ihaeyong / drama-graph

Drama-Graph repository produces both knowledge base on drama scripts and video graph for Video Turing Test (VTT).

Git clone this repository.

>> git clone https://github.com/ihaeyong/drama-graph.git

Prepare env. and install requirements:

1. create conda env.
>> conda create -n vtt_env python=3.6
>> source activate vtt_env
2. install pytorch and lib. 
>> conda install pytorch==1.1.0 torchvision==0.3.0 cudatoolkit=9.0 -c pytorch
>> pip install -r requirements.txt

Download AnotherMissOh datasets

Unzip the anothermissoh dataset,

>> mkdir ./data
>> cd data
>> unzip datasets.zip

You can set your data path to 'Yolo_v2_pytorch/src/anotherMissOh_dataset.py' as follows:

img_path = './data/AnotherMissOh/AnotherMissOh_images/AnotherMissOh01/'
json_dir = './data/AnotherMissOh/AnotherMissOh_Visual/AnotherMissOh01_visual.json'

Drama-graph model

We mainly use YOLOv2-pytorch.

You could find all trained models in this link.

We finetuned YOLOv2 w.r.t 20 persons for about 50 epoches as follows:

Train model:

train the integrated model from scratch.

For place recognition, make 'pre_model' folder and put places365 pre-trained model.

>> ./scripts/train_models.sh #gpu

train sound event model from scratch.

>> ./scripts/train_sound_event.sh #gpu

The trained model is saved in ./sound_event_detection/checkpoint/torch_model.pt

Test(drama graph generation) model:

drama graph generation on the all episodes.

>> ./scripts/eval_models.sh #gpu

Evaluation

mAP for person

>> python eval_mAP.py -rtype person

mAP for behavior

>> python eval_mAP.py -rtype behave

mAP for face

>> python eval_mAP.py -rtype face

mAP for relation

>> python eval_mAP.py -rtype relation

mAP for object

>> python eval_mAP.py -rtype object

Accuracy for sound event and its visualization. sed_vis folder should be in the directory from which you run the file(./), so here the directory is (drama-graph/sed_vis/).

>> ./scripts/eval_sound_event.sh
>> ./scripts/inference_sound_event.sh

Performances

model	trainset	validation	test
person detection	54.2%	50.6%	47.3%
face detection	43.9%	25.83	26.6%
emotion	72.6%	80.6%	66.9%
behavior	17.43%	3.9%	4.89%
object detection	2.18%	1.17%	1.33%
predicate	88.1%	88.8%	85.9%
place	61.0%	41.1%	38.8%
sound event	89.6%	69.0%	62.5%

Acknowledgements

This work was supported by Institute for Information & communications Technology Promotion(IITP) grant funded by the Korea government(MSIT) (2017-0-01780, The technology development for event recognition/relational reasoning and learning knowledge based system for video understanding)

drama-graph's Issues

Replace person_model.py with this layer

drama-graph/lib/relation_model.py

Line 29 in 56ce65e

self.conv_person = nn.Conv2d(

conv --> conv_relation

drama-graph/lib/relation_model.py

Line 29 in e44436d

self.conv = nn.Conv2d(

Define an object model

drama-graph/lib/relation_model.py

Line 31 in 56ce65e

self.conv_objects = nn.Conv2d(

face model params save

drama-graph/models/train_model.py

Line 388 in 57ec0a1

torch.save(model.state_dict(),

save face model params respectively.

face model params

drama-graph/models/train_model.py

Line 157 in 57ec0a1

face_params = [p for n, p in model_face.named_parameters()

sort face detector model parameters only.

change layer name

self.stage3_conv_my --> self.stage3_conv_person

drama-graph/Yolo_v2_pytorch/src/yolo_tunning.py

Line 95 in 77966bf

output = self.stage3_conv_my(output_fmap)

Request on Knowledge graph README

Describe the format of input in more details.

• How can we input a question into this system as you show in example

Need for explanations on the output formats.

• what do the frame.gv / triple.gv mean?
• Also, each label (e.g., Entity, Topical_entity) represents what?

Minor: Add commands to install stanfordnlpcore.

