kargaranamir
commited on
Commit
•
c53190e
1
Parent(s):
2a21b24
Update app.py
Browse files
app.py
CHANGED
@@ -153,8 +153,25 @@ def load_GlotLID_v2(model_name, file_name):
|
|
153 |
return model
|
154 |
|
155 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
156 |
model_1 = load_GlotLID_v1(constants.MODEL_NAME, "model_v1.bin")
|
157 |
model_2 = load_GlotLID_v2(constants.MODEL_NAME, "model_v2.bin")
|
|
|
|
|
|
|
158 |
|
159 |
# @st.cache_resource
|
160 |
def plot(label, prob):
|
@@ -189,7 +206,16 @@ def compute(sentences, version = 'v2'):
|
|
189 |
A list of language probablities and labels for the given sentences.
|
190 |
"""
|
191 |
progress_text = "Computing Language..."
|
192 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
193 |
my_bar = st.progress(0, text=progress_text)
|
194 |
|
195 |
probs = []
|
@@ -206,7 +232,7 @@ def compute(sentences, version = 'v2'):
|
|
206 |
output_label_language = output_label.split('_')[0]
|
207 |
|
208 |
# script control
|
209 |
-
if version in ['v2'] and output_label_language!= 'zxx':
|
210 |
main_script, all_scripts = get_script(sent)
|
211 |
output_label_script = output_label.split('_')[1]
|
212 |
|
@@ -247,8 +273,8 @@ with tab1:
|
|
247 |
|
248 |
version = st.radio(
|
249 |
"Choose model",
|
250 |
-
["v1", "v2"],
|
251 |
-
captions=["GlotLID version 1", "GlotLID version 2 (more data and languages)"],
|
252 |
index = 1,
|
253 |
key = 'version_tab1',
|
254 |
horizontal = True
|
|
|
153 |
return model
|
154 |
|
155 |
|
156 |
+
@st.cache_resource
|
157 |
+
def load_OpenLID():
|
158 |
+
model_path = hf_hub_download(repo_id='laurievb/OpenLID', filename='model.bin')
|
159 |
+
model = fasttext.load_model(model_path)
|
160 |
+
return model
|
161 |
+
|
162 |
+
|
163 |
+
@st.cache_resource
|
164 |
+
def load_NLLB():
|
165 |
+
model_path = hf_hub_download(repo_id='facebook/fasttext-language-identification', filename='model.bin')
|
166 |
+
model = fasttext.load_model(model_path)
|
167 |
+
return model
|
168 |
+
|
169 |
+
|
170 |
model_1 = load_GlotLID_v1(constants.MODEL_NAME, "model_v1.bin")
|
171 |
model_2 = load_GlotLID_v2(constants.MODEL_NAME, "model_v2.bin")
|
172 |
+
model_3 = load_OpenLID()
|
173 |
+
model_4 = load_NLLB()
|
174 |
+
|
175 |
|
176 |
# @st.cache_resource
|
177 |
def plot(label, prob):
|
|
|
206 |
A list of language probablities and labels for the given sentences.
|
207 |
"""
|
208 |
progress_text = "Computing Language..."
|
209 |
+
|
210 |
+
if version == 'nllb-218':
|
211 |
+
model_choice = model_4
|
212 |
+
elif version == 'openlid-201':
|
213 |
+
model_choice = model_3
|
214 |
+
elif version == 'v2':
|
215 |
+
model_choice = model_2
|
216 |
+
else:
|
217 |
+
model_choice = model_1
|
218 |
+
|
219 |
my_bar = st.progress(0, text=progress_text)
|
220 |
|
221 |
probs = []
|
|
|
232 |
output_label_language = output_label.split('_')[0]
|
233 |
|
234 |
# script control
|
235 |
+
if version in ['v2', 'openlid-201', 'nllb-218'] and output_label_language!= 'zxx':
|
236 |
main_script, all_scripts = get_script(sent)
|
237 |
output_label_script = output_label.split('_')[1]
|
238 |
|
|
|
273 |
|
274 |
version = st.radio(
|
275 |
"Choose model",
|
276 |
+
["nllb-218", "openlid-201", "v1", "v2"],
|
277 |
+
captions=["NLLB", "OpenLID", "GlotLID version 1", "GlotLID version 2 (more data and languages)"],
|
278 |
index = 1,
|
279 |
key = 'version_tab1',
|
280 |
horizontal = True
|