AWS BedrockのStable Diffusion XLで画像生成やーる

投稿日 2023年11月1日
著者 aws-recipe-user
カテゴリー Uncategorized

はじめに

AWS BedrockのStable Diffusion XLで画像生成をやっていきます。こちらの記事を参考にしています。

Bedrock APIを使用して画像から画像を作成してみました | DevelopersIO

3ヶ月ほど前にオンラインイベントで初めてAmazon Bedrockのことを知り、それから試してみたいと思いました。ついに先週リリースされたので、私が最初の試みでやったことを…

dev.classmethod.jp

導入

1．IAMロールを作成
2．アクセスキーとシークレットキーをメモ
3．ポリシーを作成

{
	"Version": "2012-10-17",
	"Statement": [
		{
			"Sid": "Statement1",
			"Effect": "Allow",
			"Action": [
				"bedrock:*"
			],
			"Resource": "*"
		}
	]
}

				
					
				1
2
3
4
5
6
7
8
9
10
11
12
13

						{
	"Version": "2012-10-17",
	"Statement": [
		{
			"Sid": "Statement1",
			"Effect": "Allow",
			"Action": [
				"bedrock:*"
			],
			"Resource": "*"
		}
	]
}

					

			

4．AWS BedrockのModel access->EditからStable Diffusion XLにチェック
5．AWS CLIをインストール

AWS CLI をシステムにインストールします。

docs.aws.amazon.com

6．コマンドプロンプトを開く

C:\Users\admin&gt;aws --version
aws-cli/2.13.30 Python/3.11.6 Windows/10 exe/AMD64 prompt/off

				1
2

						C:\Users\admin&gt;aws --version
aws-cli/2.13.30 Python/3.11.6 Windows/10 exe/AMD64 prompt/off

7．aws configureでアクセスキーとシークレットキーを設定

C:\Users\admin&gt;aws configure
AWS Access Key ID [****************]: 
AWS Secret Access Key [****************]: 
Default region name [ap-northeast-1]:
Default output format [None]: json

				
					
				1
2
3
4
5

						C:\Users\admin&gt;aws configure
AWS Access Key ID [****************]: 
AWS Secret Access Key [****************]: 
Default region name [ap-northeast-1]:
Default output format [None]: json

					

			

8．anaconda promptを開き、Python 3.11環境を作成

(base) C:\Users\admin\Documents\PythonProjects&gt;conda create -n py311 python=3.11
(base) C:\Users\admin\Documents\PythonProjects&gt;conda activate py311 

				1
2

						(base) C:\Users\admin\Documents\PythonProjects&gt;conda create -n py311 python=3.11
(base) C:\Users\admin\Documents\PythonProjects&gt;conda activate py311

9．ライブラリのインストール

(py311) C:\Users\admin\Documents\PythonProjects>pip install --no-build-isolation --force-reinstall boto3>=1.28.57 awscli>=1.29.57 botocore>=1.31.57
(py311) C:\Users\admin\Documents\PythonProjects>pip install pillow==9.5

				1
2

						(py311) C:\Users\admin\Documents\PythonProjects>pip install --no-build-isolation --force-reinstall boto3>=1.28.57 awscli>=1.29.57 botocore>=1.31.57
(py311) C:\Users\admin\Documents\PythonProjects>pip install pillow==9.5

10．bedrockのツールをダウンロード

11．utilsをプロジェクトフォルダに置く

12．app.pyを作成。os.environ[“BEDROCK_ASSUME_ROLE”]はコメントアウトしてOK。

# Python Built-Ins:
import base64
import io
import json
import os
import sys

# External Dependencies:
import boto3
from PIL import Image

module_path = ".."
sys.path.append(os.path.abspath(module_path))
from utils import bedrock, print_ww


# ---- ⚠️ Un-comment and edit the below lines as needed for your AWS setup ⚠️ ----

os.environ["AWS_DEFAULT_REGION"] = "us-east-1"  # E.g. "us-east-1"
# os.environ["AWS_PROFILE"] = "<your_profile>"
# os.environ["BEDROCK_ASSUME_ROLE"] = "arn:aws:iam::785208610362:user/s-fujimoto"  # E.g. "arn:aws:iam::xxxxxxxxxxxx:role/xxxxxx"  # E.g.

