Syoy commited on
Commit
7efbe93
1 Parent(s): 8414268

updated readme

Browse files
Files changed (1) hide show
  1. README.md +25 -25
README.md CHANGED
@@ -51,14 +51,14 @@ tags:
51
  ## Dataset Description
52
 
53
  - **Homepage:** [Renumics Homepage](https://renumics.com/)
54
- - **Original Dataset Upload** [ZENODO: DCASE 2023 Challenge Task 2 Development Dataset](https://zenodo.org/record/7687464#.Y_9VtdLMLmE)
 
55
  - **Paper** [MIMII DG](https://arxiv.org/abs/2205.13879)
56
  - **Paper** [ToyADMOS2](https://arxiv.org/abs/2106.02369)
 
57
 
58
  ### Dataset Summary
59
 
60
- [//]: # (todo)
61
-
62
  [//]: # (todo: verantwortlichkeit)
63
 
64
  ### Explore the data with Spotlight
@@ -78,16 +78,18 @@ df = ds.to_pandas()
78
  spotlight.show(df, dtype={'path': spotlight.Audio})
79
  ```
80
 
81
- advanced view:
82
  ```jupyterpython
 
83
  from datasets import load_dataset, load_dataset_builder
84
  from renumics import spotlight
85
 
86
- ds = load_dataset("renumics/dcase23-task2-enriched", "dev", split="train", streaming=False)
 
87
  db = load_dataset_builder("renumics/dcase23-task2-enriched", "dev")
88
 
89
- df = db.config.to_spotlight(ds)
90
- spotlight.show(df, dtype={'audio': spotlight.Audio, 'baseline-embeddings': spotlight.Embedding}, layout=db.config.get_layout())
91
  ```
92
 
93
  [//]: # (todo: add embeddings column)
@@ -99,7 +101,6 @@ spotlight.show(df, dtype={'audio': spotlight.Audio, 'baseline-embeddings': spotl
99
  For each instance, there is a Audio for the audio, a string for the path, an integer for the section, a string for the d1p (parameter), a string for the d1v (value),
100
  a ClassLabel for the label and a ClassLabel for the class.
101
 
102
- [//]: # (todo)
103
  ```
104
  {'audio': {'array': array([ 0. , 0.00024414, -0.00024414, ..., -0.00024414,
105
  0. , 0. ], dtype=float32),
@@ -110,12 +111,13 @@ a ClassLabel for the label and a ClassLabel for the class.
110
  'section': 1
111
  'd1p': 'f-n'
112
  'd1v': 'A'
113
- 'd2p': ''
114
- 'd2v': ''
115
- 'd3p': ''
116
- 'd3v': ''
 
