prueba diffuser
Browse files- README.md +1 -1
- appDiffuser.py +16 -0
- requirements.txt +4 -1
README.md
CHANGED
|
@@ -5,7 +5,7 @@ colorFrom: red
|
|
| 5 |
colorTo: green
|
| 6 |
sdk: gradio
|
| 7 |
sdk_version: 6.0.2
|
| 8 |
-
app_file:
|
| 9 |
pinned: false
|
| 10 |
license: apache-2.0
|
| 11 |
---
|
|
|
|
| 5 |
colorTo: green
|
| 6 |
sdk: gradio
|
| 7 |
sdk_version: 6.0.2
|
| 8 |
+
app_file: appDiffuser.py
|
| 9 |
pinned: false
|
| 10 |
license: apache-2.0
|
| 11 |
---
|
appDiffuser.py
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import torch
|
| 3 |
+
from PIL import Image
|
| 4 |
+
from diffusers import DiffusionPipeline
|
| 5 |
+
from transformers import pipeline
|
| 6 |
+
|
| 7 |
+
modeloObtenerTextoImagen = pipeline("image-to-text", model="Salesforce/blip-image-captioning-base")
|
| 8 |
+
modeloGenerarImagen = DiffusionPipeline.from_pretrained("sd-legacy/stable-diffusion-v1-5", torch_dtype=torch.float32)
|
| 9 |
+
|
| 10 |
+
def obtenerDescripcion(imagen):
|
| 11 |
+
resultadoModeloTI = modeloObtenerTextoImagen(Image.fromarray(imagen))
|
| 12 |
+
print(f'La frase que se ha obtenido de la imagen es {resultadoModeloTI}')
|
| 13 |
+
return modeloGenerarImagen(resultadoModeloTI[0]['generated_text']).images[0]
|
| 14 |
+
|
| 15 |
+
demo = gr.Interface(fn=obtenerDescripcion, inputs="image", outputs="image")
|
| 16 |
+
demo.launch(share=True)
|
requirements.txt
CHANGED
|
@@ -1,3 +1,6 @@
|
|
| 1 |
gradio
|
| 2 |
transformers==4.49.0
|
| 3 |
-
torch==2.6.0
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
gradio
|
| 2 |
transformers==4.49.0
|
| 3 |
+
torch==2.6.0
|
| 4 |
+
diffusers==0.32.2
|
| 5 |
+
accelerate==1.5.2
|
| 6 |
+
pydantic==2.10.6
|