boto3_bedrock = bedrock.get_bedrock_client(
    assumed_role=os.environ.get("BEDROCK_ASSUME_ROLE", None),
    region=os.environ.get("AWS_DEFAULT_REGION", None),
)

print(
    "1 if the image is generated using a prompt, 2 if the image is generated using an image and a prompt."
)
choice = int(input("Enter Choice :"))
print(choice)

if choice == 1:
    print("please enter the Prompt"),
    prompt = str(input())
    print(
        "Please enter 3 negative Prompt eg:poorly rendered,poor background details,poorly drawn face"
    )
    negative_prompts = []
    n = 3
    for i in range(0, n):
        # ele = [input(), str(input()),]
        ele = str(input())
        negative_prompts.append(ele)
    print(negative_prompts)
    style_preset = "photographic"
    request = json.dumps(
        {
            "text_prompts": (
                [{"text": prompt, "weight": 1.0}]
                + [
                    {"text": negprompt, "weight": -1.0}
                    for negprompt in negative_prompts
                ]
            ),
            "cfg_scale": 5,
            "seed": 5450,
            "steps": 70,
            "style_preset": style_preset,
        }
    )
    modelId = "stability.stable-diffusion-xl"

    response = boto3_bedrock.invoke_model(body=request, modelId=modelId)
    response_body = json.loads(response.get("body").read())

    print(response_body["result"])
    base_64_img_str = response_body["artifacts"][0].get("base64")
    print(f"{base_64_img_str[0:80]}...")

    os.makedirs("data", exist_ok=True)
    image_1 = Image.open(
        io.BytesIO(base64.decodebytes(bytes(base_64_img_str, "utf-8")))
    )
    image_1.save("data/image_1.png")
    image_1

if choice == 2:
    print("please enter the image file path"),
    image_path = str(input())
    print("please enter the Prompt"),
    change_prompt = str(input())
    print(
        "Please enter 3 negative Prompt eg:poorly rendered,poor background details,poorly drawn face"
    )
    negative_prompts = []
    n = 3
    for i in range(0, n):
        # ele = [input(), str(input()),]
        ele = str(input())
        negative_prompts.append(ele)
    print(negative_prompts)
    style_preset = "photographic"

    def image_to_base64(img) -&gt; str:
        """Convert a PIL Image or local image file path to a base64 string for Amazon Bedrock"""
        if isinstance(img, str):
            if os.path.isfile(img):
                print(f"Reading image from file: {img}")
                with open(img, "rb") as f:
                    return base64.b64encode(f.read()).decode("utf-8")
            else:
                raise FileNotFoundError(f"File {img} does not exist")
        elif isinstance(img, Image.Image):
            print("Converting PIL Image to base64 string")
            buffer = io.BytesIO()
            img.save(buffer, format="PNG")
            return base64.b64encode(buffer.getvalue()).decode("utf-8")
        else:
            raise ValueError(f"Expected str (filename) or PIL Image. Got {type(img)}")

    original_image = Image.open(image_path)
    image_1 = original_image.resize((512, 512))
    init_image_b64 = image_to_base64(image_1)
    print(init_image_b64[:80] + "...")

    request = json.dumps(
        {
            "text_prompts": (
                [{"text": change_prompt, "weight": 1.0}]
                + [
                    {"text": negprompt, "weight": -1.0}
                    for negprompt in negative_prompts
                ]
            ),
            "cfg_scale": 10,
            "init_image": init_image_b64,
            "seed": 321,
            "start_schedule": 0.6,
            "steps": 50,
            "style_preset": style_preset,
        }
    )
    modelId = "stability.stable-diffusion-xl"

    response = boto3_bedrock.invoke_model(body=request, modelId=modelId)
    response_body = json.loads(response.get("body").read())

    print(response_body["result"])
    image_2_b64_str = response_body["artifacts"][0].get("base64")
    print(f"{image_2_b64_str[0:80]}...")

    image_2 = Image.open(
        io.BytesIO(base64.decodebytes(bytes(image_2_b64_str, "utf-8")))
    )
    image_2.save("data/image.png")

				
					
				1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147

						# Python Built-Ins:
import base64
import io
import json
import os
import sys
 
# External Dependencies:
import boto3
from PIL import Image
 
module_path = ".."
sys.path.append(os.path.abspath(module_path))
from utils import bedrock, print_ww
 
