-
Notifications
You must be signed in to change notification settings - Fork 52
/
main.py
397 lines (381 loc) · 16.1 KB
/
main.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
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
import json
import logging
from collections import Counter
import openai
from config import args
from extracter import get_precision, extract_answer
from prompt import get_prompt, construct_input
from prediction_runner import basic_runner
from utils import write_json, print_now, load_data, print_exp, mkpath
now = print_now(1).split(' ')[0].replace('/', '-')
Result_Folder = 'result/{}'.format(now)
mkpath('result')
mkpath(Result_Folder)
mkpath(f'{Result_Folder}/{args.dataset}')
Log_Folder = 'log/{}'.format(now)
mkpath('log')
mkpath(Log_Folder)
mkpath(f'{Log_Folder}/{args.dataset}')
Decoder_Error_File = f'{Result_Folder}/{args.learning_type}-{args.dataset}-{args.prompt_id}-{args.engine}-SC-{args.SC}_deco.json'
Predict_File = f'{Result_Folder}/{args.dataset}/{args.learning_type}-{args.prompt_id}-{args.engine}-SC-{args.SC}.json'
Log_File = f'{Log_Folder}/{args.dataset}/{args.learning_type}-{args.prompt_id}-{args.engine}-SC-{args.SC}.log'
# logging.basicConfig(filename=Log_File)
formatter = logging.Formatter('%(asctime)s - %(pathname)s[line:%(lineno)d] - %(levelname)s: %(message)s')
sh = logging.StreamHandler()
fh = logging.FileHandler(filename=Log_File)
fh.setFormatter(formatter)
sh.setFormatter(formatter)
logger = logging.getLogger()
logger.addHandler(fh)
logger.addHandler(sh)
def zero_shot_cot():
correct = 0
apikey_list = json.load(open('apikeys.json', 'r', encoding='utf-8'))
assert len(apikey_list) >= 1, f"need 1 apikey, find {len(apikey_list)} in '/apikeys.json'"
apikey = apikey_list[0]
question, answer, ids = load_data(args)
_, prompt = get_prompt()
for idx, element in enumerate(question):
# --test_star 100 从100开始测试
# --test_end 100 只测前100个
# --test_star 100 --test_end 200 从100开始测试，到200结束
inputs = construct_input(prompt, element)
try:
get_result, pred, error_msg = basic_runner(args, inputs, args.max_length_cot, apikey)
except Exception as e:
decode_error_data = {
'question': question[idx]
}
write_json(decode_error_data, Decoder_Error_File)
logger.warning(
f"an error raised when predicting (question id: {ids[idx]}). "
f"ERROR: {getattr(e.__class__, '__name__')}:{str(e)}"
)
continue
if not get_result:
logger.warning(
f"not get predicted result (question id: {ids[idx]})."
f"ERROR Message: {error_msg if error_msg else None}"
)
continue
if args.SC:
answer_list = []
for i in range(len(pred)):
input_ = inputs + pred[i]
if 'Therefore, the answer is' in input_ or 'The answer is' in input_:
if 'The answer is' in input_:
input_2 = input_.split('The answer is')[-1]
else:
input_2 = input_.split('the answer is')[-1]
try:
pred_answer1 = extract_answer(args, input_2)
except:
pred_answer1 = None
else:
inputs2 = input_ + ' ' + args.direct_answer_trigger_for_direct
try:
get_result, pred3, error_msg = basic_runner(args, inputs2, 32, apikey)
except Exception as e:
decode_error_data = {
'question': question[idx]
}
write_json(decode_error_data, Decoder_Error_File)
logger.warning(
f"an error raised when predicting (question id: {ids[idx]}). "
f"ERROR: {getattr(e.__class__, '__name__')}:{str(e)}"
)
continue
if not get_result:
logger.warning(
f"not get predicted result (question id: {ids[idx]})."
f"ERROR Message: {error_msg if error_msg else None}"
)
continue
try:
pred_answer1 = extract_answer(args, pred3)
except:
pred_answer1 = None
answer_list.append(pred_answer1)
collection_words = Counter(answer_list)
pred_answer = collection_words.most_common(1)[0][0]
else:
if 'Therefore, the answer is' in pred or 'The answer is' in pred:
if 'The answer is' in pred:
pred2 = pred.split('The answer is')[-1]
else:
pred2 = pred.split('the answer is')[-1]
try:
pred_answer = extract_answer(args, pred2)
except:
pred_answer = None
else:
inputs2 = inputs + pred + ' ' + args.direct_answer_trigger_for_direct
try:
get_result, pred3, error_msg = basic_runner(args, inputs2, 32, apikey)
except Exception as e:
decode_error_data = {
'question': question[idx]
}
write_json(decode_error_data, Decoder_Error_File)
logger.warning(
f"an error raised when predicting (question id: {ids[idx]}). "
f"ERROR: {getattr(e.__class__, '__name__')}:{str(e)}"
)
continue
if not get_result:
logger.warning(
f"not get predicted result (question id: {ids[idx]})."
