.default())
hyper_parameters = models.TextField(default="{}") hyperparameters = models.TextField(default='''{
"DDPG_ACTOR_HIDDEN_SIZES": [128, 128, 64],
"DDPG_ACTOR_LEARNING_RATE": 0.02,
"DDPG_CRITIC_HIDDEN_SIZES": [64, 128, 64],
"DDPG_CRITIC_LEARNING_RATE": 0.001,
"DDPG_BATCH_SIZE": 32,
"DDPG_GAMMA": 0.0,
"DDPG_SIMPLE_REWARD": true,
"DDPG_UPDATE_EPOCHS": 30,
"DDPG_USE_DEFAULT": false,
"DNN_DEBUG": true,
"DNN_DEBUG_INTERVAL": 100,
"DNN_EXPLORE": false,
"DNN_EXPLORE_ITER": 500,
"DNN_GD_ITER": 100,
"DNN_NOISE_SCALE_BEGIN": 0.1,
"DNN_NOISE_SCALE_END": 0.0,
"DNN_TRAIN_ITER": 100,
"FLIP_PROB_DECAY": 0.5,
"GPR_BATCH_SIZE": 3000,
"GPR_DEBUG": true,
"GPR_EPS": 0.001,
"GPR_EPSILON": 1e-06,
"GPR_LEARNING_RATE": 0.01,
"GPR_LENGTH_SCALE": 2.0,
"GPR_MAGNITUDE": 1.0,
"GPR_MAX_ITER": 500,
"GPR_MAX_TRAIN_SIZE": 7000,
"GPR_MU_MULTIPLIER": 1.0,
"GPR_MODEL_NAME": "BasicGP",
"GPR_HP_LEARNING_RATE": 0.001,
"GPR_HP_MAX_ITER": 5000,
"GPR_RIDGE": 1.0,
"GPR_SIGMA_MULTIPLIER": 1.0,
"GPR_UCB_SCALE": 0.2,
"GPR_USE_GPFLOW": true,
"GPR_UCB_BETA": "get_beta_td",
"IMPORTANT_KNOB_NUMBER": 10000,
"INIT_FLIP_PROB": 0.3,
"NUM_SAMPLES": 30,
"TF_NUM_THREADS": 4,
"TOP_NUM_CONFIG": 10}''')
def clean(self): def clean(self):
if self.target_objective is None: if self.target_objective is None:

132
server/website/website/settings/constants.py
View File

@ -4,8 +4,6 @@
# Copyright (c) 2017-18, Carnegie Mellon University Database Group # Copyright (c) 2017-18, Carnegie Mellon University Database Group
# #
# ---------------------------------------------
# These parameters are not specified for any session, so they can only be set here # These parameters are not specified for any session, so they can only be set here
# address categorical knobs (enum, boolean) # address categorical knobs (enum, boolean)
@ -14,133 +12,3 @@ ENABLE_DUMMY_ENCODER = False
# ---PIPELINE CONSTANTS--- # ---PIPELINE CONSTANTS---
# how often to run the background tests, in seconds # how often to run the background tests, in seconds
RUN_EVERY = 300 RUN_EVERY = 300
# ---------------------------------------------
# The following parameters can be viewed and modified on the session page on the website
# ---SAMPLING CONSTANTS---
# the number of samples (staring points) in gradient descent
NUM_SAMPLES = 30
# the number of selected tuning knobs
# set it to a large value if you want to disable the knob identification
# phase (i.e. tune all session knobs)
IMPORTANT_KNOB_NUMBER = 10000
# top K config with best performance put into prediction
TOP_NUM_CONFIG = 10
# ---CONSTRAINTS CONSTANTS---
# Initial probability to flip categorical feature in apply_constraints
# server/analysis/constraints.py
INIT_FLIP_PROB = 0.3
# The probability that we flip the i_th categorical feature is
# FLIP_PROB_DECAY * (probability we flip (i-1)_th categorical feature)
FLIP_PROB_DECAY = 0.5
# ---GPR CONSTANTS---
USE_GPFLOW = True
GPR_DEBUG = True
DEFAULT_LENGTH_SCALE = 2.0
DEFAULT_MAGNITUDE = 1.0
# Max training size in GPR model
MAX_TRAIN_SIZE = 7000
# Batch size in GPR model
BATCH_SIZE = 3000
# Threads for TensorFlow config
NUM_THREADS = 4
# Value of beta for UCB
UCB_BETA = 'get_beta_td'
# Name of the GPR model to use (GPFLOW only)
GPR_MODEL_NAME = 'BasicGP'
