Delete load_dataset.txt
Browse files- load_dataset.txt +0 -46
load_dataset.txt
DELETED
|
@@ -1,46 +0,0 @@
|
|
| 1 |
-
# Dataset Loading Script
|
| 2 |
-
# Save this as load_dataset.py to use
|
| 3 |
-
|
| 4 |
-
import csv
|
| 5 |
-
import os
|
| 6 |
-
from datasets import Dataset, Audio, Value, Features
|
| 7 |
-
|
| 8 |
-
def load_dataset():
|
| 9 |
-
# Define features
|
| 10 |
-
features = Features({
|
| 11 |
-
# Preserve original sampling rates by not forcing a fixed rate
|
| 12 |
-
"audio": Audio(sampling_rate=None),
|
| 13 |
-
"text": Value("string"),
|
| 14 |
-
"speaker_id": Value("string"),
|
| 15 |
-
"emotion": Value("string"),
|
| 16 |
-
"language": Value("string")
|
| 17 |
-
})
|
| 18 |
-
|
| 19 |
-
# Load data from CSV
|
| 20 |
-
data = {
|
| 21 |
-
"audio": [],
|
| 22 |
-
"text": [],
|
| 23 |
-
"speaker_id": [],
|
| 24 |
-
"emotion": [],
|
| 25 |
-
"language": []
|
| 26 |
-
}
|
| 27 |
-
|
| 28 |
-
# Read JSONL
|
| 29 |
-
import json
|
| 30 |
-
with open("data.jsonl", "r", encoding="utf-8") as f:
|
| 31 |
-
for line in f:
|
| 32 |
-
obj = json.loads(line)
|
| 33 |
-
data["audio"].append(obj["audio"]) # relative path within repo
|
| 34 |
-
data["text"].append(obj.get("text", ""))
|
| 35 |
-
data["speaker_id"].append(obj.get("speaker_id", ""))
|
| 36 |
-
data["emotion"].append(obj.get("emotion", "neutral"))
|
| 37 |
-
data["language"].append(obj.get("language", "en"))
|
| 38 |
-
|
| 39 |
-
# Create dataset
|
| 40 |
-
dataset = Dataset.from_dict(data, features=features)
|
| 41 |
-
return dataset
|
| 42 |
-
|
| 43 |
-
# For direct loading
|
| 44 |
-
if __name__ == "__main__":
|
| 45 |
-
dataset = load_dataset()
|
| 46 |
-
print(f"Dataset loaded with {len(dataset)} examples")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|