liuyu/qlib
1
0
Fork 0
mirror of https://github.com/microsoft/qlib.git synced 2026-07-14 00:06:58 +08:00
Code Activity

update model

Browse Source
This commit is contained in:
bxdd
2020-10-04 17:16:39 +08:00
committed by you-n-g
parent 992983ff4c
commit 916d9d363e
3 changed files with 11 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
examples/estimator/estimator_config.yaml
View File

@@ -51,4 +51,3 @@ qlib_data:
# when testing, please modify the following parameters according to the specific environment # when testing, please modify the following parameters according to the specific environment
provider_uri: "~/.qlib/qlib_data/cn_data" provider_uri: "~/.qlib/qlib_data/cn_data"
region: "cn" region: "cn"
redis_port: 1222

3
examples/estimator/estimator_config_dnn.yaml
View File

@@ -48,11 +48,8 @@ backtest:
open_cost: 0.0005 open_cost: 0.0005
close_cost: 0.0015 close_cost: 0.0015
min_cost: 5 min_cost: 5
long_short_backtest_args:
topk: 50
qlib_data: qlib_data:
# when testing, please modify the following parameters according to the specific environment # when testing, please modify the following parameters according to the specific environment
provider_uri: "~/.qlib/qlib_data/cn_data" provider_uri: "~/.qlib/qlib_data/cn_data"
region: "cn" region: "cn"
redis_port: 1222

19
qlib/contrib/model/pytorch_nn.py
View File

@@ -47,7 +47,7 @@ class DNNModelPytorch(Model):
self, self,
input_dim, input_dim,
output_dim, output_dim,
layers=(256, 256, 128), layers=(256, 512, 768, 1024, 768, 512, 256, 128, 64),
lr=0.001, lr=0.001,
max_steps=300, max_steps=300,
batch_size=2000, batch_size=2000,
@@ -76,6 +76,7 @@ class DNNModelPytorch(Model):
self.optimizer = optimizer.lower() self.optimizer = optimizer.lower()
self.loss_type = loss self.loss_type = loss
self.visible_GPU = GPU self.visible_GPU = GPU
self.use_gpu = torch.cuda.is_available()
self.logger.info( self.logger.info(
"DNN parameters setting:" "DNN parameters setting:"
@@ -90,7 +91,8 @@ class DNNModelPytorch(Model):
"\noptimizer : {}" "\noptimizer : {}"
"\nloss_type : {}" "\nloss_type : {}"
"\neval_steps : {}" "\neval_steps : {}"
"\nvisible_GPU : {}".format( "\nvisible_GPU : {}"
"\nuse_GPU : {}".format(
layers, layers,
lr, lr,
max_steps, max_steps,
@@ -103,6 +105,7 @@ class DNNModelPytorch(Model):
loss, loss,
eval_steps, eval_steps,
GPU, GPU,
self.use_gpu,
) )
) )
@@ -133,7 +136,7 @@ class DNNModelPytorch(Model):
) )
self._fitted = False self._fitted = False
self.use_gpu = torch.cuda.is_available()
if self.use_gpu: if self.use_gpu:
self.dnn_model.cuda() self.dnn_model.cuda()
@@ -327,20 +330,20 @@ class AverageMeter(object):
class Net(nn.Module): class Net(nn.Module):
def __init__(self, input_dim, output_dim, layers=(256, 256, 256), loss="mse"): def __init__(self, input_dim, output_dim, layers=(256, 512, 768, 512, 256, 128, 64), loss="mse"):
super(Net, self).__init__() super(Net, self).__init__()
layers = [input_dim] + list(layers) layers = [input_dim] + list(layers)
dnn_layers = [] dnn_layers = []
drop_input = nn.Dropout(0.1) drop_input = nn.Dropout(0.05)
dnn_layers.append(drop_input) dnn_layers.append(drop_input)
for i, (input_dim, hidden_units) in enumerate(zip(layers[:-1], layers[1:])): for i, (input_dim, hidden_units) in enumerate(zip(layers[:-1], layers[1:])):
fc = nn.Linear(input_dim, hidden_units) fc = nn.Linear(input_dim, hidden_units)
activation = nn.ReLU() activation = nn.ReLU()
bn = nn.BatchNorm1d(hidden_units) bn = nn.BatchNorm1d(hidden_units)
drop = nn.Dropout(0.1) seq = nn.Sequential(fc, bn, activation)
seq = nn.Sequential(fc, bn, activation, drop)
dnn_layers.append(seq) dnn_layers.append(seq)
drop_input = nn.Dropout(0.05)
dnn_layers.append(drop_input)
if loss == "mse": if loss == "mse":
fc = nn.Linear(hidden_units, output_dim) fc = nn.Linear(hidden_units, output_dim)
dnn_layers.append(fc) dnn_layers.append(fc)