# ---GRADIENT DESCENT CONSTANTS---
# the maximum iterations of gradient descent
MAX_ITER = 500
DEFAULT_LEARNING_RATE = 0.01
# ---GRADIENT DESCENT FOR GPR---
# a small bias when using training data points as starting points.
GPR_EPS = 0.001
DEFAULT_RIDGE = 1.00
DEFAULT_EPSILON = 1e-6
DEFAULT_SIGMA_MULTIPLIER = 1.0
DEFAULT_MU_MULTIPLIER = 1.0
DEFAULT_UCB_SCALE = 0.2
# ---HYPERPARAMETER TUNING FOR GPR---
HP_MAX_ITER = 5000
HP_LEARNING_RATE = 0.001
# ---GRADIENT DESCENT FOR DNN---
DNN_TRAIN_ITER = 100
# Gradient Descent iteration for recommendation
DNN_GD_ITER = 100
DNN_EXPLORE = False
DNN_EXPLORE_ITER = 500
# noise scale for paramater space exploration
DNN_NOISE_SCALE_BEGIN = 0.1
DNN_NOISE_SCALE_END = 0.0
DNN_DEBUG = True
DNN_DEBUG_INTERVAL = 100
# ---DDPG CONSTRAINTS CONSTANTS---
# Use a simple reward
DDPG_SIMPLE_REWARD = True
# The weight of future rewards in Q value
DDPG_GAMMA = 0.0
# Batch size in DDPG model
DDPG_BATCH_SIZE = 32
# Learning rate of actor network
ACTOR_LEARNING_RATE = 0.02
# Learning rate of critic network
CRITIC_LEARNING_RATE = 0.001
# Number of update epochs per iteration
UPDATE_EPOCHS = 30
# The number of hidden units in each layer of the actor MLP
ACTOR_HIDDEN_SIZES = [128, 128, 64]
# The number of hidden units in each layer of the critic MLP
CRITIC_HIDDEN_SIZES = [64, 128, 64]
# Use the same setting from the CDBTune paper
USE_DEFAULT = False
# Overwrite the DDPG settings if using CDBTune
if USE_DEFAULT:
DDPG_SIMPLE_REWARD = False
DDPG_GAMMA = 0.99
DDPG_BATCH_SIZE = 32
ACTOR_LEARNING_RATE = 0.001
CRITIC_LEARNING_RATE = 0.001
UPDATE_EPOCHS = 1

73
server/website/website/tasks/async_tasks.py
View File

@ -322,7 +322,7 @@ def train_ddpg(result_id):
LOG.info('Add training data to ddpg and train ddpg') LOG.info('Add training data to ddpg and train ddpg')
result = Result.objects.get(pk=result_id) result = Result.objects.get(pk=result_id)
session = Result.objects.get(pk=result_id).session session = Result.objects.get(pk=result_id).session
params = JSONUtil.loads(session.hyper_parameters) params = JSONUtil.loads(session.hyperparameters)
session_results = Result.objects.filter(session=session, session_results = Result.objects.filter(session=session,
creation_time__lt=result.creation_time) creation_time__lt=result.creation_time)
result_info = {} result_info = {}
@ -402,16 +402,18 @@ def train_ddpg(result_id):
LOG.info('reward: %f', reward) LOG.info('reward: %f', reward)
# Update ddpg # Update ddpg
ddpg = DDPG(n_actions=knob_num, n_states=metric_num, alr=params['ACTOR_LEARNING_RATE'], ddpg = DDPG(n_actions=knob_num, n_states=metric_num, alr=params['DDPG_ACTOR_LEARNING_RATE'],
clr=params['CRITIC_LEARNING_RATE'], gamma=params['DDPG_GAMMA'], clr=params['DDPG_CRITIC_LEARNING_RATE'], gamma=params['DDPG_GAMMA'],
batch_size=params['DDPG_BATCH_SIZE'], a_hidden_sizes=params['ACTOR_HIDDEN_SIZES'], batch_size=params['DDPG_BATCH_SIZE'],
c_hidden_sizes=params['CRITIC_HIDDEN_SIZES'], use_default=params['USE_DEFAULT']) a_hidden_sizes=params['DDPG_ACTOR_HIDDEN_SIZES'],
c_hidden_sizes=params['DDPG_CRITIC_HIDDEN_SIZES'],
use_default=params['DDPG_USE_DEFAULT'])
if session.ddpg_actor_model and session.ddpg_critic_model: if session.ddpg_actor_model and session.ddpg_critic_model:
ddpg.set_model(session.ddpg_actor_model, session.ddpg_critic_model) ddpg.set_model(session.ddpg_actor_model, session.ddpg_critic_model)
if session.ddpg_reply_memory: if session.ddpg_reply_memory:
ddpg.replay_memory.set(session.ddpg_reply_memory) ddpg.replay_memory.set(session.ddpg_reply_memory)
ddpg.add_sample(normalized_metric_data, knob_data, reward, normalized_metric_data) ddpg.add_sample(normalized_metric_data, knob_data, reward, normalized_metric_data)
for _ in range(params['UPDATE_EPOCHS']): for _ in range(params['DDPG_UPDATE_EPOCHS']):
ddpg.update() ddpg.update()
session.ddpg_actor_model, session.ddpg_critic_model = ddpg.get_model() session.ddpg_actor_model, session.ddpg_critic_model = ddpg.get_model()
session.ddpg_reply_memory = ddpg.replay_memory.get() session.ddpg_reply_memory = ddpg.replay_memory.get()
@ -443,7 +445,7 @@ def configuration_recommendation_ddpg(result_info): # pylint: disable=invalid-n
result_list = Result.objects.filter(pk=result_id) result_list = Result.objects.filter(pk=result_id)
result = result_list.first() result = result_list.first()
session = result.session session = result.session
params = JSONUtil.loads(session.hyper_parameters) params = JSONUtil.loads(session.hyperparameters)
agg_data = DataUtil.aggregate_data(result_list) agg_data = DataUtil.aggregate_data(result_list)
metric_data, _ = clean_metric_data(agg_data['y_matrix'], agg_data['y_columnlabels'], session) metric_data, _ = clean_metric_data(agg_data['y_matrix'], agg_data['y_columnlabels'], session)
metric_data = metric_data.flatten() metric_data = metric_data.flatten()
@ -455,8 +457,9 @@ def configuration_recommendation_ddpg(result_info): # pylint: disable=invalid-n
metric_num = len(metric_data) metric_num = len(metric_data)
ddpg = DDPG(n_actions=knob_num, n_states=metric_num, ddpg = DDPG(n_actions=knob_num, n_states=metric_num,
a_hidden_sizes=params['ACTOR_HIDDEN_SIZES'], a_hidden_sizes=params['DDPG_ACTOR_HIDDEN_SIZES'],
c_hidden_sizes=params['CRITIC_HIDDEN_SIZES'], use_default=params['USE_DEFAULT']) c_hidden_sizes=params['DDPG_CRITIC_HIDDEN_SIZES'],
use_default=params['DDPG_USE_DEFAULT'])
if session.ddpg_actor_model is not None and session.ddpg_critic_model is not None: if session.ddpg_actor_model is not None and session.ddpg_critic_model is not None:
ddpg.set_model(session.ddpg_actor_model, session.ddpg_critic_model) ddpg.set_model(session.ddpg_actor_model, session.ddpg_critic_model)
if session.ddpg_reply_memory is not None: if session.ddpg_reply_memory is not None:
@ -492,7 +495,7 @@ def combine_workload(target_data):
newest_result = Result.objects.get(pk=target_data['newest_result_id']) newest_result = Result.objects.get(pk=target_data['newest_result_id'])
session = newest_result.session session = newest_result.session
params = JSONUtil.loads(session.hyper_parameters) params = JSONUtil.loads(session.hyperparameters)
cleaned_workload_knob_data = clean_knob_data(workload_knob_data["data"], cleaned_workload_knob_data = clean_knob_data(workload_knob_data["data"],
workload_knob_data["columnlabels"], workload_knob_data["columnlabels"],
newest_result.session) newest_result.session)
@ -661,7 +664,7 @@ def configuration_recommendation(recommendation_input):
LOG.info('configuration_recommendation called') LOG.info('configuration_recommendation called')
