Spaces:
Running
on
Zero
Running
on
Zero
BounharAbdelaziz
commited on
v0.1: removed save to eval dataset for now
Browse files
utils.py
CHANGED
@@ -15,10 +15,15 @@ from datasets import (
|
|
15 |
Audio,
|
16 |
)
|
17 |
|
|
|
|
|
18 |
|
19 |
# Hugging Face evaluation dataset
|
20 |
HF_DATASET_NAME = "BounharAbdelaziz/Moroccan-STT-Eval-Dataset"
|
21 |
|
|
|
|
|
|
|
22 |
# Models paths
|
23 |
MODEL_PATHS = {
|
24 |
"NANO": "BounharAbdelaziz/Morocco-Darija-STT-tiny",
|
@@ -26,6 +31,9 @@ MODEL_PATHS = {
|
|
26 |
"LARGE": "BounharAbdelaziz/Morocco-Darija-STT-large-v1.2",
|
27 |
}
|
28 |
|
|
|
|
|
|
|
29 |
# Access token to models
|
30 |
STT_MODEL_TOKEN = os.environ.get("STT_MODEL_TOKEN")
|
31 |
|
@@ -56,7 +64,6 @@ def create_html_image(image_path):
|
|
56 |
"""
|
57 |
return html_string
|
58 |
|
59 |
-
|
60 |
# ---------------------------------------------------------------------------- #
|
61 |
# ---------------------------------------------------------------------------- #
|
62 |
|
@@ -105,6 +112,9 @@ def load_model(model_name):
|
|
105 |
# ---------------------------------------------------------------------------- #
|
106 |
|
107 |
def process_audio(audio, model_choice, save_data):
|
|
|
|
|
|
|
108 |
pipe = load_model(model_choice)
|
109 |
audio_signal = audio[1]
|
110 |
sample_rate = audio[0]
|
@@ -161,7 +171,7 @@ def create_interface():
|
|
161 |
with gr.Row():
|
162 |
save_data = gr.Checkbox(
|
163 |
label="Contribute to the evaluation benchmark",
|
164 |
-
value=
|
165 |
)
|
166 |
|
167 |
submit_btn = gr.Button("Transcribe 🔥")
|
|
|
15 |
Audio,
|
16 |
)
|
17 |
|
18 |
+
# ---------------------------------------------------------------------------- #
|
19 |
+
# ---------------------------------------------------------------------------- #
|
20 |
|
21 |
# Hugging Face evaluation dataset
|
22 |
HF_DATASET_NAME = "BounharAbdelaziz/Moroccan-STT-Eval-Dataset"
|
23 |
|
24 |
+
# ---------------------------------------------------------------------------- #
|
25 |
+
# ---------------------------------------------------------------------------- #
|
26 |
+
|
27 |
# Models paths
|
28 |
MODEL_PATHS = {
|
29 |
"NANO": "BounharAbdelaziz/Morocco-Darija-STT-tiny",
|
|
|
31 |
"LARGE": "BounharAbdelaziz/Morocco-Darija-STT-large-v1.2",
|
32 |
}
|
33 |
|
34 |
+
# ---------------------------------------------------------------------------- #
|
35 |
+
# ---------------------------------------------------------------------------- #
|
36 |
+
|
37 |
# Access token to models
|
38 |
STT_MODEL_TOKEN = os.environ.get("STT_MODEL_TOKEN")
|
39 |
|
|
|
64 |
"""
|
65 |
return html_string
|
66 |
|
|
|
67 |
# ---------------------------------------------------------------------------- #
|
68 |
# ---------------------------------------------------------------------------- #
|
69 |
|
|
|
112 |
# ---------------------------------------------------------------------------- #
|
113 |
|
114 |
def process_audio(audio, model_choice, save_data):
|
115 |
+
|
116 |
+
# Force to false for now, issue with dataset
|
117 |
+
save_data = False
|
118 |
pipe = load_model(model_choice)
|
119 |
audio_signal = audio[1]
|
120 |
sample_rate = audio[0]
|
|
|
171 |
with gr.Row():
|
172 |
save_data = gr.Checkbox(
|
173 |
label="Contribute to the evaluation benchmark",
|
174 |
+
value=False,
|
175 |
)
|
176 |
|
177 |
submit_btn = gr.Button("Transcribe 🔥")
|