 
# ---- ⚠️ Un-comment and edit the below lines as needed for your AWS setup ⚠️ ----
 
os.environ["AWS_DEFAULT_REGION"] = "us-east-1"  # E.g. "us-east-1"
# os.environ["AWS_PROFILE"] = "<your_profile>"
# os.environ["BEDROCK_ASSUME_ROLE"] = "arn:aws:iam::785208610362:user/s-fujimoto"  # E.g. "arn:aws:iam::xxxxxxxxxxxx:role/xxxxxx"  # E.g.
 
boto3_bedrock = bedrock.get_bedrock_client(
    assumed_role=os.environ.get("BEDROCK_ASSUME_ROLE", None),
    region=os.environ.get("AWS_DEFAULT_REGION", None),
)
 
print(
    "1 if the image is generated using a prompt, 2 if the image is generated using an image and a prompt."
)
choice = int(input("Enter Choice :"))
print(choice)
 
if choice == 1:
    print("please enter the Prompt"),
    prompt = str(input())
    print(
        "Please enter 3 negative Prompt eg:poorly rendered,poor background details,poorly drawn face"
    )
    negative_prompts = []
    n = 3
    for i in range(0, n):
        # ele = [input(), str(input()),]
        ele = str(input())
        negative_prompts.append(ele)
    print(negative_prompts)
    style_preset = "photographic"
    request = json.dumps(
        {
            "text_prompts": (
                [{"text": prompt, "weight": 1.0}]
                + [
                    {"text": negprompt, "weight": -1.0}
                    for negprompt in negative_prompts
                ]
            ),
            "cfg_scale": 5,
            "seed": 5450,
            "steps": 70,
            "style_preset": style_preset,
        }
    )
    modelId = "stability.stable-diffusion-xl"
 
    response = boto3_bedrock.invoke_model(body=request, modelId=modelId)
    response_body = json.loads(response.get("body").read())
 
    print(response_body["result"])
    base_64_img_str = response_body["artifacts"][0].get("base64")
    print(f"{base_64_img_str[0:80]}...")
 
    os.makedirs("data", exist_ok=True)
    image_1 = Image.open(
        io.BytesIO(base64.decodebytes(bytes(base_64_img_str, "utf-8")))
    )
    image_1.save("data/image_1.png")
    image_1
 
if choice == 2:
    print("please enter the image file path"),
    image_path = str(input())
    print("please enter the Prompt"),
    change_prompt = str(input())
    print(
        "Please enter 3 negative Prompt eg:poorly rendered,poor background details,poorly drawn face"
    )
    negative_prompts = []
    n = 3
    for i in range(0, n):
        # ele = [input(), str(input()),]
        ele = str(input())
        negative_prompts.append(ele)
    print(negative_prompts)
    style_preset = "photographic"
 
    def image_to_base64(img) -&gt; str:
        """Convert a PIL Image or local image file path to a base64 string for Amazon Bedrock"""
        if isinstance(img, str):
            if os.path.isfile(img):
                print(f"Reading image from file: {img}")
                with open(img, "rb") as f:
                    return base64.b64encode(f.read()).decode("utf-8")
            else:
                raise FileNotFoundError(f"File {img} does not exist")
        elif isinstance(img, Image.Image):
            print("Converting PIL Image to base64 string")
            buffer = io.BytesIO()
            img.save(buffer, format="PNG")
            return base64.b64encode(buffer.getvalue()).decode("utf-8")
        else:
            raise ValueError(f"Expected str (filename) or PIL Image. Got {type(img)}")
 
    original_image = Image.open(image_path)
    image_1 = original_image.resize((512, 512))
    init_image_b64 = image_to_base64(image_1)
    print(init_image_b64[:80] + "...")
 
    request = json.dumps(
        {
            "text_prompts": (
                [{"text": change_prompt, "weight": 1.0}]
                + [
                    {"text": negprompt, "weight": -1.0}
                    for negprompt in negative_prompts
                ]
            ),
            "cfg_scale": 10,
            "init_image": init_image_b64,
            "seed": 321,
            "start_schedule": 0.6,
            "steps": 50,
            "style_preset": style_preset,
        }
    )
    modelId = "stability.stable-diffusion-xl"
 
    response = boto3_bedrock.invoke_model(body=request, modelId=modelId)
    response_body = json.loads(response.get("body").read())
 
    print(response_body["result"])
    image_2_b64_str = response_body["artifacts"][0].get("base64")
    print(f"{image_2_b64_str[0:80]}...")
 
    image_2 = Image.open(
        io.BytesIO(base64.decodebytes(bytes(image_2_b64_str, "utf-8")))
    )
    image_2.save("data/image.png")

					

			

