質問編集履歴

2

修正

2022/11/11 23:03

投稿

shirouto3
shirouto3

スコア1

test CHANGED
File without changes
test CHANGED
@@ -3,45 +3,31 @@
3
3
  解決方法が知りたいです。質問が分かりにくいかもしれませんがよろしくお願いします。
4
4
 
5
5
  ```python
6
+ import torch
7
+ from torch import autocast
6
- >>> from diffusers import StableDiffusionPipeline, LMSDiscreteScheduler
8
+ from diffusers import StableDiffusionPipeline, LMSDiscreteScheduler
9
+
7
- >>> pipe = StableDiffusionPipeline.from_pretrained(
10
+ pipe = StableDiffusionPipeline.from_pretrained(
8
- ... "hakurei/waifu-diffusion",
11
+ "hakurei/waifu-diffusion",
9
- ... torch_dtype=torch.float16,
12
+ torch_dtype=torch.float16,
10
- ... scheduler=LMSDiscreteScheduler(
13
+ scheduler=LMSDiscreteScheduler(
11
- ... beta_start=0.00085,
14
+ beta_start=0.00085,
12
- ... beta_end=0.012,
15
+ beta_end=0.012,
13
- ... beta_schedule='scaled_linear',
16
+ beta_schedule='scaled_linear',
14
- ... num_train_timesteps=1000
17
+ num_train_timesteps=1000
15
- ... ),
18
+ ),
16
- ... ).to('cuda')
19
+ ).to('cuda')
17
-
18
- >>> import torch
20
+
19
- >>> from torch import autocast
20
- >>> from diffusers import StableDiffusionPipeline, LMSDiscreteScheduler
21
- >>>
22
- >>> pipe = StableDiffusionPipeline.from_pretrained(
23
- ... "hakurei/waifu-diffusion",
24
- ... torch_dtype=torch.float16, # GPUに余裕があるなら「torch.float32」
25
- ... scheduler=LMSDiscreteScheduler(
26
- ... beta_start=0.00085,
27
- ... beta_end=0.012,
28
- ... beta_schedule='scaled_linear',
29
- ... num_train_timesteps=1000
30
- ... ),
31
- ... ).to('cuda')
32
- >>>
33
- >>> prompt = "1girl, black eyes, black hair, black sweater, blue background, bob cut, closed mouth, glasses, medium hair, red-framed eyewear, simple background, solo, sweater, upper body, wide-eyed"
21
+ prompt = "1girl, black eyes, black hair, black sweater, blue background, bob cut, closed mouth, glasses, medium hair, red-framed eyewear, simple background, solo, sweater, upper body, wide-eyed"
34
- >>>
22
+
35
- >>> # ====ここから====
36
- >>> def null_safety(images, **kwargs):
23
+ def null_safety(images, **kwargs):
37
- ... return images, False
24
+ return images, False
38
- ...
25
+
39
- ...
26
+
40
- ... pipe.safety_checker = null_safety
27
+ pipe.safety_checker = null_safety
41
- >>> # ====ここまで====
28
+
42
- >>>
43
- >>> with autocast("cuda"):
29
+ with autocast("cuda"):
44
- ... image = pipe(prompt, guidance_scale=6)["images"][0]
30
+ image = pipe(prompt, guidance_scale=6)["images"][0]
45
31
  ```
46
32
 
47
33
  実行結果

1

修正

2022/11/10 23:18

投稿

shirouto3
shirouto3

スコア1

test CHANGED
File without changes
test CHANGED
@@ -14,23 +14,7 @@
14
14
  ... num_train_timesteps=1000
15
15
  ... ),
16
16
  ... ).to('cuda')
17
- Downloading: 100%|█████████████████████████████████████████████████████████████████████| 550/550 [00:00<00:00, 549kB/s]
17
+
18
- Downloading: 100%|█████████████████████████████████████████████████████████████████████| 342/342 [00:00<00:00, 339kB/s]
19
- Downloading: 100%|████████████████████████████████████████████████████████████████████████| 4.67k/4.67k [00:00<?, ?B/s]
20
- Downloading: 100%|████████████████████████████████████████████████████████████████| 1.22G/1.22G [04:24<00:00, 4.59MB/s]
21
- Downloading: 100%|█████████████████████████████████████████████████████████████████████| 215/215 [00:00<00:00, 214kB/s]
22
- Downloading: 100%|█████████████████████████████████████████████████████████████████████| 612/612 [00:00<00:00, 612kB/s]
23
- Downloading: 100%|██████████████████████████████████████████████████████████████████| 492M/492M [01:38<00:00, 5.02MB/s]
24
- Downloading: 100%|███████████████████████████████████████████████████████████████████| 525k/525k [00:04<00:00, 120kB/s]
25
- Downloading: 100%|█████████████████████████████████████████████████████████████████████| 472/472 [00:00<00:00, 472kB/s]
26
- Downloading: 100%|█████████████████████████████████████████████████████████████████████| 806/806 [00:00<00:00, 806kB/s]
27
- Downloading: 100%|█████████████████████████████████████████████████████████████████| 1.06M/1.06M [00:07<00:00, 137kB/s]
28
- Downloading: 100%|█████████████████████████████████████████████████████████████████████| 743/743 [00:00<00:00, 737kB/s]
29
- Downloading: 100%|████████████████████████████████████████████████████████████████| 3.44G/3.44G [12:43<00:00, 4.50MB/s]
30
- Downloading: 100%|█████████████████████████████████████████████████████████████████████| 547/547 [00:00<00:00, 543kB/s]
31
- Downloading: 100%|██████████████████████████████████████████████████████████████████| 335M/335M [00:50<00:00, 6.58MB/s]
32
- Fetching 15 files: 100%|███████████████████████████████████████████████████████████████| 15/15 [20:12<00:00, 80.86s/it]
33
- ftfy or spacy is not installed using BERT BasicTokenizer instead of ftfy.
34
18
  >>> import torch
35
19
  >>> from torch import autocast
36
20
  >>> from diffusers import StableDiffusionPipeline, LMSDiscreteScheduler
@@ -45,8 +29,6 @@
45
29
  ... num_train_timesteps=1000
46
30
  ... ),
47
31
  ... ).to('cuda')
48
- Fetching 15 files: 100%|█████████████████████████████████████████████████████████████| 15/15 [00:00<00:00, 7496.08it/s]
49
- ftfy or spacy is not installed using BERT BasicTokenizer instead of ftfy.
50
32
  >>>
51
33
  >>> prompt = "1girl, black eyes, black hair, black sweater, blue background, bob cut, closed mouth, glasses, medium hair, red-framed eyewear, simple background, solo, sweater, upper body, wide-eyed"
52
34
  >>>
@@ -56,14 +38,19 @@
56
38
  ...
57
39
  ...
58
40
  ... pipe.safety_checker = null_safety
59
- File "<stdin>", line 5
60
- pipe.safety_checker = null_safety
61
- ^^^^
62
- SyntaxError: invalid syntax
63
41
  >>> # ====ここまで====
64
42
  >>>
65
43
  >>> with autocast("cuda"):
66
44
  ... image = pipe(prompt, guidance_scale=6)["images"][0]
67
45
  ```
68
46
 
47
+ 実行結果
48
+ ```python
49
+ pipe.safety_checker = null_safety
50
+ File "<stdin>", line 5
51
+ pipe.safety_checker = null_safety
52
+ ^^^^
53
+ SyntaxError: invalid syntax
54
+ ```
69
55
 
56
+