qanastek commited on
Commit
eb6442c
·
verified ·
1 Parent(s): 5aa2397

Upload 12 files

Browse files
README.md CHANGED
@@ -8,24 +8,26 @@ configs:
8
  path: data/train-00000-of-00001.parquet
9
  ---
10
 
11
- # Background Noise Dataset
12
 
13
- This dataset contains **3** audio recordings of **2** different background noise classes.
14
 
15
  ## Dataset Statistics
16
 
17
- - **Total Audio Files**: 3
18
- - **Total Classes**: 2
19
  - **Format**: WAV (converted to Parquet for Hugging Face Datasets)
20
 
21
  ## Classes and Descriptions
22
 
23
- The dataset covers the following background noises:
24
 
25
  | Label | Description |
26
  | :--- | :--- |
27
- | `fan_noise` | Fan noise background |
28
- | `white_noise` | White noise background |
 
 
29
 
30
  ## Structure
31
 
@@ -35,13 +37,13 @@ A `metadata.csv` file provides the file paths, labels, and descriptions.
35
  ### Columns
36
 
37
  - `audio`: Path to the audio file.
38
- - `label`: The category label of the background noise.
39
- - `description`: A text description of the background noise.
40
 
41
  ## Usage
42
 
43
  ```python
44
  from datasets import load_dataset
45
 
46
- ds = load_dataset("sdialog/background")
47
  ```
 
8
  path: data/train-00000-of-00001.parquet
9
  ---
10
 
11
+ # Foreground Noise Dataset
12
 
13
+ This dataset contains **4** audio recordings of **4** different foreground noise classes (AC noise variations).
14
 
15
  ## Dataset Statistics
16
 
17
+ - **Total Audio Files**: 4
18
+ - **Total Classes**: 4
19
  - **Format**: WAV (converted to Parquet for Hugging Face Datasets)
20
 
21
  ## Classes and Descriptions
22
 
23
+ The dataset covers the following foreground noises:
24
 
25
  | Label | Description |
26
  | :--- | :--- |
27
+ | `ac_noise_loud` | Loud AC noise |
28
+ | `ac_noise_low` | Low AC noise |
29
+ | `ac_noise_medium` | Medium AC noise |
30
+ | `ac_noise_minimal` | Minimal AC noise |
31
 
32
  ## Structure
33
 
 
37
  ### Columns
38
 
39
  - `audio`: Path to the audio file.
40
+ - `label`: The category label of the foreground noise.
41
+ - `description`: A text description of the foreground noise.
42
 
43
  ## Usage
44
 
45
  ```python
46
  from datasets import load_dataset
47
 
48
+ ds = load_dataset("sdialog/foreground")
49
  ```
ac_noise_loud/351790__reiyamanor__small-room-tone-vintage-air-condition-at.wav ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87fd0268b220978cef82ff911f86e372da3c145a63cf3b83bd352b96f7bfee88
3
+ size 40824738
ac_noise_low/351790__reiyamanor__small-room-tone-vintage-air-condition-at-15dB.wav ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b05aa59f4c2e0a37e83e04e5b40886e5e560af4d2bc5d14abefd9abc7272a08a
3
+ size 27072114
ac_noise_medium/351790__reiyamanor__small-room-tone-vintage-air-condition-at-10dB.wav ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:809378f31f1d0148efb76543b16a26a61c08d2a51505c70cbb99479f41bef3a7
3
+ size 27072114
ac_noise_minimal/351790__reiyamanor__small-room-tone-vintage-air-condition-at-20dB.wav ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a5cc3b5b246476a82c585043d68238e0f8e734418f925d5814284f1cc4ab4b9
3
+ size 27072114
convert_to_parquet.py CHANGED
@@ -1,28 +1,32 @@
1
- from datasets import Dataset, Audio, Features, Value, Version
2
  import os
3
  import csv
 
4
 
5
  # Point to the current directory
6
  dataset_path = os.path.dirname(os.path.abspath(__file__))
7
 
8
  # Dataset metadata
9
  VERSION = Version("1.0.0")
10
- DESCRIPTION = "Background noise dataset."
11
- HOMEPAGE = "https://huggingface.co/datasets/sdialog/background"
12
  LICENSE = "CC BY-NC-SA 4.0"
13
 
14
  # Define features
15
  features = Features({
16
  "audio": Audio(sampling_rate=16000),
 
17
  "label": Value("string"),
18
  "description": Value("string"),
19
  })
20
 
 
21
  def dataset_generator():
22
  csv_path = os.path.join(dataset_path, "metadata.csv")
23
  if not os.path.exists(csv_path):
24
- raise FileNotFoundError(f"{csv_path} not found. Run create_metadata.py first.")
25
-
 
 
26
  with open(csv_path, "r", encoding="utf-8") as f:
27
  reader = csv.DictReader(f)
28
  for row in reader:
@@ -30,17 +34,19 @@ def dataset_generator():
30
  # Ensure the path is absolute for reading
31
  # audio_path is relative to the dataset root
32
  full_path = os.path.join(dataset_path, audio_path)
33
-
34
  # Read audio bytes to ensure embedding
35
  with open(full_path, "rb") as audio_file:
36
  audio_bytes = audio_file.read()
37
-
38
  yield {
39
  "audio": {"path": None, "bytes": audio_bytes},
 