newest_result = Result.objects.get(pk=target_data['newest_result_id']) newest_result = Result.objects.get(pk=target_data['newest_result_id'])
session = newest_result.session session = newest_result.session
params = JSONUtil.loads(session.hyper_parameters) params = JSONUtil.loads(session.hyperparameters)
if target_data['bad'] is True: if target_data['bad'] is True:
target_data_res = create_and_save_recommendation( target_data_res = create_and_save_recommendation(
@ -726,17 +729,17 @@ def configuration_recommendation(recommendation_input):
elif algorithm == AlgorithmType.GPR: elif algorithm == AlgorithmType.GPR:
# default gpr model # default gpr model
if params['USE_GPFLOW']: if params['GPR_USE_GPFLOW']:
model_kwargs = {} model_kwargs = {}
model_kwargs['model_learning_rate'] = params['HP_LEARNING_RATE'] model_kwargs['model_learning_rate'] = params['GPR_HP_LEARNING_RATE']
model_kwargs['model_maxiter'] = params['HP_MAX_ITER'] model_kwargs['model_maxiter'] = params['GPR_HP_MAX_ITER']
opt_kwargs = {} opt_kwargs = {}
opt_kwargs['learning_rate'] = params['DEFAULT_LEARNING_RATE'] opt_kwargs['learning_rate'] = params['GPR_LEARNING_RATE']
opt_kwargs['maxiter'] = params['MAX_ITER'] opt_kwargs['maxiter'] = params['GPR_MAX_ITER']
opt_kwargs['bounds'] = [X_min, X_max] opt_kwargs['bounds'] = [X_min, X_max]
opt_kwargs['debug'] = params['GPR_DEBUG'] opt_kwargs['debug'] = params['GPR_DEBUG']
opt_kwargs['ucb_beta'] = ucb.get_ucb_beta(params['UCB_BETA'], opt_kwargs['ucb_beta'] = ucb.get_ucb_beta(params['GPR_UCB_BETA'],
scale=params['DEFAULT_UCB_SCALE'], scale=params['GPR_UCB_SCALE'],
t=i + 1., ndim=X_scaled.shape[1]) t=i + 1., ndim=X_scaled.shape[1])
tf.reset_default_graph() tf.reset_default_graph()
graph = tf.get_default_graph() graph = tf.get_default_graph()
@ -745,17 +748,17 @@ def configuration_recommendation(recommendation_input):
**model_kwargs) **model_kwargs)
res = tf_optimize(m.model, X_samples, **opt_kwargs) res = tf_optimize(m.model, X_samples, **opt_kwargs)
else: else:
model = GPRGD(length_scale=params['DEFAULT_LENGTH_SCALE'], model = GPRGD(length_scale=params['GPR_LENGTH_SCALE'],
magnitude=params['DEFAULT_MAGNITUDE'], magnitude=params['GPR_MAGNITUDE'],
max_train_size=params['MAX_TRAIN_SIZE'], max_train_size=params['GPR_MAX_TRAIN_SIZE'],
batch_size=params['BATCH_SIZE'], batch_size=params['GPR_BATCH_SIZE'],
num_threads=params['NUM_THREADS'], num_threads=params['TF_NUM_THREADS'],
learning_rate=params['DEFAULT_LEARNING_RATE'], learning_rate=params['GPR_LEARNING_RATE'],
epsilon=params['DEFAULT_EPSILON'], epsilon=params['GPR_EPSILON'],
max_iter=params['MAX_ITER'], max_iter=params['GPR_MAX_ITER'],
sigma_multiplier=params['DEFAULT_SIGMA_MULTIPLIER'], sigma_multiplier=params['GPR_SIGMA_MULTIPLIER'],
mu_multiplier=params['DEFAULT_MU_MULTIPLIER'], mu_multiplier=params['GPR_MU_MULTIPLIER'],
ridge=params['DEFAULT_RIDGE']) ridge=params['GPR_RIDGE'])
model.fit(X_scaled, y_scaled, X_min, X_max) model.fit(X_scaled, y_scaled, X_min, X_max)
res = model.predict(X_samples, constraint_helper=constraint_helper) res = model.predict(X_samples, constraint_helper=constraint_helper)
@ -814,7 +817,7 @@ def map_workload(map_workload_input):
