thomasavare commited on
Commit
45e467a
·
verified ·
1 Parent(s): e277424

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +8 -4
README.md CHANGED
@@ -15,13 +15,13 @@ dataset_info:
15
  dtype: float64
16
  splits:
17
  - name: train
18
- num_bytes: 515094059.0
19
  num_examples: 750
20
  - name: test
21
- num_bytes: 171698025.0
22
  num_examples: 250
23
  download_size: 590736573
24
- dataset_size: 686792084.0
25
  configs:
26
  - config_name: default
27
  data_files:
@@ -29,8 +29,12 @@ configs:
29
  path: data/train-*
30
  - split: test
31
  path: data/test-*
 
 
 
 
32
  ---
33
 
34
  concatentation of "thomasavare/waste-classification-audio-deepl" and "thomasavare/waste-classification-audio-deepl2", so its easier to use.
35
 
36
- there are 3 duplicates, idk why but it's too long to remake the datasets (maybe later) and not necessary.
 
15
  dtype: float64
16
  splits:
17
  - name: train
18
+ num_bytes: 515094059
19
  num_examples: 750
20
  - name: test
21
+ num_bytes: 171698025
22
  num_examples: 250
23
  download_size: 590736573
24
+ dataset_size: 686792084
25
  configs:
26
  - config_name: default
27
  data_files:
 
29
  path: data/train-*
30
  - split: test
31
  path: data/test-*
32
+ license: cc-by-nc-nd-4.0
33
+ language:
34
+ - it
35
+ - en
36
  ---
37
 
38
  concatentation of "thomasavare/waste-classification-audio-deepl" and "thomasavare/waste-classification-audio-deepl2", so its easier to use.
39
 
40
+ there are 3 duplicates, idk why but it's too long to remake the datasets (maybe later) and not necessary.