Upload 3 files

app.py

@@ -3,7 +3,8 @@ import torch
 import gradio as gr
 import pytube as pt
 from transformers import pipeline
-from diffusers import DiffusionPipeline
+from diffusers import StableDiffusionPipeline
+
 
 
 MODEL_NAME = "whispy/whisper_italian"
@@ -22,18 +23,10 @@ pipe = pipeline(
     device=device,
 )
 
-diffuser_pipeline = DiffusionPipeline.from_pretrained(
-    "CompVis/stable-diffusion-v1-4",
-    #custom_pipeline="speech_to_image_diffusion",
-    #speech_model=model,
-    #speech_processor=processor,
-    #use_auth_token=MY_SECRET_TOKEN,
-    #revision="fp16",
-    #torch_dtype=torch.float16,
-)
+YOUR_TOKEN="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"
+image_pipe = StableDiffusionPipeline.from_pretrained("CompVis/stable-diffusion-v1-4", revision="fp16", torch_dtype=torch.float16, use_auth_token=YOUR_TOKEN)
 
-#diffuser_pipeline.enable_attention_slicing()
-#diffuser_pipeline = diffuser_pipeline.to(device)
+image_pipe.to("cuda")
 
 translator = pipeline("translation", model="Helsinki-NLP/opus-mt-it-en")
 
@@ -55,8 +48,9 @@ def transcribe(microphone, file_upload):
     translate = translator(text)
     translate = translate[0]["translation_text"]
 
-    output = diffuser_pipeline(translate)
-    image = output.images[0]
+    image = image_pipe(translate)["sample"][0]
+    #output = diffuser_pipeline(translate)
+    #image = output.images[0]
 
     return warn_output + text, translate, image
 

requirements.txt

@@ -1,5 +1,5 @@
 transformers
 torch
 pytube
-diffusers
-sentencepiece
+sentencepiece
+diffusers