newest_result = Result.objects.get(pk=target_data['newest_result_id']) newest_result = Result.objects.get(pk=target_data['newest_result_id'])
session = newest_result.session session = newest_result.session
params = JSONUtil.loads(session.hyper_parameters) params = JSONUtil.loads(session.hyperparameters)
target_workload = newest_result.workload target_workload = newest_result.workload
X_columnlabels = np.array(target_data['X_columnlabels']) X_columnlabels = np.array(target_data['X_columnlabels'])
y_columnlabels = np.array(target_data['y_columnlabels']) y_columnlabels = np.array(target_data['y_columnlabels'])
@ -929,11 +932,11 @@ def map_workload(map_workload_input):
# and then predict the performance of each metric for each of # and then predict the performance of each metric for each of
# the knob configurations attempted so far by the target. # the knob configurations attempted so far by the target.
y_col = y_col.reshape(-1, 1) y_col = y_col.reshape(-1, 1)
model = GPRNP(length_scale=params['DEFAULT_LENGTH_SCALE'], model = GPRNP(length_scale=params['GPR_LENGTH_SCALE'],
magnitude=params['DEFAULT_MAGNITUDE'], magnitude=params['GPR_MAGNITUDE'],
max_train_size=params['MAX_TRAIN_SIZE'], max_train_size=params['GPR_MAX_TRAIN_SIZE'],
batch_size=params['BATCH_SIZE']) batch_size=params['GPR_BATCH_SIZE'])
model.fit(X_scaled, y_col, ridge=params['DEFAULT_RIDGE']) model.fit(X_scaled, y_col, ridge=params['GPR_RIDGE'])
predictions[:, j] = model.predict(X_target).ypreds.ravel() predictions[:, j] = model.predict(X_target).ypreds.ravel()
# Bin each of the predicted metric columns by deciles and then # Bin each of the predicted metric columns by deciles and then
# compute the score (i.e., distance) between the target workload # compute the score (i.e., distance) between the target workload

6
server/website/website/templates/edit_session.html
View File

@ -49,9 +49,9 @@
<td>{{ form.target_objective.label_tag }}</td> <td>{{ form.target_objective.label_tag }}</td>
<td>{{ form.target_objective }}</td> <td>{{ form.target_objective }}</td>
</tr> </tr>
<tr id="target_obj_row"> <tr id="hyperparameters_row">
<td>{{ form.hyper_parameters.label_tag }}</td> <td>{{ form.hyperparameters.label_tag }}</td>
<td>{{ form.hyper_parameters }}</td> <td>{{ form.hyperparameters }}</td>
</tr> </tr>
<tr id="upload_code_row"> <tr id="upload_code_row">
<td>{{ form.gen_upload_code.label_tag }}</td> <td>{{ form.gen_upload_code.label_tag }}</td>

17
server/website/website/views.py
View File

@ -330,14 +330,12 @@ def create_or_edit_session(request, project_id, session_id=''):
else: else:
# Return a new form with defaults for creating a new session # Return a new form with defaults for creating a new session
session = None session = None
hyper_parameters = JSONUtil.dumps(utils.get_constants())
form_kwargs.update( form_kwargs.update(
initial={ initial={
'dbms': DBMSCatalog.objects.get( 'dbms': DBMSCatalog.objects.get(
type=DBMSType.POSTGRES, version='9.6'), type=DBMSType.POSTGRES, version='9.6'),
'algorithm': AlgorithmType.GPR, 'algorithm': AlgorithmType.GPR,
'target_objective': target_objectives.default(), 'target_objective': target_objectives.default()
'hyper_parameters': hyper_parameters
}) })
form = SessionForm(**form_kwargs) form = SessionForm(**form_kwargs)
