Delete run.py
Browse files
run.py
DELETED
@@ -1,101 +0,0 @@
|
|
1 |
-
import os
|
2 |
-
import pandas as pd
|
3 |
-
|
4 |
-
import torch
|
5 |
-
from torch.nn import functional as F
|
6 |
-
from transformers import AutoTokenizer
|
7 |
-
|
8 |
-
from util.utils import *
|
9 |
-
|
10 |
-
from tqdm import tqdm
|
11 |
-
from train import markerModel
|
12 |
-
os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
|
13 |
-
os.environ["CUDA_VISIBLE_DEVICES"] = '0,1'
|
14 |
-
|
15 |
-
device_count = torch.cuda.device_count()
|
16 |
-
device_biomarker = torch.device('cuda' if torch.cuda.is_available() else "cpu")
|
17 |
-
|
18 |
-
device = torch.device('cpu')
|
19 |
-
d_model_name = 'DeepChem/ChemBERTa-10M-MTR'
|
20 |
-
p_model_name = 'DeepChem/ChemBERTa-10M-MLM'
|
21 |
-
|
22 |
-
tokenizer = AutoTokenizer.from_pretrained(d_model_name)
|
23 |
-
prot_tokenizer = AutoTokenizer.from_pretrained(p_model_name)
|
24 |
-
|
25 |
-
#--biomarker Model
|
26 |
-
##-- hyper param config file Load --##
|
27 |
-
config = load_hparams('config/predict.json')
|
28 |
-
config = DictX(config)
|
29 |
-
model = markerModel.load_from_checkpoint(config.load_checkpoint,strict=False)
|
30 |
-
|
31 |
-
# model = BiomarkerModel.load_from_checkpoint('./biomarker_bindingdb_train8595_pretopre/3477h3wf/checkpoints/epoch=30-step=7284.ckpt').to(device_biomarker)
|
32 |
-
|
33 |
-
model.eval()
|
34 |
-
model.freeze()
|
35 |
-
|
36 |
-
if device_biomarker.type == 'cuda':
|
37 |
-
model = torch.nn.DataParallel(model)
|
38 |
-
|
39 |
-
def get_biomarker(drug_inputs, prot_inputs):
|
40 |
-
output_preds = model(drug_inputs, prot_inputs)
|
41 |
-
|
42 |
-
predict = torch.squeeze((output_preds)).tolist()
|
43 |
-
|
44 |
-
# output_preds = torch.relu(output_preds)
|
45 |
-
# predict = torch.tanh(output_preds)
|
46 |
-
# predict = predict.squeeze(dim=1).tolist()
|
47 |
-
|
48 |
-
return predict
|
49 |
-
|
50 |
-
|
51 |
-
def biomarker_prediction(smile_acc, smile_don):
|
52 |
-
try:
|
53 |
-
aas_input = smile_acc
|
54 |
-
|
55 |
-
|
56 |
-
das_input =smile_don
|
57 |
-
d_inputs = tokenizer(aas_input, padding='max_length', max_length=510, truncation=True, return_tensors="pt")
|
58 |
-
# d_inputs = tokenizer(smiles, truncation=True, return_tensors="pt")
|
59 |
-
drug_input_ids = d_inputs['input_ids'].to(device)
|
60 |
-
drug_attention_mask = d_inputs['attention_mask'].to(device)
|
61 |
-
drug_inputs = {'input_ids': drug_input_ids, 'attention_mask': drug_attention_mask}
|
62 |
-
|
63 |
-
p_inputs = prot_tokenizer(das_input, padding='max_length', max_length=510, truncation=True, return_tensors="pt")
|
64 |
-
# p_inputs = prot_tokenizer(aas_input, truncation=True, return_tensors="pt")
|
65 |
-
prot_input_ids = p_inputs['input_ids'].to(device)
|
66 |
-
prot_attention_mask = p_inputs['attention_mask'].to(device)
|
67 |
-
prot_inputs = {'input_ids': prot_input_ids, 'attention_mask': prot_attention_mask}
|
68 |
-
|
69 |
-
output_predict = get_biomarker(drug_inputs, prot_inputs)
|
70 |
-
|
71 |
-
return output_predict
|
72 |
-
|
73 |
-
except Exception as e:
|
74 |
-
print(e)
|
75 |
-
return {'Error_message': e}
|
76 |
-
|
77 |
-
|
78 |
-
def smiles_aas_test(smile_acc,smile_don):
|
79 |
-
|
80 |
-
batch_size = 1
|
81 |
-
try:
|
82 |
-
output_pred = biomarker_prediction((smile_acc), (smile_don))
|
83 |
-
|
84 |
-
datas = output_pred
|
85 |
-
|
86 |
-
## -- Export result data to csv -- ##
|
87 |
-
# df = pd.DataFrame(datas)
|
88 |
-
# df.to_csv('./results/predict_test.csv', index=None)
|
89 |
-
|
90 |
-
# print(df)
|
91 |
-
return datas
|
92 |
-
|
93 |
-
except Exception as e:
|
94 |
-
print(e)
|
95 |
-
return {'Error_message': e}
|
96 |
-
|
97 |
-
|
98 |
-
|
99 |
-
|
100 |
-
|
101 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|