-
Notifications
You must be signed in to change notification settings - Fork 16
/
eval_checkpoint.py
173 lines (139 loc) · 7.64 KB
/
eval_checkpoint.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
# Copyright (c) Microsoft Corporation.
# Licensed under the MIT license.
import argparse
import os
import logging
from dpr.options import add_encoder_params, add_cuda_params, setup_args_gpu, set_seed, print_args
from generate_dense_embeddings import main as dense_encoding
from dense_retriever import main as dense_retrieve
from train_dense_encoder_modified import BiEncoderTrainer
from dense_retriever import *
from dpr.utils.model_utils import load_states_from_checkpoint_only_model
logger = logging.getLogger()
logger.setLevel(logging.INFO)
if logger.hasHandlers():
logger.handlers.clear()
console = logging.StreamHandler()
logger.addHandler(console)
def recall_k(result_ctx_ids: List[Tuple[List[object], List[float]]], topK : int) -> float:
top_k_hits = 0
total_ref = len(result_ctx_ids)
for idx, res in enumerate(result_ctx_ids):
assert(topK <= len(res[0]))
candidates = set(res[0][:topK])
if str(idx+1) in candidates:
top_k_hits += 1
recall = float(top_k_hits/total_ref)
logger.info(f'Validation results: recall@{topK}:{recall:.3f}')
return recall
parser = argparse.ArgumentParser()
add_cuda_params(parser)
add_encoder_params(parser)
parser.add_argument('--eval_mode', type=str, default='qa', help="evaluation mode for model")
parser.add_argument('--ctx_file', type=str, default=None, help='Path to passages set .tsv file for encoding')
parser.add_argument('--ict_file', type=str, default=None, help='Path to passages set .tsv file for ict')
parser.add_argument('--qa_file', required=True, type=str, default=None,
help="Question and answers file of the format: question \\t ['answer1','answer2', ...]")
parser.add_argument("--hard_negatives", default=1, type=int,
help="amount of hard negative ctx per question")
parser.add_argument("--other_negatives", default=1, type=int,
help="amount of other negative ctx per question")
parser.add_argument('--out_folder', required=False, type=str, default=None,
help='output .tsv file path to write results to ')
parser.add_argument('--seed', type=int, default=42, help="seed for random number generator")
parser.add_argument('--shard_id', type=int, default=0, help="Number(0-based) of data shard to process")
parser.add_argument('--num_shards', type=int, default=1, help="Total amount of data shards")
parser.add_argument('--batch_size', type=int, default=32, help="Batch size for the passage encoder forward pass")
parser.add_argument('--match', type=str, default='string', choices=['regex', 'string'],
help="Answer matching logic type")
parser.add_argument('--n_docs', type=int, default=5, help="Amount of top docs to return")
parser.add_argument('--validation_workers', type=int, default=16,
help="Number of parallel processes to validate results")
parser.add_argument('--index_buffer', type=int, default=50000,
help="Temporal memory data buffer size (in samples) for indexer")
parser.add_argument("--hnsw_index", action='store_true', help='If enabled, use inference time efficient HNSW index')
parser.add_argument("--encoding", action='store_true', help='If enabled, encode the doc features offline')
parser.add_argument("--shard_folder", action='store_true', help='If enabled, encode the doc features are sharded')
parser.add_argument("--debug", action='store_true', help='If enabled, debug mode')
parser.add_argument("--load_trained_model", action='store_true', help='If enabled, debug mode')
parser.add_argument("--log_batch_step", default=100, type=int, help="")
args = parser.parse_args()
assert args.model_file, 'Please specify --model_file checkpoint to init model weights'
setup_args_gpu(args)
set_seed(args)
print_args(args)
args.do_lower_case = True
if args.eval_mode == 'qa':
args.out_file = os.path.join(args.out_folder, 'dense_embedding', os.path.basename(args.ctx_file))
if args.encoding:
dense_encoding(args)
args.encoded_ctx_file = args.out_file + '_*'
if 'ance' in args.encoder_model_type:
args.sequence_length = 64
args.out_file = os.path.join(args.out_folder, '.'.join(['qa_eval', os.path.basename(args.qa_file)]))
hit_res = dense_retrieve(args)
elif args.eval_mode == 'ict':
args.learning_rate, args.adam_eps, args.weight_decay, args.gradient_accumulation_steps = 0.0, 0.0, 0.0, 1
args.adam_betas = '(0.0, 0.0)'
args.ict = True
args.batch_size = 128
args.dev_file = '../data/DPR/data/wikipedia_split_small/wiki_test.tsv'
args.dev_batch_size = args.batch_size
trainer = BiEncoderTrainer(args)
nll_res = trainer.validate_nll()
elif args.eval_mode == 'rank':
if not args.load_trained_model:
saved_state = load_states_from_checkpoint(args.model_file)
else:
saved_state = load_states_from_checkpoint_only_model(args.model_file)
set_encoder_params_from_state(saved_state.encoder_params, args)
tensorizer, encoder, _ = init_biencoder_components(args.encoder_model_type, args, inference_only=True)
encoder = encoder.question_model
encoder, _ = setup_for_distributed_mode(encoder, None, args.device, args.n_gpu,
args.local_rank,
args.fp16)
encoder.eval()
model_to_load = get_model_obj(encoder)
logger.info('Loading saved model state ...')
prefix_len = len('question_model.')
question_encoder_state = {key[prefix_len:]: value for (key, value) in saved_state.model_dict.items() if
key.startswith('question_model.')}
model_to_load.load_state_dict(question_encoder_state)
vector_size = model_to_load.get_out_size()
logger.info('Encoder vector_size=%d', vector_size)
index_buffer_sz = args.index_buffer
if args.hnsw_index:
index = DenseHNSWFlatIndexer(vector_size)
index_buffer_sz = -1
else:
index = DenseFlatIndexer(vector_size)
retriever = DenseRetriever(encoder, args.batch_size, tensorizer, index)
args.out_file = os.path.join(os.path.dirname(args.model_file), 'dense_embedding', os.path.basename(args.model_file) + '.' + os.path.basename(args.ctx_file))
if args.shard_folder:
args.out_file = os.path.join(os.path.dirname(args.model_file), 'dense_embedding/shard', os.path.basename(args.model_file) + '.' + os.path.basename(args.ctx_file))
if args.encoding:
dense_encoding(args)
args.encoded_ctx_file = args.out_file + '_*'
ctx_files_pattern = args.encoded_ctx_file
input_paths = glob.glob(ctx_files_pattern)
logger.info('Reading all passages data from files: %s', input_paths)
retriever.index_encoded_data(input_paths, buffer_size=index_buffer_sz)
questions = []
question_answers = []
for ds_item in parse_qa_csv_file(args.qa_file):
question, answers = ds_item
questions.append(question)
question_answers.append(answers)
questions_tensor = retriever.generate_question_vectors(questions)
top_ids_and_scores = retriever.get_top_docs(questions_tensor.numpy(), args.n_docs, is_hnsw = args.hnsw_index)
all_passages = load_passages(args.ctx_file)
if len(all_passages) == 0:
raise RuntimeError('No passages data found. Please specify ctx_file param properly.')
log_file = os.path.join(os.path.dirname(args.model_file),os.path.basename(args.model_file) + '.' + os.path.basename(args.qa_file) + ".log.txt")
logger.info('Save to: %s', log_file)
with open(log_file,'w') as log_f:
for k in range(10, args.n_docs+1, 20):
recall_at_k = recall_k(top_ids_and_scores, topK=k)
log_f.write(f'Validation results: recall@{k}:{recall_at_k :.3f}\n')
else:
raise NotImplementedError()