p1atdev commited on
Commit
faaa32d
1 Parent(s): 7a11b50

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +82 -0
README.md CHANGED
@@ -246,6 +246,88 @@ If you want to avoid too realistic faces, try this!
246
  </details>
247
 
248
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
249
  ## License
250
  PVC v3 is released under the Fair AI Public License 1.0-SD (https://freedevproject.org/faipl-1.0-sd/). If any derivative of this model is made, please share your changes accordingly. Special thanks to ronsor/undeleted (https://undeleted.ronsor.com/) for help with the license.
251
 
 
246
  </details>
247
 
248
 
249
+ ## 🧨 Diffusers
250
+
251
+ Using the [🤗's Diffusers library](https://github.com/huggingface/diffusers) to run Stable Diffusion 2 in a simple and efficient manner.
252
+
253
+ ```bash
254
+ pip install diffusers transformers accelerate scipy safetensors
255
+ pip install --pre xformers
256
+ ```
257
+
258
+ Using StableDiffusionPipeline:
259
+
260
+ ```py
261
+ import torch
262
+ from diffusers import StableDiffusionPipeline
263
+
264
+ model_id = "p1atdev/pvc-v3"
265
+ revision = "fp16" # "main" or "fp16"
266
+
267
+ pipe = StableDiffusionPipeline.from_pretrained(
268
+ model_id,
269
+ revision=revision,
270
+ torch_dtype=torch.float16,
271
+ )
272
+ pipe = pipe.to("cuda")
273
+ pipe.enable_attention_slicing()
274
+ pipe.enable_xformers_memory_efficient_attention() # required
275
+
276
+ prompt = "pvc, masterpiece, best quality, exceptional, 1girl, cat ears, red hair, long hair, hairpin, swept bangs, yellow eyes, black jacket, white shirt, blue tie, white gloves, hand up, upper body, looking at viewer, buildings"
277
+ negative_prompt = "nsfw, nude, worst quality, low quality, oldest, bad anatomy"
278
+ image = pipe(
279
+ prompt,
280
+ negative_prompt=negative_prompt,
281
+ guidance_scale=7.0,
282
+ num_inference_steps=20
283
+ ).images[0]
284
+
285
+ # save image
286
+ image.save("pvc_figure.png")
287
+
288
+ # or just display it
289
+ # display(image)
290
+ ```
291
+
292
+ Using StableDiffusionLongPromptWeightingPipeline:
293
+
294
+ ```py
295
+ import torch
296
+ from diffusers import DiffusionPipeline
297
+
298
+ model_id = "p1atdev/pvc-v3"
299
+ revision = "fp16" # "main" or "fp16"
300
+
301
+ pipe = DiffusionPipeline.from_pretrained(
302
+ model_id,
303
+ revision=revision,
304
+ torch_dtype=torch.float16,
305
+ custom_pipeline="lpw_stable_diffusion"
306
+ )
307
+ pipe = pipe.to("cuda")
308
+ pipe.enable_attention_slicing()
309
+ pipe.enable_xformers_memory_efficient_attention() # required
310
+
311
+ prompt = """
312
+ pvc, anime, masterpiece, best quality, exceptional,
313
+ 1girl, bangs, bare shoulders, beret, black hair, black shorts, blue hair, bracelet, breasts, buttons,
314
+ colored inner hair, double-breasted, eyewear removed, green headwear, green jacket, grey eyes, grey sky,
315
+ hat, jacket, jewelry, long hair, looking at viewer, multicolored hair, neck ring, o-ring, off shoulder, rain,
316
+ round eyewear, shorts, sidelocks, small breasts, solo, sunglasses, wavy hair, wet, zipper
317
+ """ # long prompt
318
+
319
+ negative_prompt = "nsfw, nude, worst quality, low quality, oldest, bad anatomy"
320
+ image = pipe(
321
+ prompt,
322
+ negative_prompt=negative_prompt,
323
+ guidance_scale=7.0,
324
+ num_inference_steps=20
325
+ ).images[0]
326
+
327
+ display(image)
328
+ ```
329
+
330
+
331
  ## License
332
  PVC v3 is released under the Fair AI Public License 1.0-SD (https://freedevproject.org/faipl-1.0-sd/). If any derivative of this model is made, please share your changes accordingly. Special thanks to ronsor/undeleted (https://undeleted.ronsor.com/) for help with the license.
333