-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy path2_train_rt.py
220 lines (205 loc) · 9.63 KB
/
2_train_rt.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch
import os
from fastNLP.core.metrics import MetricBase,seq_len_to_mask
from fastNLP.core.losses import LossBase
from preprocess import PPeptidePipe
from transformers import BertConfig,RobertaConfig
from model_gly import *
from Bertmodel import _2deepchargeModelms2_bert_irt
import pandas as pd
from pathlib import Path
from utils import *
# ----------------------- training time begin ------------------------------#
from timeit import default_timer as timer
train_time_start = timer()
import datetime
starttime = datetime.datetime.now()
print(f"starttime {starttime}",end="\n\n")
# ----------------------- model parameter for optimization ------------------------------#
import argparse
def parsering():
parser = argparse.ArgumentParser()
parser.add_argument('--lr', type=float,
default=0.0001,
help='learning rate')
parser.add_argument('--warmupsteps', type=int,
default=0,
help='warmupsteps')
parser.add_argument('--weight_decay', type=float,
default=1e-2,
help='weight_decay')
parser.add_argument('--model_ablation', type=str, default="DeepFLR", help='model for ablation (BERT, DeepFLR, protein bert)')
parser.add_argument('--device', type=int, default=4, help='cudadevice')
parser.add_argument("--task_name",type=str, default="mouse_five_tissues")
parser.add_argument('--testdata', type=str, default="alltest", help='data for test')
parser.add_argument('--folder_path', type=str,
default="/remote-home/yxwang/test/zzb/DeepGlyco/DeepSweet_v1/data/mouse/",
help='the folder path for the training data')
parser.add_argument("--trainpathcsv",type=str,
default="Five_tissues/Mouse_five_tissues_data_rt_1st_combine.csv")
parser.add_argument("--irt",type=str, default="no")
parser.add_argument("--irt_csv",type=str, default="/remote-home/yxwang/test/DeepGP_code/data/mouse/All_adjust_irt.csv")
parser.add_argument("--pattern",type=str,
default='*data_rt_1st.csv')
parser.add_argument("--DeepFLR_modelpath",type=str,default="/remote-home/yxwang/test/DeepGP_code/model/DeepFLR/best__2deepchargeModelms2_bert_mediancos_2021-09-20-01-17-50-729399")
args = parser.parse_args()
return args
args=parsering()
lr=args.lr
warmupsteps=args.warmupsteps
weight_decay=args.weight_decay
testdata=args.testdata
pattern=args.pattern
irt_csv=args.irt_csv
DeepFLR_modelpath=args.DeepFLR_modelpath
device = torch.device('cuda', args.device) if torch.cuda.is_available() else torch.device('cpu')
# ----------------------- model parameter------------------------------#
set_seed(seed)
batch_size=128 #128
task_name=args.task_name
check_point_name=str(starttime)+"_rt_"+task_name+"_test_"+testdata+"_checkpoint"
print(f"Project name check_point_name {check_point_name} !",end="\n\n")
print(f"hyper parameter tested lr {lr} !",end="\n\n")
print(f"hyper parameter tested BATCH_SIZE {BATCH_SIZE} !",end="\n\n")
print(f"hyper parameter tested warmupsteps {warmupsteps} !",end="\n\n")
print(f"hyper parameter tested batch_size {batch_size} !",end="\n\n")
print(f"hyper parameter tested weight decay {weight_decay} !",end="\n\n")
# ----------------------- input pre-processing------------------------------#
folder_path = args.folder_path
import folder_walk
trainpathcsv_list=folder_walk.trainpathcsv_list(folder_path=folder_path,pattern=pattern)
print(f"Please check trainpathcsv_list {trainpathcsv_list}. The {len(trainpathcsv_list)} files contains all the training data!")
#将获得的文件合并以后并且导出，按照TotalFDR排序并去重
traincsv=pd.DataFrame()
trainpathcsv=folder_path+args.trainpathcsv
trainpathcsv_path = Path(trainpathcsv)
for x in trainpathcsv_list:
if testdata in x:
print(f"The test data is {x}! It is removed from the training data!")
else:
train=pd.read_csv(x)
traincsv=pd.concat([traincsv,train])
traincsv.sort_values(by='TotalFDR',ascending=True,inplace=True)
traincsv.drop_duplicates(subset=['iden_pep'],inplace=True)
traincsv.reset_index(drop=True,inplace=True)
print(f"with the addition of {x}, the combined file contains {len(traincsv)} lines")
if args.irt=="yes":
print("begin rt adjustment!")
irt=pd.read_csv(irt_csv)
print(traincsv.columns)
traincsv.rename(columns={'RT': 'rt_old'}, inplace=True)
traincsv["run"]=traincsv["GlySpec"].apply(lambda x: x.split("-")[0])
traincsv=pd.merge(traincsv,irt,on=["run","iden_pep"],how="left")
traincsv=traincsv[['GlySpec', 'Charge', 'rt_new' ,'Peptide', 'Mod', 'PlausibleStruct',\
'GlySite', 'iden_pep', 'TotalFDR', 'PrecursorMZ' ]]
traincsv.rename(columns={'rt_new': 'RT'}, inplace=True)
traincsv.to_csv(trainpathcsv,index=False)
else:
traincsv.to_csv(trainpathcsv,index=False)
traindatajson=trainpathcsv[:-4]+"processed_onlyirt.json"
traindatajson_path = Path(traindatajson)
print("Begin matrixwithdict to produce result...")
