Spaces:
Running
Running
Fouzi Takelait
commited on
Commit
•
0dfd702
1
Parent(s):
b1c0f8d
Update app.py
Browse files
app.py
CHANGED
@@ -23,11 +23,7 @@ from transformer_mt_roberta.modeling_transformer_final import TransfomerEncoderD
|
|
23 |
# source_tokenizer = PreTrainedTokenizerFast.from_pretrained("da_en_output_dir/da_tokenizer")
|
24 |
# target_tokenizer = PreTrainedTokenizerFast.from_pretrained("da_en_output_dir/en_tokenizer")
|
25 |
# model = TransfomerEncoderDecoderModel.from_pretrained("da_en_output_dir")
|
26 |
-
|
27 |
-
#
|
28 |
-
=======
|
29 |
-
#
|
30 |
-
>>>>>>> adb80531e202c58b4ab91375bc391ab50bbc882f
|
31 |
# input_ids = source_tokenizer.encode(text_in, return_tensors="pt")
|
32 |
# output_ids = model.generate(
|
33 |
# input_ids,
|
@@ -36,11 +32,7 @@ from transformer_mt_roberta.modeling_transformer_final import TransfomerEncoderD
|
|
36 |
# eos_token_id=target_tokenizer.eos_token_id,
|
37 |
# pad_token_id=target_tokenizer.pad_token_id,
|
38 |
# )
|
39 |
-
|
40 |
-
#
|
41 |
-
=======
|
42 |
-
#
|
43 |
-
>>>>>>> adb80531e202c58b4ab91375bc391ab50bbc882f
|
44 |
# return target_tokenizer.decode(output_ids[0])
|
45 |
|
46 |
def translator_fn_roberta(text_in):
|
@@ -49,7 +41,7 @@ def translator_fn_roberta(text_in):
|
|
49 |
model_pretrained_roberta = mt_roberta.from_pretrained("da_en_RoBERTa_pretrained")
|
50 |
|
51 |
input_ids_pretrained_roberta = source_tokenizer_pretrained_roberta.encode(text_in, return_tensors="pt")
|
52 |
-
output_ids_pretrained_roberta =
|
53 |
input_ids_pretrained_roberta,
|
54 |
max_length=10,
|
55 |
bos_token_id=target_tokenizer_pretrained_roberta.bos_token_id,
|
|
|
23 |
# source_tokenizer = PreTrainedTokenizerFast.from_pretrained("da_en_output_dir/da_tokenizer")
|
24 |
# target_tokenizer = PreTrainedTokenizerFast.from_pretrained("da_en_output_dir/en_tokenizer")
|
25 |
# model = TransfomerEncoderDecoderModel.from_pretrained("da_en_output_dir")
|
26 |
+
|
|
|
|
|
|
|
|
|
27 |
# input_ids = source_tokenizer.encode(text_in, return_tensors="pt")
|
28 |
# output_ids = model.generate(
|
29 |
# input_ids,
|
|
|
32 |
# eos_token_id=target_tokenizer.eos_token_id,
|
33 |
# pad_token_id=target_tokenizer.pad_token_id,
|
34 |
# )
|
35 |
+
|
|
|
|
|
|
|
|
|
36 |
# return target_tokenizer.decode(output_ids[0])
|
37 |
|
38 |
def translator_fn_roberta(text_in):
|
|
|
41 |
model_pretrained_roberta = mt_roberta.from_pretrained("da_en_RoBERTa_pretrained")
|
42 |
|
43 |
input_ids_pretrained_roberta = source_tokenizer_pretrained_roberta.encode(text_in, return_tensors="pt")
|
44 |
+
output_ids_pretrained_roberta = model_pretrained_roberta.generate(
|
45 |
input_ids_pretrained_roberta,
|
46 |
max_length=10,
|
47 |
bos_token_id=target_tokenizer_pretrained_roberta.bos_token_id,
|