BounharAbdelaziz
commited on
v0.1: remove create dataset, only load it.
Browse files
utils.py
CHANGED
@@ -56,58 +56,15 @@ def create_html_image(image_path):
|
|
56 |
"""
|
57 |
return html_string
|
58 |
|
59 |
-
# ---------------------------------------------------------------------------- #
|
60 |
-
# ---------------------------------------------------------------------------- #
|
61 |
-
|
62 |
-
def load_or_create_dataset():
|
63 |
-
try:
|
64 |
-
dataset = load_dataset(HF_DATASET_NAME, token=STT_EVAL_DATASET_TOKEN)
|
65 |
-
return dataset
|
66 |
-
except Exception as e:
|
67 |
-
print(f"[INFO] Dataset not found or error loading: {e}. Creating a new one.")
|
68 |
-
features = Features({
|
69 |
-
"timestamp": Value("string"),
|
70 |
-
"audio": Audio(sampling_rate=16000),
|
71 |
-
"model_used": Value("string"),
|
72 |
-
"transcription": Value("string")
|
73 |
-
})
|
74 |
-
dataset = Dataset.from_dict({
|
75 |
-
"timestamp": [],
|
76 |
-
"audio": [],
|
77 |
-
"model_used": [],
|
78 |
-
"transcription": []
|
79 |
-
}, features=features)
|
80 |
-
dataset = DatasetDict({
|
81 |
-
"train": dataset,
|
82 |
-
})
|
83 |
-
return dataset
|
84 |
|
85 |
# ---------------------------------------------------------------------------- #
|
86 |
# ---------------------------------------------------------------------------- #
|
87 |
|
88 |
def save_to_hf_dataset(audio_signal, model_choice, transcription):
|
89 |
print("[INFO] Loading dataset...")
|
90 |
-
|
91 |
-
|
92 |
-
|
93 |
-
except Exception as e:
|
94 |
-
print(f"[INFO] Dataset not found or error loading. Creating a new one.")
|
95 |
-
dataset = DatasetDict({
|
96 |
-
"train": Dataset.from_dict(
|
97 |
-
{
|
98 |
-
"audio": [],
|
99 |
-
"transcription": [],
|
100 |
-
"model_used": [],
|
101 |
-
"timestamp": [],
|
102 |
-
},
|
103 |
-
features=Features({
|
104 |
-
"audio": Audio(sampling_rate=16000),
|
105 |
-
"transcription": Value("string"),
|
106 |
-
"model_used": Value("string"),
|
107 |
-
"timestamp": Value("string"),
|
108 |
-
})
|
109 |
-
)
|
110 |
-
})
|
111 |
|
112 |
timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
|
113 |
new_entry = {
|
|
|
56 |
"""
|
57 |
return html_string
|
58 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
59 |
|
60 |
# ---------------------------------------------------------------------------- #
|
61 |
# ---------------------------------------------------------------------------- #
|
62 |
|
63 |
def save_to_hf_dataset(audio_signal, model_choice, transcription):
|
64 |
print("[INFO] Loading dataset...")
|
65 |
+
|
66 |
+
dataset = load_dataset(HF_DATASET_NAME, token=STT_EVAL_DATASET_TOKEN)
|
67 |
+
print("[INFO] Dataset loaded successfully.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
68 |
|
69 |
timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
|
70 |
new_entry = {
|