Spaces:

fffiloni
/

speech-to-image

Paused

fffiloni commited on Nov 14, 2022

Commit

4ca4534

1 Parent(s): 9b1c0d7

disabled torch dtype as GPU is not available

Files changed (1) hide show

app.py CHANGED Viewed

@@ -20,8 +20,8 @@ diffuser_pipeline = DiffusionPipeline.from_pretrained(
     speech_model=model,
     speech_processor=processor,
     use_auth_token=MY_SECRET_TOKEN,
-    revision="fp16",
-    torch_dtype=torch.float16,
 )
 diffuser_pipeline.enable_attention_slicing()

     speech_model=model,
     speech_processor=processor,
     use_auth_token=MY_SECRET_TOKEN,
+    #revision="fp16",
+    #torch_dtype=torch.float16,
 )
 diffuser_pipeline.enable_attention_slicing()