f"ERROR Message: {error_msg if error_msg else None}"
)
continue
try:
pred_answer = extract_answer(args, pred3)
except:
pred_answer = None
ans = False
if pred_answer is not None:
if args.dataset.lower() in ["svamp", "gsm8k", "multiarith", "addsub", "singleeq"]:
if abs(pred_answer - answer[idx]) < 1e-3:
correct += 1
ans = True
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
if isinstance(pred_answer, float) and isinstance(answer[idx], float):
precision = min(get_precision(pred_answer), get_precision(answer[idx]))
if round(pred_answer, precision) == round(answer[idx], precision):
correct += 1
ans = True
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
ans = False
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
if pred_answer == answer[idx]:
correct += 1
ans = True
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
logger.info(
f"correct:{correct} tested:{idx + 1} {correct / (idx + 1)} total:{len(question)}"
)
return correct
def few_shot_cot():
correct = 0
apikey_list = json.load(open('apikeys.json', 'r', encoding='utf-8'))
assert len(apikey_list) >= 1
apikey = apikey_list[0]
question, answer, ids = load_data(args)
demos, prompt = get_prompt()
for idx, element in enumerate(question):
# --test_star 100 从100开始测试
# --test_end 100 只测前100个
# --test_star 100 --test_end 200 从100开始测试，到200结束
inputs = construct_input(prompt, element)
inputs = demos + '\n' + inputs
try:
get_result, pred, error_msg = basic_runner(args, inputs, args.max_length_cot, apikey)
except Exception as e:
decode_error_data = {
'question': question[idx]
}
write_json(decode_error_data, Decoder_Error_File)
logger.warning(
f"an error raised when predicting (question id: {ids[idx]}). "
f"ERROR: {getattr(e.__class__, '__name__')}:{str(e)}"
)
continue
if not get_result:
logger.warning(
f"not get predicted result (question id: {ids[idx]})."
f"ERROR Message: {error_msg if error_msg else None}"
)
continue
if 'The answer is' in pred or 'the answer is' in pred:
if 'The answer is' in pred:
pred2 = pred.split('The answer is')[-1]
else:
pred2 = pred.split('the answer is')[-1]
try:
pred_answer = extract_answer(args, pred2)
except:
pred_answer = None
else:
inputs2 = inputs + pred + ' ' + args.direct_answer_trigger_for_direct
try:
get_result, pred3, error_msg = basic_runner(args, inputs2, 32, apikey)
except Exception as e:
decode_error_data = {
'question': question[idx]
}
write_json(decode_error_data, Decoder_Error_File)
logger.warning(
f"an error raised when predicting (question id: {ids[idx]}). "
f"ERROR: {getattr(e.__class__, '__name__')}:{str(e)}"
)
continue
if not get_result:
logger.warning(
f"not get predicted result (question id: {ids[idx]})."
f"ERROR Message: {error_msg if error_msg else None}"
)
continue
try:
pred_answer = extract_answer(args, pred3)
except:
pred_answer = None
ans = False
if pred_answer is not None:
if args.dataset.lower() in ["svamp", "gsm8k", "multiarith", "addsub", "singleeq"]:
if abs(pred_answer - answer[idx]) < 1e-3:
correct += 1
ans = True
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
if isinstance(pred_answer, float) and isinstance(answer[idx], float):
precision = min(get_precision(pred_answer), get_precision(answer[idx]))
if round(pred_answer, precision) == round(answer[idx], precision):
correct += 1
ans = True
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
ans = False
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
if pred_answer == answer[idx]:
correct += 1
ans = True
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
else:
json_data = {
"ID": ids[idx],
"question": question[idx],
"chain-of-thought": pred,
"pred": pred_answer,
"answer": answer[idx],
"ans": ans
}
write_json(json_data, Predict_File)
logger.info(
f"correct:{correct} tested:{idx + 1} {correct / (idx + 1)} total:{len(question)}"
)
return correct
if __name__ == '__main__':
print_exp(args)
if args.learning_type == 'zero_shot':
alls = zero_shot_cot()
else:
alls = few_shot_cot()
logger.info(f"correct num: {alls}")