IAN/test_script.py at main · Once2gain/IAN · GitHub

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
# import torch
# import re
# import numpy as np
# from fairseq.data import (
#     Dictionary,
# )
# from fairseq.data.encoders.gpt2_bpe import get_encoder
# import torch
#
# tokenizer = get_encoder("gpt2_bpe/encoder.json", "gpt2_bpe/vocab.bpe")
#
# from fairseq import hub_utils
#
# x = hub_utils.from_pretrained(
#     '/data/tangqirui/fairseq/outputs/2023-08-15/14-34-54/checkpoints',
#     'checkpoint_best.pt',
#     '/data/tangqirui/fairseq/DATA-bin/kaggle-mbti/mbti-3-5/fold-1/doc_dataset',
#     archive_map={},
#     bpe="gpt2",
#     load_checkpoint_heads=True,
# )
# model = x["models"][0]
#
# folder = '/data/tangqirui/fairseq/DATA-bin/kaggle-mbti/mbti-3-5/fold-1/doc_dataset'
# data = folder + '/input'
#
# dictionary = Dictionary.load(data + '/dict.txt')
# dictionary.add_symbol("<mask>")
#
# input_f = folder + '/train.input'
# f = open(input_f, 'r', encoding='utf-8')
# input_l = f.readlines()
# src_tokens = []
# for line in input_l[100:104]:
#     for seg in line.split('-100'):
#         seg = seg.strip().split()
#         seg = list(map(int, seg))
#         print(tokenizer.decode(seg))
#
#     item = dictionary.encode_doc(line.strip())
#     prepend = item.new(item.size(0), 1).fill_(dictionary.bos())
#     item = torch.cat([prepend, item], dim=1)
#     src_tokens.append(item)
#
# input_f = folder + '/train.label'
# f = open(input_f, 'r', encoding='utf-8')
# input_l = f.readlines()
# targets = []
# for line in input_l[100:104]:
#     targets.append(line.strip())
#
#
# doc_mairesse_f = folder + '/feats/doc/mairesse/train'
# doc_senticnet_f = folder + '/feats/doc/senticnet/train'
# doc_emotion_f = folder + '/feats/doc/emotion/train'
# doc_senticnet_dis_f = folder + '/feats/doc/senticnet_dis/train'
# seg_mairesse_f = folder + '/feats/seg/mairesse/train'
# seg_senticnet_f = folder + '/feats/seg/senticnet/train'
# seg_emotion_f = folder + '/feats/seg/emotion/train'
#
#
# def read_8(vec_path, keep_raw=False):
#     if keep_raw:
#         samples = np.load(vec_path + '.npy', allow_pickle=True)
#     else:
#         samples = np.load(vec_path + '.npy', allow_pickle=True).astype(np.float32)
#     samples = samples[100:104]
#     if keep_raw:
#         new_samples = []
#         for item in samples:
#             new_item = []
#             for array in item:
#                 new_item.append(torch.tensor(array.astype(np.float32), dtype=torch.float32))
#             new_item = torch.stack(new_item)
#             new_samples.append(new_item)
#         return new_samples
#     return torch.tensor(np.array(samples), dtype=torch.float32)
#
#
# doc_mairesse = read_8(doc_mairesse_f)
# doc_senticnet = read_8(doc_senticnet_f)
# doc_emotion = read_8(doc_emotion_f)
# doc_senticnet_dis = read_8(doc_senticnet_dis_f)
# seg_mairesse = read_8(seg_mairesse_f, keep_raw=True)
# seg_senticnet = read_8(seg_senticnet_f, keep_raw=True)
# seg_emotion = read_8(seg_emotion_f, keep_raw=True)
#
#
# model.eval()
# with torch.no_grad():
#     logits, extra = model(
#         src_tokens,  # batch_size_1(documents_num) * batch_size_2(sentence_num) * seq_length
#         doc_mairesse=doc_mairesse,
#         doc_senticnet=doc_senticnet,
#         doc_emotion=doc_emotion,
#         doc_senticnet_dis=doc_senticnet_dis,
#         seg_mairesse=seg_mairesse,
#         seg_senticnet=seg_senticnet,
#         seg_emotion=seg_emotion,
#     )
# preds = logits.argmax(dim=1)
# print(preds)
#
#
# tmp = []
# for target in targets:
#     tmp.append(target.split()[0])
# targets = torch.tensor(list(map(int, tmp)))
# print(targets)
# print((preds == targets).sum())

# import os
# import pickle
#
# feats_filename = os.path.join("prepare/tools/pandora_baseline/data/feats.pickle")
# vocab_filename = os.path.join("prepare/tools/pandora_baseline/data/vocab.pickle")
# unames_filename = os.path.join("prepare/tools/pandora_baseline/data/unames.pickle")
# gram_feats = pickle.load(open(feats_filename, "rb"))
# gram_feat_names = pickle.load(open(vocab_filename, "rb"))
# unames = pickle.load(open(unames_filename, "rb"))
# print(list(gram_feat_names))

import numpy as np
a = np.zeros([16, 16, 4])
print(a.shape)
print(a[..., 0].shape)

JP_pred = np.array(
    list(chain.from_iterable(log.get("JP_pred", []) for log in logging_outputs))
)
JP_targ = np.array(
    list(chain.from_iterable(log.get("JP_targ", []) for log in logging_outputs))
)
metrics.log_concat_tensor("JP_pred", torch.from_numpy(JP_pred), dim=0)
metrics.log_concat_tensor("JP_targ", torch.from_numpy(JP_targ), dim=0)