Datasets:

Formats:
parquet
Languages:
English
ArXiv:
Tags:
image
Libraries:
Datasets
Dask
License:
padge ppbrown commited on
Commit
867988b
·
verified ·
1 Parent(s): 9c269aa

Update tutorials/metadata.md (#6)

Browse files

- Update tutorials/metadata.md (b18cb828e5ec68865409a238bd944569e511f8d8)


Co-authored-by: Philip Brown <ppbrown@users.noreply.huggingface.co>

Files changed (1) hide show
  1. tutorials/metadata.md +20 -0
tutorials/metadata.md CHANGED
@@ -1,4 +1,13 @@
1
  # Working with the Metadata
 
 
 
 
 
 
 
 
 
2
  The metadata files are in parquet format, and contain the following attributes:
3
  - `id`: A unique identifier for the image.
4
  - `url`: The URL of the image.
@@ -24,3 +33,14 @@ Once you have opened a maetadata file with pandas, you can get the URLs of the i
24
  urls = df['url']
25
  ```
26
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  # Working with the Metadata
2
+
3
+ ## Downloading all the metadata files at once
4
+
5
+ Install the huggingface-cli utility (via pip). You may then use the following command:
6
+
7
+ huggingface-cli download Spawning/PD12M --repo-type dataset --local-dir metadata --include "metadata/*"
8
+
9
+ ## metadata format
10
+
11
  The metadata files are in parquet format, and contain the following attributes:
12
  - `id`: A unique identifier for the image.
13
  - `url`: The URL of the image.
 
33
  urls = df['url']
34
  ```
35
 
36
+ ### Download all files mentioned in metadata
37
+
38
+ If you want to just grab all files referenced by a metadata collection, you may try this (adjust to taste):
39
+
40
+ img2dataset --url_list $file --input_format "parquet" \
41
+ --url_col "url" --caption_col "caption" --output_format files \
42
+ --output_folder $dir --processes_count 16 --thread_count 64 \
43
+ --skip_reencode true --min_image_sizel 654 --max_aspect_ratio=1.77
44
+
45
+
46
+