add debug
Browse files
app.py
CHANGED
@@ -126,22 +126,32 @@ def opus_trans(article, target_language):
|
|
126 |
|
127 |
|
128 |
def nllb_trans(article, target_language):
|
|
|
|
|
129 |
tokenizer = AutoTokenizer.from_pretrained("facebook/nllb-200-distilled-600M")
|
130 |
model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M")
|
131 |
inputs = tokenizer(article, return_tensors="pt")
|
132 |
|
133 |
if target_language == "English":
|
134 |
target_lang = "eng_Latn"
|
|
|
135 |
elif target_language == "Chinese":
|
136 |
target_lang = "zho_Hans"
|
|
|
137 |
|
138 |
-
|
139 |
-
|
140 |
-
|
141 |
-
|
142 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
143 |
|
144 |
-
return
|
145 |
|
146 |
|
147 |
def translate(article, toolkit, target_language):
|
|
|
126 |
|
127 |
|
128 |
def nllb_trans(article, target_language):
|
129 |
+
result_lang = detect_lang(article)
|
130 |
+
|
131 |
tokenizer = AutoTokenizer.from_pretrained("facebook/nllb-200-distilled-600M")
|
132 |
model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M")
|
133 |
inputs = tokenizer(article, return_tensors="pt")
|
134 |
|
135 |
if target_language == "English":
|
136 |
target_lang = "eng_Latn"
|
137 |
+
target_language = "en"
|
138 |
elif target_language == "Chinese":
|
139 |
target_lang = "zho_Hans"
|
140 |
+
target_language = "zh"
|
141 |
|
142 |
+
if result_lang != target_language:
|
143 |
+
translated_tokens = model.generate(
|
144 |
+
**inputs,
|
145 |
+
forced_bos_token_id=tokenizer.lang_code_to_id[target_lang],
|
146 |
+
max_length=30,
|
147 |
+
)
|
148 |
+
translated = tokenizer.batch_decode(
|
149 |
+
translated_tokens, skip_special_tokens=True
|
150 |
+
)[0]
|
151 |
+
else:
|
152 |
+
translated = "Error: You chose the same language as the article detected language. Please reselect language and try again."
|
153 |
|
154 |
+
return translated
|
155 |
|
156 |
|
157 |
def translate(article, toolkit, target_language):
|