117
  'label': 0 (normal)
118
- 'class': 0 (fan)
119
  }
120
  ```
121
 
@@ -130,19 +132,18 @@ The length of each audio file is 10 seconds.
130
  - `section`: an integer
131
  - `d*p`:
132
  - `d*v`:
 
133
  - `class`:
134
  - `label`: an integer whose value may be either _0_, indicating that the audio sample is _normal_, _1_, indicating that the audio sample contains an _anomaly_.
135
 
136
- [//]: # (todo)
137
-
138
  ### Data Splits
139
 
140
  The development dataset has 2 splits: _train_ and _test_.
141
 
142
- | Dataset Split | Number of Instances in Split |
143
- | ------------- |----------------------------- |
144
- | Train | TODO |
145
- | Test | TODO |
146
 
147
  The information for the evaluation dataset will follow after release.
148
 
@@ -221,19 +222,17 @@ Not applicable.
221
 
222
  ### Baseline system
223
 
224
- [//]: # (todo)
225
- Two baseline systems are available on the Github repository. The baseline systems provide a simple entry-level approach that gives a reasonable performance in the dataset of Task 2. They are good starting points, especially for entry-level researchers who want to get familiar with the anomalous-sound-detection task.
226
 
227
  ### Dataset Curators
228
 
229
  [//]: # (todo)
230
  Example: The SNLI corpus was developed by Samuel R. Bowman, Gabor Angeli, Christopher Potts, and Christopher D. Manning as part of the [Stanford NLP group](https://nlp.stanford.edu/).
231
 
232
- It was supported by a Google Faculty Research Award, a gift from Bloomberg L.P., the Defense Advanced Research Projects Agency (DARPA) Deep Exploration and Filtering of Text (DEFT) Program under Air Force Research Laboratory (AFRL) contract no. FA8750-13-2-0040, the National Science Foundation under grant no. IIS 1159679, and the Department of the Navy, Office of Naval Research, under grant no. N00014-10-1-0109.
233
-
234
  ### Licensing Information - Condition of use
235
 
236
- The [original dataset](https://zenodo.org/record/7687464#.Y_9dd9LMLmH) was created jointly by **Hitachi, Ltd.** and **NTT Corporation** and is available under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license.
 
237
 
238
 
239
  ### Citation Information (original)
@@ -242,6 +241,7 @@ If you use this dataset, please cite all the following papers. We will publish a
242
 
243
  - Kota Dohi, Tomoya Nishida, Harsh Purohit, Ryo Tanabe, Takashi Endo, Masaaki Yamamoto, Yuki Nikaido, and Yohei Kawaguchi. MIMII DG: sound dataset for malfunctioning industrial machine investigation and inspection for domain generalization task. In arXiv e-prints: 2205.13879, 2022. [[URL](https://arxiv.org/abs/2205.13879)]
244
  - Noboru Harada, Daisuke Niizumi, Daiki Takeuchi, Yasunori Ohishi, Masahiro Yasuda, and Shoichiro Saito. ToyADMOS2: another dataset of miniature-machine operating sounds for anomalous sound detection under domain shift conditions. In Proceedings of the 6th Detection and Classification of Acoustic Scenes and Events 2021 Workshop (DCASE2021), 1–5. Barcelona, Spain, November 2021. [[URL](https://dcase.community/documents/workshop2021/proceedings/DCASE2021Workshop_Harada_6.pdf)]
 
245
 
246
  ```
247
  @dataset{kota_dohi_2023_7687464,
@@ -262,4 +262,4 @@ If you use this dataset, please cite all the following papers. We will publish a
262
  doi = {10.5281/zenodo.7687464},
263
  url = {https://doi.org/10.5281/zenodo.7687464}
264
  }
265
- ```
 
51
  ## Dataset Description
52
 
53
  - **Homepage:** [Renumics Homepage](https://renumics.com/)
54
+ - **Homepage** [DCASE23 Task 2 Challenge](https://dcase.community/challenge2023/task-first-shot-unsupervised-anomalous-sound-detection-for-machine-condition-monitoring#evaluation)
55
+ - **Original Dataset Upload (Dev)** [ZENODO: DCASE 2023 Challenge Task 2 Development Dataset](https://zenodo.org/record/7687464#.Y_9VtdLMLmE)
56
  - **Paper** [MIMII DG](https://arxiv.org/abs/2205.13879)
57
  - **Paper** [ToyADMOS2](https://arxiv.org/abs/2106.02369)
58
+ - **Paper** [First-shot anomaly detection for machine condition monitoring: A domain generalization baseline](https://arxiv.org/pdf/2303.00455.pdf)
59
 
60
  ### Dataset Summary
61
 
 
 
62
  [//]: # (todo: verantwortlichkeit)
63
 
64
  ### Explore the data with Spotlight
 
78
  spotlight.show(df, dtype={'path': spotlight.Audio})
79
  ```
80
 
81
+ to load a more advanced view, we can utilize spotlight's layout function and add clear names for the machine_types in the dataset.
82
  ```jupyterpython
83
+ import pandas as pd
84
  from datasets import load_dataset, load_dataset_builder
85
  from renumics import spotlight
86
 
87
+ train = load_dataset("renumics/dcase23-task2-enriched", "dev", split="train", streaming=False)
88
+ test = load_dataset("renumics/dcase23-task2-enriched", "dev", split="test", streaming=False)
89
  db = load_dataset_builder("renumics/dcase23-task2-enriched", "dev")
90
 
91
+ df = pd.concat([db.config.to_spotlight(train), db.config.to_spotlight(test)])
92
+ spotlight.show(df, dtype={'audio': spotlight.Audio}, layout=db.config.get_layout())
93
  ```
94
 
95
  [//]: # (todo: add embeddings column)
 
101
  For each instance, there is a Audio for the audio, a string for the path, an integer for the section, a string for the d1p (parameter), a string for the d1v (value),
102
  a ClassLabel for the label and a ClassLabel for the class.
103
 
