Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

feat: upd dp-ner with extended version #189

Merged
merged 10 commits into from
Aug 11, 2022
Merged

feat: upd dp-ner with extended version #189

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
4a933d2
feat: upd dp-ner with extended version
dilyararimovna Aug 5, 2022
b4da4e5
fix: upd tests
dilyararimovna Aug 8, 2022
ceb744f
fix: working for tags
dilyararimovna Aug 8, 2022
2580748
fix: codestyle
dilyararimovna Aug 8, 2022
dc4deea
fix: user new model
dilyararimovna Aug 8, 2022
b323d6b
feat: woking
dilyararimovna Aug 8, 2022
7c1c9b5
fix: config
dilyararimovna Aug 9, 2022
a0f6bec
fix: upd ner dockerfile
dilyararimovna Aug 9, 2022
9b6e4d0
fix: revert format list
dilyararimovna Aug 11, 2022
b070b4b
fix: change ner for all dists
dilyararimovna Aug 11, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions annotators/NER_deeppavlov/Dockerfile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,5 +16,6 @@ COPY $SRC_DIR /src
WORKDIR /src

RUN python -m deeppavlov install $CONFIG
RUN python -m deeppavlov download $CONFIG

CMD gunicorn --workers=1 --timeout 500 server:app -b 0.0.0.0:8021
157 changes: 157 additions & 0 deletions annotators/NER_deeppavlov/ner_case_agnostic_multilingual_bert_base_extended.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,157 @@
{
"dataset_reader": {
"class_name": "conll2003_reader",
"data_path": "{DOWNLOADS_PATH}/conll2003/",
"dataset_name": "conll2003",
"provide_pos": false
},
"dataset_iterator": {
"class_name": "data_learning_iterator"
},
"chainer": {
"in": [
"x"
],
"in_y": [
"y"
],
"pipe": [
{
"class_name": "torch_transformers_ner_preprocessor",
"vocab_file": "{TRANSFORMER}",
"do_lower_case": false,
"max_seq_length": 512,
"max_subword_length": 15,
"token_masking_prob": 0.0,
"in": [
"x"
],
"out": [
"x_tokens",
"x_subword_tokens",
"x_subword_tok_ids",
"startofword_markers",
"attention_mask"
]
},
{
"id": "tag_vocab",
"class_name": "simple_vocab",
"unk_token": [
"O"
],
"pad_with_zeros": true,
"save_path": "{MODEL_PATH}/tag.dict",
"load_path": "{MODEL_PATH}/tag.dict",
"fit_on": [
"y"
],
"in": [
"y"
],
"out": [
"y_ind"
]
},
{
"class_name": "torch_transformers_sequence_tagger",
"n_tags": "#tag_vocab.len",
"pretrained_bert": "{TRANSFORMER}",
"attention_probs_keep_prob": 0.5,
"return_probas": false,
"use_crf": true,
"encoder_layer_ids": [
-1
],
"optimizer": "AdamW",
"optimizer_parameters": {
"lr": 2e-05,
"weight_decay": 1e-06,
"betas": [
0.9,
0.999
],
"eps": 1e-06
},
"clip_norm": 1.0,
"min_learning_rate": 1e-07,
"learning_rate_drop_patience": 20,
"learning_rate_drop_div": 1.5,
"load_before_drop": true,
"save_path": "{MODEL_PATH}/model",
"load_path": "{MODEL_PATH}/model",
"in": [
"x_subword_tok_ids",
"attention_mask",
"startofword_markers"
],
"in_y": [
"y_ind"
],
"out": [
"y_pred_ind",
"probas"
]
},
{
"ref": "tag_vocab",
"in": [
"y_pred_ind"
],
"out": [
"y_pred"
]
}
],
"out": [
"x_tokens",
"y_pred"
]
},
"train": {
"epochs": 50,
"batch_size": 100,
"metrics": [
{
"name": "ner_f1",
"inputs": [
"y",
"y_pred"
]
},
{
"name": "ner_token_f1",
"inputs": [
"y",
"y_pred"
],
"print_results": true
}
],
"validation_patience": 100,
"val_every_n_batches": 50,
"log_every_n_batches": 50,
"show_examples": false,
"pytest_max_batches": 2,
"pytest_batch_size": 8,
"evaluation_targets": [
"test"
],
"class_name": "torch_trainer"
},
"metadata": {
"variables": {
"ROOT_PATH": "~/.deeppavlov",
"DOWNLOADS_PATH": "~/.deeppavlov/downloads",
"MODELS_PATH": "~/.deeppavlov/models",
"TRANSFORMER": "bert-base-multilingual-cased",
"MODEL_PATH": "{MODELS_PATH}/ner/mbert_dream_with_numbers_rus_ext"
},
"download": [
{
"url": "http://files.deeppavlov.ai/v1/ner/mbert_dream_with_numbers.tar.gz",
"subdir": "{MODEL_PATH}"
}
]
}
}
156 changes: 156 additions & 0 deletions annotators/NER_deeppavlov/ner_case_agnostic_multilingual_distilbert_extended.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,156 @@
{
"dataset_reader": {
"class_name": "conll2003_reader",
"data_path": "{DOWNLOADS_PATH}/conll2003/",
"dataset_name": "conll2003",
"provide_pos": false
},
"dataset_iterator": {
"class_name": "data_learning_iterator"
},
"chainer": {
"in": [
"x"
],
"in_y": [
"y"
],
"pipe": [
{
"class_name": "torch_transformers_ner_preprocessor",
"vocab_file": "{TRANSFORMER}",
"do_lower_case": false,
"max_seq_length": 512,
"max_subword_length": 15,
"token_masking_prob": 0.0,
"in": [
"x"
],
"out": [
"x_tokens",
"x_subword_tokens",
"x_subword_tok_ids",
"startofword_markers",
"attention_mask"
]
},
{
"id": "tag_vocab",
"class_name": "simple_vocab",
"unk_token": [
"O"
],
"pad_with_zeros": true,
"save_path": "{MODEL_PATH}/tag.dict",
"load_path": "{MODEL_PATH}/tag.dict",
"fit_on": [
"y"
],
"in": [
"y"
],
"out": [
"y_ind"
]
},
{
"class_name": "torch_transformers_sequence_tagger",
"n_tags": "#tag_vocab.len",
"pretrained_bert": "{TRANSFORMER}",
"attention_probs_keep_prob": 0.5,
"return_probas": false,
"use_crf": true,
"encoder_layer_ids": [
-1
],
"optimizer": "AdamW",
"optimizer_parameters": {
"lr": 2e-05,
"weight_decay": 1e-06,
"betas": [
0.9,
0.999
],
"eps": 1e-06
},
"clip_norm": 1.0,
"min_learning_rate": 1e-07,
"learning_rate_drop_patience": 20,
"learning_rate_drop_div": 1.5,
"load_before_drop": true,
"save_path": "{MODEL_PATH}/model",
"load_path": "{MODEL_PATH}/model",
"in": [
"x_subword_tok_ids",
"attention_mask",
"startofword_markers"
],
"in_y": [
"y_ind"
],
"out": [
"y_pred_ind",
"probas"
]
},
{
"ref": "tag_vocab",
"in": [
"y_pred_ind"
],
"out": [
"y_pred"
]
}
],
"out": [
"x_tokens",
"y_pred"
]
},
"train": {
"epochs": 50,
"batch_size": 150,
"metrics": [
{
"name": "ner_f1",
"inputs": [
"y",
"y_pred"
]
},
{
"name": "ner_token_f1",
"inputs": [
"y",
"y_pred"
]
}
],
"validation_patience": 100,
"val_every_n_batches": 50,
"log_every_n_batches": 50,
"show_examples": false,
"pytest_max_batches": 2,
"pytest_batch_size": 8,
"evaluation_targets": [
"test"
],
"class_name": "torch_trainer"
},
"metadata": {
"variables": {
"ROOT_PATH": "~/.deeppavlov",
"DOWNLOADS_PATH": "~/.deeppavlov/downloads",
"MODELS_PATH": "~/.deeppavlov/models",
"TRANSFORMER": "distilbert-base-multilingual-cased",
"MODEL_PATH": "{MODELS_PATH}/ner/mbert_dream_distil_with_numbers_rus_ext"
},
"download": [
{
"url": "http://files.deeppavlov.ai/v1/ner/mbert_dream_distil_with_numbers.tar.gz",
"subdir": "{MODEL_PATH}"
}
]
}
}
59 changes: 46 additions & 13 deletions annotators/NER_deeppavlov/server.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,15 +27,50 @@
raise e


