def do_translate(src_sentence: str) -> str:
hypotheses = beam_search(model, src_sentence,
beam_size=5,
max_decoding_time_step=70)
tgt_sentence = []
for src_sent, hyps in zip(src_sentence, hypotheses):
top_hyp = hyps[0]
hyp_sent = ' '.join(top_hyp.value)
tgt_sentence.append(hyp_sent + '\n')
return ''.join(tgt_sentence)
def send_ASL():
print(request.json['englishInputText'])
line = request.json['englishInputText']
outputText = ''
try:
outputText = ' '.join(
beam_search(nmt_model, [nltk.word_tokenize(line)],
beam_size=5,
max_decoding_time_step=70)[0][0].value)
except Exception as e:
print(e)
outputText = 'Could not translate...'
return {'outputText': outputText}
def decode(args: Dict[str, str]):
""" Performs decoding on the autograder test set
Make sure to run this code before submitting the code to the auto`grader
@param args (Dict): args from cmd line
"""
test_data_src = read_corpus(args['SOURCE_FILE'], source='src')
model = NMT.load(args['MODEL_PATH'])
if args['CUDA']:
model = model.to(torch.device("cuda:0"))
hypotheses = beam_search(model,
test_data_src,
beam_size=int(args['BEAM_SIZE']),
max_decoding_time_step=int(
args['MAX_DECODING_TIME_STEP']))
with open(args['OUTPUT_FILE'], 'w') as f:
for src_sent, hyps in zip(test_data_src, hypotheses):
top_hyp = hyps[0]
hyp_sent = ' '.join(top_hyp.value)
f.write(hyp_sent + '\n')