実行

オプション1は、プロンプトから画像生成します

(py311) C:\Users\admin\Documents\PythonProjects\Bedrocks&gt;python app.py
Create new client
  Using region: us-east-1
boto3 Bedrock client successfully created!
bedrock-runtime(https://bedrock-runtime.us-east-1.amazonaws.com)
1 if the image is generated using a prompt, 2 if the image is generated using an image and a prompt.
Enter Choice :1
1
please enter the Prompt
japanese mountain
Please enter 3 negative Prompt eg:poorly rendered,poor background details,poorly drawn face
poorly rendered
poor background details
poorly drawn face
['poorly rendered', 'poor background details', 'poorly drawn face']
success
iVBORw0KGgoAAAANSUhEUgAAAgAAAAIACAIAAAB7GkOtAAADWmVYSWZNTQAqAAAACAAGAQAABAAAAAEA...

				
					
				1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

						(py311) C:\Users\admin\Documents\PythonProjects\Bedrocks&gt;python app.py
Create new client
  Using region: us-east-1
boto3 Bedrock client successfully created!
bedrock-runtime(https://bedrock-runtime.us-east-1.amazonaws.com)
1 if the image is generated using a prompt, 2 if the image is generated using an image and a prompt.
Enter Choice :1
1
please enter the Prompt
japanese mountain
Please enter 3 negative Prompt eg:poorly rendered,poor background details,poorly drawn face
poorly rendered
poor background details
poorly drawn face
['poorly rendered', 'poor background details', 'poorly drawn face']
success
iVBORw0KGgoAAAANSUhEUgAAAgAAAAIACAIAAAB7GkOtAAADWmVYSWZNTQAqAAAACAAGAQAABAAAAAEA...

					

			

japanese mountain

オプションの2は、画像とプロンプトを入力して画像生成します

(py311) C:\Users\admin\Documents\PythonProjects\Bedrocks&gt;python app.py
Create new client
  Using region: us-east-1
boto3 Bedrock client successfully created!
bedrock-runtime(https://bedrock-runtime.us-east-1.amazonaws.com)
1 if the image is generated using a prompt, 2 if the image is generated using an image and a prompt.
Enter Choice :2
2
please enter the image file path
image.jpg
please enter the Prompt
cute girl
Please enter 3 negative Prompt eg:poorly rendered,poor background details,poorly drawn face
poorly rendered
poor background details
poorly drawn face
['poorly rendered', 'poor background details', 'poorly drawn face']
Converting PIL Image to base64 string
iVBORw0KGgoAAAANSUhEUgAAAgAAAAIACAIAAAB7GkOtAAABdmlDQ1BJQ0MgUHJvZmlsZQAAeJylkLFL...
success
iVBORw0KGgoAAAANSUhEUgAAAgAAAAIACAIAAAB7GkOtAACj0GVYSWZNTQAqAAAACAAGAQAABAAAAAEA...

(py311) C:\Users\admin\Documents\PythonProjects\Bedrocks>python app.py

Create new client

Using region: us-east-1

boto3 Bedrock client successfully created!

bedrock-runtime(https://bedrock-runtime.us-east-1.amazonaws.com)

1 if the image is generated using a prompt, 2 if the image is generated using an image and a prompt.

Enter Choice :2

please enter the image file path

image.jpg

please enter the Prompt

cute girl

Please enter 3 negative Prompt eg:poorly rendered,poor background details,poorly drawn face

poorly rendered

poor background details

poorly drawn face

['poorly rendered', 'poor background details', 'poorly drawn face']

Converting PIL Image to base64 string

iVBORw0KGgoAAAANSUhEUgAAAgAAAAIACAIAAAB7GkOtAAABdmlDQ1BJQ0MgUHJvZmlsZQAAeJylkLFL...

success

iVBORw0KGgoAAAANSUhEUgAAAgAAAAIACAIAAAB7GkOtAACj0GVYSWZNTQAqAAAACAAGAQAABAAAAAEA...

だいぶ変えてきたなｗ

この記事を書いた人

aws-recipe-user

記事一覧

AWS BedrockのStable Diffusion XLで画像生成やーる

はじめに

導入

実行

この記事を書いた人

aws-recipe-user

AWS CDK Python boto3でEC2の構築をしてみた

Amazon Bedrockをすぐに触れるPlaygroundを試してみる：Te…

Amazon SageMakerサービスまとめ　令和4年12月版

「AWS AI Week For Developers」を開催！