Hugging Face's logo

Datasets:
jainr3
/
diffusiondb-pixelart

Tasks:
Text-to-Image
Image-to-Text
Sub-tasks:
image-captioning
Languages:
English
Multilinguality:
multilingual
Size Categories:
n>1T
Language Creators:
found
Annotations Creators:
no-annotation
Source Datasets:
modified
ArXiv:
Tags:
stable diffusion
prompt engineering
prompts
License:
Dataset card Viewer Files Community
jainr3 commited on
Commit
1c26606
•
1 Parent(s): 92873cb

Update diffusiondb-pixelart.py

Browse files
Files changed (1) hide show
  1. diffusiondb-pixelart.py +50 -50
diffusiondb-pixelart.py CHANGED
@@ -158,18 +158,18 @@ class DiffusionDB(datasets.GeneratorBasedBuilder):
158
  features = datasets.Features(
159
  {
160
  "image_name": datasets.Value("string"),
161
- "prompt": datasets.Value("string"),
162
- "part_id": datasets.Value("uint16"),
163
- "seed": datasets.Value("uint32"),
164
- "step": datasets.Value("uint16"),
165
- "cfg": datasets.Value("float32"),
166
- "sampler": datasets.Value("string"),
167
- "width": datasets.Value("uint16"),
168
- "height": datasets.Value("uint16"),
169
- "user_name": datasets.Value("string"),
170
- "timestamp": datasets.Value("timestamp[us, tz=UTC]"),
171
- "image_nsfw": datasets.Value("float32"),
172
- "prompt_nsfw": datasets.Value("float32"),
173
  },
174
  )
175
 
@@ -177,17 +177,17 @@ class DiffusionDB(datasets.GeneratorBasedBuilder):
177
  features = datasets.Features(
178
  {
179
  "image": datasets.Image(),
180
- "prompt": datasets.Value("string"),
181
- "seed": datasets.Value("uint32"),
182
- "step": datasets.Value("uint16"),
183
- "cfg": datasets.Value("float32"),
184
- "sampler": datasets.Value("string"),
185
- "width": datasets.Value("uint16"),
186
- "height": datasets.Value("uint16"),
187
- "user_name": datasets.Value("string"),
188
- "timestamp": datasets.Value("timestamp[us, tz=UTC]"),
189
- "image_nsfw": datasets.Value("float32"),
190
- "prompt_nsfw": datasets.Value("float32"),
191
  },
192
  )
193
 
@@ -252,20 +252,20 @@ class DiffusionDB(datasets.GeneratorBasedBuilder):
252
  for _, row in metadata_df.iterrows():
253
  yield row["image_name"], {
254
  "image_name": row["image_name"],
255
- "prompt": row["prompt"],
256
- "part_id": row["part_id"],
257
- "seed": row["seed"],
258
- "step": row["step"],
259
- "cfg": row["cfg"],
260
- "sampler": _SAMPLER_DICT[int(row["sampler"])],
261
- "width": row["width"],
262
- "height": row["height"],
263
- "user_name": row["user_name"],
264
- "timestamp": None
265
- if pd.isnull(row["timestamp"])
266
- else row["timestamp"],
267
- "image_nsfw": row["image_nsfw"],
268
- "prompt_nsfw": row["prompt_nsfw"],
269
  }
270
 
271
  else:
@@ -308,17 +308,17 @@ class DiffusionDB(datasets.GeneratorBasedBuilder):
308
  "path": img_path,
309
  "bytes": open(img_path, "rb").read(),
310
  },
311
- "prompt": img_params["p"],
312
- "seed": int(img_params["se"]),
313
- "step": int(img_params["st"]),
314
- "cfg": float(img_params["c"]),
315
- "sampler": img_params["sa"],
316
- "width": query_result["width"].to_list()[0],
317
- "height": query_result["height"].to_list()[0],
318
- "user_name": query_result["user_name"].to_list()[0],
319
- "timestamp": None
320
- if pd.isnull(query_result["timestamp"].to_list()[0])
321
- else query_result["timestamp"].to_list()[0],
322
- "image_nsfw": query_result["image_nsfw"].to_list()[0],
323
- "prompt_nsfw": query_result["prompt_nsfw"].to_list()[0],
324
  }
 
158
  features = datasets.Features(
159
  {
160
  "image_name": datasets.Value("string"),
161
+ "text": datasets.Value("string"),
162
+ #"part_id": datasets.Value("uint16"),
163
+ #"seed": datasets.Value("uint32"),
164
+ #"step": datasets.Value("uint16"),
165
+ #"cfg": datasets.Value("float32"),
166
+ #"sampler": datasets.Value("string"),
167
+ #"width": datasets.Value("uint16"),
168
+ #"height": datasets.Value("uint16"),
169
+ #"user_name": datasets.Value("string"),
170
+ #"timestamp": datasets.Value("timestamp[us, tz=UTC]"),
171
+ #"image_nsfw": datasets.Value("float32"),
172
+ #"prompt_nsfw": datasets.Value("float32"),
173
  },
174
  )
175
 
 
177
  features = datasets.Features(
178
  {
179
  "image": datasets.Image(),
180
+ "text": datasets.Value("string"),
181
+ #"seed": datasets.Value("uint32"),
182
+ #"step": datasets.Value("uint16"),
183
+ #"cfg": datasets.Value("float32"),
184
+ #"sampler": datasets.Value("string"),
185
+ #"width": datasets.Value("uint16"),
186
+ #"height": datasets.Value("uint16"),
187
+ #"user_name": datasets.Value("string"),
188
+ #"timestamp": datasets.Value("timestamp[us, tz=UTC]"),
189
+ #"image_nsfw": datasets.Value("float32"),
190
+ #"prompt_nsfw": datasets.Value("float32"),
191
  },
192
  )
193
 
 
252
  for _, row in metadata_df.iterrows():
253
  yield row["image_name"], {
254
  "image_name": row["image_name"],
255
+ "text": row["prompt"],
256
+ #"part_id": row["part_id"],
257
+ #"seed": row["seed"],
258
+ #"step": row["step"],
259
+ #"cfg": row["cfg"],
260
+ #"sampler": _SAMPLER_DICT[int(row["sampler"])],
261
+ #"width": row["width"],
262
+ #"height": row["height"],
263
+ #"user_name": row["user_name"],
264
+ #"timestamp": None
265
+ #if pd.isnull(row["timestamp"])
266
+ #else row["timestamp"],
267
+ #"image_nsfw": row["image_nsfw"],
268
+ #"prompt_nsfw": row["prompt_nsfw"],
269
  }
270
 
271
  else:
 
308
  "path": img_path,
309
  "bytes": open(img_path, "rb").read(),
310
  },
311
+ "text": img_params["p"],
312
+ #"seed": int(img_params["se"]),
313
+ #"step": int(img_params["st"]),
314
+ #"cfg": float(img_params["c"]),
315
+ #"sampler": img_params["sa"],
316
+ #"width": query_result["width"].to_list()[0],
317
+ #"height": query_result["height"].to_list()[0],
318
+ #"user_name": query_result["user_name"].to_list()[0],
319
+ #"timestamp": None
320
+ #if pd.isnull(query_result["timestamp"].to_list()[0])
321
+ #else query_result["timestamp"].to_list()[0],
322
+ #"image_nsfw": query_result["image_nsfw"].to_list()[0],
323
+ #"prompt_nsfw": query_result["prompt_nsfw"].to_list()[0],
324
  }