Request on knowledge graph

코드 실행중 다음과 같은 에러가 발생합니다.

root@4a30c3e1a0c7:/new_drama_graph/knowledge_graph# ./run_knowledge_graph.sh
2020-11-03 18:37:37,602 __main__     INFO     Config: {'mode': 'subtitle', 'output_dir': 'data/output', 'preprocessing': {'load': False, 'substitle_file': 'data/input/AnotherMissOh_Scene_Subtitle', 'output_path': 'data/output/preprocessed.json', 'coreference': 'gold'}, 'extraction': {'load': False, 'oie': 'stanford', 'frame': 'frameBERT', 'output_path': 'data/output/extracted.json'}, 'background': {'load': False, 'input_path': 'data/background.txt', 'output_path': 'data/output/background_KB.json'}, 'graph': {'character_name': 'data/character.txt', 'use_backKB': True, 'visualization': 'all', 'package_path': '/usr/share/graphviz', 'graph_path': 'data/output/graphs', 'json_path': 'data/output/graph.json', 'character': 'all', 'only_use': 'None'}}
preprocessing..
Can't match Parentheses: Hey! What are you(Heeran) going to do if I drink this in one shot?)
Can't match Parentheses: You(Heayoung1)'re so gross!)
done..
extracting knowledge..
Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/urllib3/connection.py", line 141, in _new_conn
    (self.host, self.port), self.timeout, **extra_kw)
  File "/usr/lib/python3/dist-packages/urllib3/util/connection.py", line 83, in create_connection
    raise err
  File "/usr/lib/python3/dist-packages/urllib3/util/connection.py", line 73, in create_connection
    sock.connect(sa)
ConnectionRefusedError: [Errno 111] Connection refused

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/urllib3/connectionpool.py", line 601, in urlopen
    chunked=chunked)
  File "/usr/lib/python3/dist-packages/urllib3/connectionpool.py", line 357, in _make_request
    conn.request(method, url, **httplib_request_kw)
  File "/usr/lib/python3.6/http/client.py", line 1264, in request
    self._send_request(method, url, body, headers, encode_chunked)
  File "/usr/lib/python3.6/http/client.py", line 1310, in _send_request
    self.endheaders(body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.6/http/client.py", line 1259, in endheaders
    self._send_output(message_body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.6/http/client.py", line 1038, in _send_output
    self.send(msg)
  File "/usr/lib/python3.6/http/client.py", line 976, in send
    self.connect()
  File "/usr/lib/python3/dist-packages/urllib3/connection.py", line 166, in connect
    conn = self._new_conn()
  File "/usr/lib/python3/dist-packages/urllib3/connection.py", line 150, in _new_conn
    self, "Failed to establish a new connection: %s" % e)
urllib3.exceptions.NewConnectionError: <urllib3.connection.HTTPConnection object at 0x7fde63ab7630>: Failed to establish a new connection: [Errno 111] Connection refused

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/lib/python3.6/dist-packages/requests/adapters.py", line 449, in send
    timeout=timeout
  File "/usr/lib/python3/dist-packages/urllib3/connectionpool.py", line 639, in urlopen
    _stacktrace=sys.exc_info()[2])
  File "/usr/lib/python3/dist-packages/urllib3/util/retry.py", line 398, in increment
    raise MaxRetryError(_pool, url, error or ResponseError(cause))
urllib3.exceptions.MaxRetryError: HTTPConnectionPool(host='143.248.135.188', port=1106): Max retries exceeded with url: /frameBERT (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7fde63ab7630>: Failed to establish a new connection: [Errno 111] Connection refused',))

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "main.py", line 38, in <module>
    extractor = extractor(config, preprocessor.output)
  File "/new_drama_graph/knowledge_graph/knowledge_extraction/extractor.py", line 15, in __init__
    self.frame = frame(config, self.input)
  File "/new_drama_graph/knowledge_graph/knowledge_extraction/frame.py", line 15, in __init__
    self.output = self.run()
  File "/new_drama_graph/knowledge_graph/knowledge_extraction/frame.py", line 27, in run
    results = eng_frameBERT(sent['statement'])
  File "/new_drama_graph/knowledge_graph/knowledge_extraction/frame.py", line 8, in eng_frameBERT
    response = requests.post(host, data=data, verify=False)
  File "/usr/local/lib/python3.6/dist-packages/requests/api.py", line 119, in post
    return request('post', url, data=data, json=json, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/requests/api.py", line 61, in request
    return session.request(method=method, url=url, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/requests/sessions.py", line 530, in request
    resp = self.send(prep, **send_kwargs)
  File "/usr/local/lib/python3.6/dist-packages/requests/sessions.py", line 643, in send
    r = adapter.send(request, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/requests/adapters.py", line 516, in send
    raise ConnectionError(e, request=request)
requests.exceptions.ConnectionError: HTTPConnectionPool(host='143.248.135.188', port=1106): Max retries exceeded with url: /frameBERT (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7fde63ab7630>: Failed to establish a new connection: [Errno 111] Connection refused',))

서버가 닫혀서 생긴 문제인 것 같은데 해결 요청 드릴 수 있을까요?