40
  "label": row["label"],
41
  "description": row["description"]
42
  }
43
 
 
44
  print("Creating dataset from generator (embedding bytes)...")
45
  ds = Dataset.from_generator(dataset_generator, features=features)
46
 
 
 
1
  import os
2
  import csv
3
+ from datasets import Dataset, Audio, Features, Value, Version
4
 
5
  # Point to the current directory
6
  dataset_path = os.path.dirname(os.path.abspath(__file__))
7
 
8
  # Dataset metadata
9
  VERSION = Version("1.0.0")
10
+ DESCRIPTION = "Foreground noise dataset."
11
+ HOMEPAGE = "https://huggingface.co/datasets/sdialog/foreground"
12
  LICENSE = "CC BY-NC-SA 4.0"
13
 
14
  # Define features
15
  features = Features({
16
  "audio": Audio(sampling_rate=16000),
17
+ "file_path": Value("string"),
18
  "label": Value("string"),
19
  "description": Value("string"),
20
  })
21
 
22
+
23
  def dataset_generator():
24
  csv_path = os.path.join(dataset_path, "metadata.csv")
25
  if not os.path.exists(csv_path):
26
+ raise FileNotFoundError(
27
+ f"{csv_path} not found. Run create_metadata.py first."
28
+ )
29
+
30
  with open(csv_path, "r", encoding="utf-8") as f:
31
  reader = csv.DictReader(f)
32
  for row in reader:
 
34
  # Ensure the path is absolute for reading
35
  # audio_path is relative to the dataset root
36
  full_path = os.path.join(dataset_path, audio_path)
37
+
38
  # Read audio bytes to ensure embedding
39
  with open(full_path, "rb") as audio_file:
40
  audio_bytes = audio_file.read()
41
+
42
  yield {
43
  "audio": {"path": None, "bytes": audio_bytes},
44
+ "file_path": audio_path,
45
  "label": row["label"],
46
  "description": row["description"]
47
  }
48
 
49
+
50
  print("Creating dataset from generator (embedding bytes)...")
51
  ds = Dataset.from_generator(dataset_generator, features=features)
52
 
create_metadata.py CHANGED
@@ -34,8 +34,7 @@ if os.path.exists(train_dir):
34
  if not os.path.isdir(label_dir) or label.startswith('.'):
35
  continue
36
 
37
- # Only process directories that are likely audio labels (simple heuristic or check against tags)
38
- # Here we just check if it's not the 'data' output folder or hidden
39
  if label == "data":
40
  continue
41
 
 
34
  if not os.path.isdir(label_dir) or label.startswith('.'):
35
  continue
36
 
37
+ # Only process directories that are likely audio labels
 
38
  if label == "data":
39
  continue
40
 
data/train-00000-of-00001.parquet CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e5760fdb2716b1c3b5ca88f4294eb90cca543bb23e15cec8c8f5c2293c98d3b4
3
- size 56836437
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15defd1a8836dbf9537a3088741348d75cb4aee1d37d85070bc70bfd237a453a
3
+ size 71949792
metadata.csv CHANGED
@@ -1,4 +1,5 @@
1
  audio,label,description
2
- fan_noise/210098__yuval__room-big-fan.wav,fan_noise,Fan noise background
3
- fan_noise/674563__klankbeeld__room-tone-fan-801am-220813_0497.wav,fan_noise,Fan noise background
4
- white_noise/white_zero_zero.wav,white_noise,White noise background
 
 
1
  audio,label,description
2
+ ac_noise_loud/351790__reiyamanor__small-room-tone-vintage-air-condition-at.wav,ac_noise_loud,Loud AC noise
3
+ ac_noise_low/351790__reiyamanor__small-room-tone-vintage-air-condition-at-15dB.wav,ac_noise_low,Low AC noise
4
+ ac_noise_medium/351790__reiyamanor__small-room-tone-vintage-air-condition-at-10dB.wav,ac_noise_medium,Medium AC noise
5
+ ac_noise_minimal/351790__reiyamanor__small-room-tone-vintage-air-condition-at-20dB.wav,ac_noise_minimal,Minimal AC noise
tags_labels.csv CHANGED
@@ -1,2 +1,4 @@
1
- fan_noise;Fan noise background
2
- white_noise;White noise background
 
 
 
1
+ ac_noise_loud;Loud AC noise
2
+ ac_noise_low;Low AC noise
3
+ ac_noise_medium;Medium AC noise
4
+ ac_noise_minimal;Minimal AC noise
test_dataset.py ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from datasets import load_dataset
2
+ import os
3
+
4
+ # Point to the parquet file generated
5
+ dataset_path = os.path.join(os.path.dirname(__file__), "data/train-00000-of-00001.parquet")
6
+
7
+ if not os.path.exists(dataset_path):
8
+ print(f"Dataset file not found at {dataset_path}")
9
+ print("Please run convert_to_parquet.py first.")
10
+ exit(1)
11
+
12
+ print(f"Loading foreground dataset from {dataset_path}...")
13
+ ds = load_dataset("parquet", data_files={"train": dataset_path}, split="train")
14
+
15
+ print("\nDataset Info:")
16
+ print(ds)
17
+
18
+ print("\nFeatures:")
19
+ print(ds.features)
20
+
21
+ print("\nFirst Example:")
22
+ item = ds[0]
23
+ print(item)
24
+
25
+ print("\nChecking file_path:")
26
+ print(f"File Path: {item['file_path']}")