def convert_prediction(s, token, tag):
start_pos = s.find(token)
return {
"confidence": 1,
"text": token,
"type": tag.replace("B-", "").replace("I-", ""),
"start_pos": start_pos,
"end_pos": start_pos + len(token),
}
def convert_prediction(sents, pred_labels):
entities = []
for sent, tags in zip(sents, pred_labels):
entities.append([])
start = end = -1
for i, (word, tag) in enumerate(zip(sent, tags)):
if tag[0] == "B":
if start != -1:
entities[-1].append(
{
"start_pos": start,
"end_pos": end,
"type": tags[start].split("-")[1],
"text": " ".join(sent[start:end]),
"confidence": 1,
}
)
start = i
end = i + 1
elif tag[0] == "I":
end = i + 1
else:
if start != -1:
entities[-1].append(
{
"start_pos": start,
"end_pos": end,
"type": tags[start].split("-")[1],
"text": " ".join(sent[start:end]),
"confidence": 1,
}
)
start = -1
if start != -1:
entities[-1].append(
{
"start_pos": start,
"end_pos": end,
"type": tags[start].split("-")[1],
"text": " ".join(sent[start:end]),
"confidence": 1,
}
)
return entities


def get_result(request):
Expand All @@ -51,10 +86,8 @@ def get_result(request):
dialog_ids.append(i)

tokens_batch, tags_batch = ner_model(samples)
good_preds = [
[convert_prediction(s, token, tag) for token, tag in zip(tokens, tags) if tag != "O"]
for s, tokens, tags in zip(samples, tokens_batch, tags_batch)
]
logger.info(f"NER model predictions: tokens: {tokens_batch}, tags: {tags_batch}")
good_preds = convert_prediction(tokens_batch, tags_batch)
dialog_ids = np.array(dialog_ids)

ret = []
Expand Down
Loading