 
104
  ```
105
  {'audio': {'array': array([ 0. , 0.00024414, -0.00024414, ..., -0.00024414,
106
  0. , 0. ], dtype=float32),
 
111
  'section': 1
112
  'd1p': 'f-n'
113
  'd1v': 'A'
114
+ 'd2p': 'nan'
115
+ 'd2v': 'nan'
116
+ 'd3p': 'nan'
117
+ 'd3v': 'nan'
118
+ 'domain': 0 (source)
119
  'label': 0 (normal)
120
+ 'class': 1 (fan)
121
  }
122
  ```
123
 
 
132
  - `section`: an integer
133
  - `d*p`:
134
  - `d*v`:
135
+ - `domain`:
136
  - `class`:
137
  - `label`: an integer whose value may be either _0_, indicating that the audio sample is _normal_, _1_, indicating that the audio sample contains an _anomaly_.
138
 
 
 
139
  ### Data Splits
140
 
141
  The development dataset has 2 splits: _train_ and _test_.
142
 
143
+ | Dataset Split | Number of Instances in Split | Source Domain / Target Domain Samples |
144
+ | ------------- |------------------------------|---------------------------------------|
145
+ | Train | 7000 | 6930 / 70 |
146
+ | Test | 1400 | 700 / 700 |
147
 
148
  The information for the evaluation dataset will follow after release.
149
 
 
222
 
223
  ### Baseline system
224
 
225
+ The baseline system is available on the Github repository [dcase2023_task2_baseline_ae](https://github.com/nttcslab/dase2023_task2_baseline_ae).The baseline systems provide a simple entry-level approach that gives a reasonable performance in the dataset of Task 2. They are good starting points, especially for entry-level researchers who want to get familiar with the anomalous-sound-detection task.
 
226
 
227
  ### Dataset Curators
228
 
229
  [//]: # (todo)
230
  Example: The SNLI corpus was developed by Samuel R. Bowman, Gabor Angeli, Christopher Potts, and Christopher D. Manning as part of the [Stanford NLP group](https://nlp.stanford.edu/).
231
 
 
 
232
  ### Licensing Information - Condition of use
233
 
234
+ This is a feature/embeddings-enriched version of the "DCASE 2023 Challenge Task 2 Development Dataset".
235
+ The [original dataset](https://dcase.community/challenge2023/task-first-shot-unsupervised-anomalous-sound-detection-for-machine-condition-monitoring#audio-datasets) was created jointly by **Hitachi, Ltd.** and **NTT Corporation** and is available under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license.
236
 
237
 
238
  ### Citation Information (original)
 
241
 
242
  - Kota Dohi, Tomoya Nishida, Harsh Purohit, Ryo Tanabe, Takashi Endo, Masaaki Yamamoto, Yuki Nikaido, and Yohei Kawaguchi. MIMII DG: sound dataset for malfunctioning industrial machine investigation and inspection for domain generalization task. In arXiv e-prints: 2205.13879, 2022. [[URL](https://arxiv.org/abs/2205.13879)]
243
  - Noboru Harada, Daisuke Niizumi, Daiki Takeuchi, Yasunori Ohishi, Masahiro Yasuda, and Shoichiro Saito. ToyADMOS2: another dataset of miniature-machine operating sounds for anomalous sound detection under domain shift conditions. In Proceedings of the 6th Detection and Classification of Acoustic Scenes and Events 2021 Workshop (DCASE2021), 1–5. Barcelona, Spain, November 2021. [[URL](https://dcase.community/documents/workshop2021/proceedings/DCASE2021Workshop_Harada_6.pdf)]
244
+ - Noboru Harada, Daisuke Niizumi, Daiki Takeuchi, Yasunori Ohishi, and Masahiro Yasuda. First-shot anomaly detection for machine condition monitoring: a domain generalization baseline. In arXiv e-prints: 2303.00455, 2023. [[URL](https://arxiv.org/abs/2303.00455.pdf)]
245
 
246
  ```
247
  @dataset{kota_dohi_2023_7687464,
 
262
  doi = {10.5281/zenodo.7687464},
263
  url = {https://doi.org/10.5281/zenodo.7687464}
264
  }
265
+ ```