Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -1,15 +1,15 @@
|
|
1 |
-
|
2 |
-
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
|
3 |
import torch
|
4 |
-
from clip_interrogator import Config, Interrogator
|
5 |
-
import random
|
6 |
import re
|
|
|
7 |
import requests
|
8 |
import shutil
|
|
|
|
|
9 |
from PIL import Image
|
|
|
10 |
|
11 |
-
|
12 |
-
# Definir la funciΓ³n para generar prompt desde imagen
|
13 |
config = Config()
|
14 |
config.device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
15 |
config.blip_offload = False if torch.cuda.is_available() else True
|
@@ -19,6 +19,7 @@ config.blip_num_beams = 64
|
|
19 |
config.clip_model_name = "ViT-H-14/laion2b_s32b_b79k"
|
20 |
ci = Interrogator(config)
|
21 |
|
|
|
22 |
def get_prompt_from_image(image, mode):
|
23 |
image = image.convert('RGB')
|
24 |
if mode == 'best':
|
@@ -31,29 +32,18 @@ def get_prompt_from_image(image, mode):
|
|
31 |
prompt = ci.interrogate_negative(image)
|
32 |
return prompt
|
33 |
|
34 |
-
|
35 |
-
# Definir la funciΓ³n para generar prompt desde texto
|
36 |
-
model = AutoModelForSeq2SeqLM.from_pretrained('Helsinki-NLP/opus-mt-zh-en').eval()
|
37 |
-
tokenizer = AutoTokenizer.from_pretrained('Helsinki-NLP/opus-mt-zh-en')
|
38 |
-
|
39 |
-
def translate(text):
|
40 |
-
with torch.no_grad():
|
41 |
-
encoded = tokenizer([text], return_tensors='pt')
|
42 |
-
sequences = model.generate(**encoded)
|
43 |
-
return tokenizer.batch_decode(sequences, skip_special_tokens=True)[0]
|
44 |
-
|
45 |
text_pipe = pipeline('text-generation', model='succinctly/text2image-prompt-generator')
|
46 |
|
47 |
def text_generate(input):
|
48 |
seed = random.randint(100, 1000000)
|
49 |
set_seed(seed)
|
50 |
-
text_in_english = translate(input)
|
51 |
for count in range(6):
|
52 |
-
sequences = text_pipe(
|
53 |
list = []
|
54 |
for sequence in sequences:
|
55 |
line = sequence['generated_text'].strip()
|
56 |
-
if line !=
|
57 |
list.append(line)
|
58 |
|
59 |
result = "\n".join(list)
|
@@ -64,25 +54,24 @@ def text_generate(input):
|
|
64 |
if count == 5:
|
65 |
return result
|
66 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
with open('./image.jpg', 'wb') as f:
|
73 |
-
response.raw.decode_content = True
|
74 |
-
shutil.copyfileobj(response.raw, f)
|
75 |
-
return Image.open('./image.jpg')
|
76 |
-
else:
|
77 |
-
raise ValueError("No se pudo cargar la imagen")
|
78 |
|
|
|
|
|
79 |
|
80 |
-
|
81 |
-
with gr.Interface(
|
82 |
-
[get_prompt_from_image, text_generate],
|
83 |
-
[
|
84 |
-
gr.inputs.Image(type='pil', label='Imagen'),
|
85 |
-
gr.inputs.Radio(['best', 'fast', 'classic', 'negative'], value='best', label='Modo'),
|
86 |
-
gr.inputs.Textbox(lines=6, label='Texto de entrada'),
|
87 |
-
],
|
88 |
-
[
|
|
|
1 |
+
# Importar bibliotecas
|
|
|
2 |
import torch
|
|
|
|
|
3 |
import re
|
4 |
+
import random
|
5 |
import requests
|
6 |
import shutil
|
7 |
+
from clip_interrogator import Config, Interrogator
|
8 |
+
from transformers import pipeline, set_seed, AutoTokenizer, AutoModelForSeq2SeqLM
|
9 |
from PIL import Image
|
10 |
+
import gradio as gr
|
11 |
|
12 |
+
# Configurar CLIP
|
|
|
13 |
config = Config()
|
14 |
config.device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
15 |
config.blip_offload = False if torch.cuda.is_available() else True
|
|
|
19 |
config.clip_model_name = "ViT-H-14/laion2b_s32b_b79k"
|
20 |
ci = Interrogator(config)
|
21 |
|
22 |
+
# FunciΓ³n para generar prompt desde imagen
|
23 |
def get_prompt_from_image(image, mode):
|
24 |
image = image.convert('RGB')
|
25 |
if mode == 'best':
|
|
|
32 |
prompt = ci.interrogate_negative(image)
|
33 |
return prompt
|
34 |
|
35 |
+
# FunciΓ³n para generar texto
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
36 |
text_pipe = pipeline('text-generation', model='succinctly/text2image-prompt-generator')
|
37 |
|
38 |
def text_generate(input):
|
39 |
seed = random.randint(100, 1000000)
|
40 |
set_seed(seed)
|
|
|
41 |
for count in range(6):
|
42 |
+
sequences = text_pipe(input, max_length=random.randint(60, 90), num_return_sequences=8)
|
43 |
list = []
|
44 |
for sequence in sequences:
|
45 |
line = sequence['generated_text'].strip()
|
46 |
+
if line != input and len(line) > (len(input) + 4) and line.endswith((':', '-', 'β')) is False:
|
47 |
list.append(line)
|
48 |
|
49 |
result = "\n".join(list)
|
|
|
54 |
if count == 5:
|
55 |
return result
|
56 |
|
57 |
+
# Crear interfaz gradio
|
58 |
+
with gr.Blocks() as block:
|
59 |
+
with gr.Column():
|
60 |
+
gr.HTML('<h1>MidJourney / SD2 Helper Tool</h1>')
|
61 |
+
with gr.Tab('Generate from Image'):
|
62 |
+
with gr.Row():
|
63 |
+
input_image = gr.Image(type='pil')
|
64 |
+
with gr.Column():
|
65 |
+
input_mode = gr.Radio(['best', 'fast', 'classic', 'negative'], value='best', label='Mode')
|
66 |
+
img_btn = gr.Button('Discover Image Prompt')
|
67 |
+
output_image = gr.Textbox(lines=6, label='Generated Prompt')
|
68 |
|
69 |
+
with gr.Tab('Generate from Text'):
|
70 |
+
input_text = gr.Textbox(lines=6, label='Your Idea', placeholder='Enter your content here...')
|
71 |
+
output_text = gr.Textbox(lines=6, label='Generated Prompt')
|
72 |
+
text_btn = gr.Button('Generate Prompt')
|
|
|
|
|
|
|
|
|
|
|
|
|
73 |
|
74 |
+
img_btn.click(fn=get_prompt_from_image, inputs=[input_image, input_mode], outputs=output_image)
|
75 |
+
text_btn.click(fn=text_generate, inputs=input_text, outputs=output_text)
|
76 |
|
77 |
+
block.queue(max_size=64).launch(show_api=False, enable_queue=True, debug=True, share=True, server_name='0.0.0.0')
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|