os.system("python matrixwithdict.py \
--do_irt \
--DDAfile {} \
--outputfile {}".format(trainpathcsv,traindatajson))
#traindata
filename=traindatajson
databundle=PPeptidePipe(vocab=vocab).process_from_file(paths=filename)
totaldata=databundle.get_dataset("train")
print("totaldata",totaldata)
vocab=databundle.get_vocab("peptide_tokens")
traindata,devdata=totaldata.split(0.1)
def savingFastnlpdataset_DataFrame(dataset):
dataset_field=dataset.field_arrays.keys()
frame=pd.DataFrame(columns=dataset_field)
for i in range(len(dataset)):
c_list=[]
for name in dataset_field:
target=dataset.field_arrays[name][i]
if name=="target":
c_list.append(target.cpu().numpy().tolist())
else:
c_list.append(target)
frame.loc[i]=c_list
return frame
# ipdb.set_trace()
# devframe=savingFastnlpdataset_DataFrame(devdata)
# devframe.to_json("20230127_test_model_validata.json")
# torch.save(devframe,"20230127_test_model_validata")
# trainframe=savingFastnlpdataset_DataFrame(traindata)
# trainframe.to_json("20230127_test_model_train_data.json")
# torch.save(trainframe,"20230127_test_model_train_data")
# ipdb.set_trace()
# ----------------------- model ------------------------------#
model_ablation=args.model_ablation #DeepFLR
print(f"hyper parameter tested model_ablation {model_ablation} !",end="\n\n")
if model_ablation=="BERT":
pretrainmodel="bert-base-uncased"
deepms2=_2deepchargeModelms2_bert_irt.from_pretrained(pretrainmodel)
if model_ablation=="DeepFLR":
config=BertConfig.from_pretrained("bert-base-uncased")
bestmodelpath=DeepFLR_modelpath
model_sign=bestmodelpath.split("/")[-1]
deepms2=_2deepchargeModelms2_bert_irt(config)
bestmodel=torch.load(bestmodelpath).state_dict()
origin_model=deepms2.state_dict()
for key in origin_model.keys():
if key in bestmodel.keys():
if bestmodel[key].shape !=origin_model[key].shape:
origin_model[key]=bestmodel[key][:origin_model[key].shape[0],] #linear尺寸匹配不上的部分进行裁剪
print(f"size different key: {key}")
else:
origin_model[key]=bestmodel[key]
else:
print(f"not found key: {key}") #GNN匹配不是的参数就保持原样
deepms2.load_state_dict(origin_model)
#model info
import torchinfo
from torchinfo import summary
summary(deepms2)
# ----------------------- Trainer ------------------------------#
from fastNLP import Const
metrics=CossimilarityMetricfortest_outputrt(savename=None,pred="predirt",target="irt",seq_len='seq_len',
num_col=num_col,sequence='sequence',charge="charge",
decoration="decoration")
from fastNLP import MSELoss
loss=MSELoss(pred="predirt",target="irt")
import torch.optim as optim
optimizer=optim.AdamW(deepms2.parameters(),lr=lr,weight_decay=weight_decay)
from fastNLP import WarmupCallback,SaveModelCallback
save_path=filename[:-5]+"/checkpoints"
callback=[WarmupCallback(warmupsteps)]
callback.append(WandbCallback(project="Deepsweet",name=check_point_name,config={"lr":lr,"seed":seed,
"Batch_size":BATCH_SIZE,"warmupsteps":warmupsteps,"temperature":None,"weight_decay":None}))
callback.append(SaveModelCallback(save_path,top=3))
#trainer
from fastNLP import Trainer
if vocab_save:
vocab.save(os.path.join(save_path,"vocab"))
pptrainer=Trainer(model=deepms2, train_data=traindata,
device=device, dev_data=devdata,
save_path=save_path,
loss=loss,metrics=metrics,callbacks=callback,
optimizer=optimizer,n_epochs=N_epochs,batch_size=batch_size,update_every=int(BATCH_SIZE/batch_size),dev_batch_size=batch_size)
pptrainer.train()
# ----------------------- Time ------------------------------#
train_time_end = timer()
def print_train_time(start: float, end: float, device: torch.device = None):
"""Prints difference between start and end time.
Args:
start (float): Start time of computation (preferred in timeit format).
end (float): End time of computation.
device ([type], optional): Device that compute is running on. Defaults to None.
Returns:
float: time between start and end in seconds (higher is longer).
"""
total_time = end - start
print(f"Train time on {device}: {total_time:.3f} seconds")
return total_time
total_train_time_model_2 = print_train_time(start=train_time_start,
end=train_time_end,
device=device)