merge-20k-subset-to-main

README.md

@@ -10,7 +10,7 @@ tags:
 - diffusiondb
 pretty_name: DiffusionDB text-to-image subset
 size_categories:
-- 10K<n<100K
+- 1K<n<10K
 ---
 
 # DiffusionDB text-to-image subset
@@ -18,36 +18,46 @@ size_categories:
 A cleaned, safety-filtered image-prompt dataset for training a text-to-image
 model, built from [DiffusionDB](https://huggingface.co/datasets/poloclub/diffusiondb).
 
-Built on Hugging Face Jobs directly from `poloclub/diffusiondb`. It covers
-`part_id` 1-20 (20,000 source images) before filtering. The same content is
-also kept on the `20k-subset` branch.
+**Note on the repo name**: despite "2M" in the name (matching the source
+Kaggle mirror's naming, `diffusiondb-2m-part-0001-to-0100-of-2000`), this
+release is a small sample drawn from `part_id` 1-100 of DiffusionDB's 2000
+parts — not the full ~2,000,000-image DiffusionDB-2M collection.
 
-Load it with:
+## Splits
 
-```python
-load_dataset("whosouravsharma/text-to-image-diffusiondb-2M")
-```
-
-**Note on the repo name**: despite "2M" in the name, this is a small slice of
-DiffusionDB's 2000 parts, not the full ~2,000,000-image collection.
+- `train`: 8,950 examples
+- `validation`: 182 examples
 
 ## Filtering applied
 
+Starting from the 100,000 images in `part_id` 1-100, a deterministic random
+sample of 10,000 (seed 42) was drawn, then:
 - Dropped missing/empty or very short (<3 char) prompts
 - Dropped images below 256px on either dimension
-- Dropped NSFW-flagged rows (image_nsfw >= 0.5 or prompt_nsfw >= 0.5)
+- Dropped NSFW-flagged rows (`image_nsfw >= 0.5` or `prompt_nsfw >= 0.5`)
 - Dropped exact-duplicate images (SHA256 hash match)
 
+leaving 9,132 examples, split 98/2 into train/validation.
+
 ## Columns
 
-`image`, `prompt`, `part_id`, `seed`, `step`, `cfg`, `sampler`, `width`,
-`height`, `image_nsfw`, `prompt_nsfw`. `user_name`/`timestamp` from the
-original DiffusionDB metadata are intentionally excluded.
+- `image`: the Stable Diffusion-generated image (PNG)
+- `prompt`: the user-written text prompt
+- `part_id`, `seed`, `step`, `cfg`, `sampler`, `width`, `height`: generation
+  parameters (`sampler` is the resolved name, e.g. `k_lms`, not the numeric
+  code used in the original metadata)
+- `image_nsfw`, `prompt_nsfw`: NSFW scores from the original DiffusionDB metadata
+
+`user_name` and `timestamp` from the original DiffusionDB metadata are
+intentionally excluded from this release.
 
 A standalone `metadata.parquet` (all rows, both splits, with a `split`
-column) is also included at the repo root.
+column) is also included at the repo root for reference outside the
+dataset loader.
 
 ## Source & license
 
 Derived from [poloclub/diffusiondb](https://huggingface.co/datasets/poloclub/diffusiondb),
-CC0-1.0 (public domain). Wang et al., 2022, arXiv:2210.14896.
+released under CC0-1.0 (public domain). See the original dataset card and
+paper (Wang et al., 2022, arXiv:2210.14896) for details on how DiffusionDB
+was collected.

data/{train-00000-of-00024.parquet → train-00000-of-00012.parquet}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:84bced5392c16a93b9a6f27f40f2db8d6acb1fa8e432e6dcd69bf854d0fd516b
-size 490771957
+oid sha256:9ea43223dd600efd009327c0e6dea370204feaaf1adc9b841c75ce08c004b0c6
+size 460002650

data/{train-00001-of-00024.parquet → train-00001-of-00012.parquet}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:79307ba368c6baa04580271be7c0620c622ce003f1408ca7f5d683101d52cf4b
-size 470788687
+oid sha256:e2743885afac5628c0c596cc66f3ad81bcdd46670f6cb2d21aaf77a3b13d3288
+size 468529823

data/{train-00002-of-00024.parquet → train-00002-of-00012.parquet}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:aee87148419833724cd582510bce3a2632fb0ced0637ff4ea76a7708dd0b2c3a
-size 479568147
+oid sha256:9d001751bdee73fe2c5b68934080b8505ae961745c195df75b5fc64a4acd7dfc
+size 456130158

data/{train-00003-of-00024.parquet → train-00003-of-00012.parquet}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:064bdb0b53fdaa031c3e831883aa414e7cf85f37b51b728d1d8cdd99f450ff27
-size 479903029
+oid sha256:854e76f247bd97e0b621a5e30eaf3687bf634831ad194bf5a21d0f06be577585
+size 457132991