context = { context = {
@ -1457,36 +1455,31 @@ def create_test_website(request): # pylint: disable=unused-argument
password='ottertune_test_user') password='ottertune_test_user')
test_project = Project.objects.create(user=test_user, name='ottertune_test_project', test_project = Project.objects.create(user=test_user, name='ottertune_test_project',
creation_time=now(), last_update=now()) creation_time=now(), last_update=now())
hyper_parameters = JSONUtil.dumps(utils.get_constants())
# create no tuning session # create no tuning session
s1 = Session.objects.create(name='test_session_no_tuning', tuning_session='no_tuning_session', s1 = Session.objects.create(name='test_session_no_tuning', tuning_session='no_tuning_session',
dbms_id=1, hardware=test_hardware, project=test_project, dbms_id=1, hardware=test_hardware, project=test_project,
creation_time=now(), last_update=now(), user=test_user, creation_time=now(), last_update=now(), user=test_user,
upload_code='ottertuneTestNoTuning', upload_code='ottertuneTestNoTuning')
hyper_parameters=hyper_parameters)
set_default_knobs(s1) set_default_knobs(s1)
# create gpr session # create gpr session
s2 = Session.objects.create(name='test_session_gpr', tuning_session='tuning_session', s2 = Session.objects.create(name='test_session_gpr', tuning_session='tuning_session',
dbms_id=1, hardware=test_hardware, project=test_project, dbms_id=1, hardware=test_hardware, project=test_project,
creation_time=now(), last_update=now(), algorithm=AlgorithmType.GPR, creation_time=now(), last_update=now(), algorithm=AlgorithmType.GPR,
upload_code='ottertuneTestTuningGPR', user=test_user, upload_code='ottertuneTestTuningGPR', user=test_user)
hyper_parameters=hyper_parameters)
set_default_knobs(s2) set_default_knobs(s2)
# create dnn session # create dnn session
s3 = Session.objects.create(name='test_session_dnn', tuning_session='tuning_session', s3 = Session.objects.create(name='test_session_dnn', tuning_session='tuning_session',
dbms_id=1, hardware=test_hardware, project=test_project, dbms_id=1, hardware=test_hardware, project=test_project,
creation_time=now(), last_update=now(), algorithm=AlgorithmType.DNN, creation_time=now(), last_update=now(), algorithm=AlgorithmType.DNN,
upload_code='ottertuneTestTuningDNN', user=test_user, upload_code='ottertuneTestTuningDNN', user=test_user)
hyper_parameters=hyper_parameters)
set_default_knobs(s3) set_default_knobs(s3)
# create ddpg session # create ddpg session
s4 = Session.objects.create(name='test_session_ddpg', tuning_session='tuning_session', s4 = Session.objects.create(name='test_session_ddpg', tuning_session='tuning_session',
dbms_id=1, hardware=test_hardware, project=test_project, dbms_id=1, hardware=test_hardware, project=test_project,
creation_time=now(), last_update=now(), user=test_user, creation_time=now(), last_update=now(), user=test_user,
upload_code='ottertuneTestTuningDDPG', upload_code='ottertuneTestTuningDDPG',
algorithm=AlgorithmType.DDPG, algorithm=AlgorithmType.DDPG)
hyper_parameters=hyper_parameters)
set_default_knobs(s4) set_default_knobs(s4)
response = HttpResponse("Success: create test website successfully") response = HttpResponse("Success: create test website successfully")
return response return response