Convert dataset to Parquet

Convert dataset to Parquet.

2m_random_1k/train-00000-of-00002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1158a834c319706ef214f44a6afcd36914be50139f026a64b819d871ed4c6312
+size 333323939

2m_random_1k/train-00001-of-00002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f215aa1b640f44c03b7af7b00404e9477ef9e346653257ec77c6539607915d8
+size 302226882

README.md

@@ -1,34 +1,72 @@
 ---
-layout: default
-title: Home
-nav_order: 1
-has_children: false
-
 annotations_creators:
 - no-annotation
-language:
-- en
 language_creators:
 - found
+language:
+- en
 license:
 - cc0-1.0
 multilinguality:
 - multilingual
-pretty_name: DiffusionDB
 size_categories:
 - n>1T
 source_datasets:
 - original
-tags:
-- stable diffusion
-- prompt engineering
-- prompts
-- research paper
 task_categories:
 - text-to-image
 - image-to-text
 task_ids:
 - image-captioning
+pretty_name: DiffusionDB
+layout: default
+title: Home
+nav_order: 1
+has_children: false
+tags:
+- stable diffusion
+- prompt engineering
+- prompts
+- research paper
+dataset_info:
+  config_name: 2m_random_1k
+  features:
+  - name: image
+    dtype: image
+  - name: prompt
+    dtype: string
+  - name: seed
+    dtype: uint32
+  - name: step
+    dtype: uint16
+  - name: cfg
+    dtype: float32
+  - name: sampler
+    dtype: string
+  - name: width
+    dtype: uint16
+  - name: height
+    dtype: uint16
+  - name: user_name
+    dtype: string
+  - name: timestamp
+    dtype: timestamp[us, tz=UTC]
+  - name: image_nsfw
+    dtype: float32
+  - name: prompt_nsfw
+    dtype: float32
+  splits:
+  - name: train
+    num_bytes: 635615655.0
+    num_examples: 1000
+  download_size: 635550821
+  dataset_size: 635615655.0
+configs:
+- config_name: 2m_random_1k
+  data_files:
+  - split: train
+    path: 2m_random_1k/train-*
+  default: true
 ---
 
 # DiffusionDB