data/train-00004-of-00012.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07632ae3ea8a47b55dda39b3dccd15a5b27bdc5cd5f7dd8977b305dd741be83b
+size 451229523

data/train-00004-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d7643754f7ccc182d71530f4261ca556b77b2645e0d7fb9f627f06decdec6234
-size 490878202

data/train-00005-of-00012.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e48abc55318653e40ca8c2c8b8c6f93e4108716628df1f33e44925e81421f4e9
+size 445939529

data/train-00005-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:09b6e1b8aba04d4f5ee7755367cc13542e149f186a8e2a6f64d77aaee77c71b8
-size 475450502

data/train-00006-of-00012.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02beb1de2e5121179790383370c118c127d20155bf823e58a348f9ee38d07449
+size 461711848

data/train-00006-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d1b7e5fed1b359de2cc55215f05fdaed8362a1d0f810ca8262a7e7c953253c1b
-size 468591605

data/train-00007-of-00012.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8cabb2b0a871d5e333fd4a7f7b7400feddf6f274cff0b4630fd2f4bf42dcec18
+size 468352156

data/train-00007-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c669e6789acd4d23d4b032704f3898c2cfeb01793b3e0994e3d7c111e4d1e0e0
-size 474088233

data/train-00008-of-00012.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03a7716dc5fbe7a1530394f8fd678a7d0a868fa6cc431d47b4cb451ca98286a3
+size 456448196

data/train-00008-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c9a5f3b8d2bdb2ce1d16b1057c7469f698d5aed56f246b3a2d490899b353e4bd
-size 475958933

data/train-00009-of-00012.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9d5c2f07cd8783ff9492bc7150eafa4b3fa117f248a2dbb92995d72eb641097
+size 467021520

data/train-00009-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:39ab851fea808e05376ce0d39808605e1ee433b6548b2a91aded8a2673668570
-size 467872419

data/train-00010-of-00012.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b169fb30078b20752a506a337b2771c1603b2ef547ebb3d2f8bf8251607853a0
+size 463259643

data/train-00010-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3511fb2d324e0128ebe0fd17d856aa4762098d3ef03cd65863d20770e1db61aa
-size 468611810

data/train-00011-of-00012.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7457cc9ba69c6ba8f7bc6158b7ed9b772f87461fad181f6ca6ac2abc35bec610
+size 471466411

data/train-00011-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3f90e89485b8c117090ec2ae15591fb642e2867b23528b2ea0696fa613414aad
-size 480352001

data/train-00012-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c168456f3aae6f6358c637017086494833de9cff0cf7606956638b63034d217b
-size 457305734

data/train-00013-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e4df2dff5687f74d93dbe444b5068589986b3522904a9b61cafb140a5e50e9ee
-size 474671660

data/train-00014-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:740d95ca4bb3b2ba02c357c01d526913521884d9ce3565e75cdab8f452eede80
-size 476901033

data/train-00015-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:652c2cacd90228513ddc2c90f09a06933b2465b1f8536bfedb93c1cb7215ae37
-size 492081682

data/train-00016-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:21d694a26155ba31eed77f689ac97af6e88f4ee9c02cd197741636d693a7e940
-size 484437521

data/train-00017-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c8c3ef7b26532239dd118bf540d08b7692f79e1e2ef1383044e04dc8b4a1b739
-size 472663511

data/train-00018-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:52608d633ca1535624c6f32b4098d7820d3af12eaa67cce6a0ab0be7396bb955
-size 482875407

data/train-00019-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9e0aadcd581046eddb0b0396776952cfa044da4eb36f50e245006c3158557376
-size 482863694

data/train-00020-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bc3cfc25b082d7e0b694173968aac085e69e64bbc6aa49b8f52e83229fc5405c
-size 463656366

data/train-00021-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b62e8218be2093e4034319920824a4ed8d90b929c8c88358050c962cd3a6f76f
-size 470895543

data/train-00022-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4a281df40f879438247027dda393aa8caea87f562612f7e531c09db2ffbbf2f3
-size 469882000

data/train-00023-of-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6ab0fd3a8c39edbb76529dadd03bb14deb357d36a364c1f49e47f3500433145e
-size 484482317

data/validation-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:55af1ac9a1e7804813963e439dce1dec059f0f5cfe45d15ec70def52b0cd11fb
-size 230606581
+oid sha256:c34a1fb94384ef900267a082a92c0fb3e1d00384227a4033e4cda85ea34efa25
+size 116107793

metadata.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b7497e2e7297992074404aec755710b463ff06d2395d3278cb1c4d6d460821f1
-size 2664055
+oid sha256:1ecfd12fa836f48e05970b9050c611b36fc6a03d3e23e33a95a54b3eccc53aad
+size 1486994