EricChang
/

openai-whisper-large-v2-Lora

Automatic Speech Recognition

Model card Files Files and versions Community

cathyi commited on Jun 25, 2023

Commit

b291dd7

·

1 Parent(s): 20f594c

update handler

Files changed (1) hide show

handler.py +1 -1

handler.py CHANGED Viewed

@@ -44,4 +44,4 @@ class EndpointHandler():
         inputs = data.pop("inputs", data)
         with torch.cuda.amp.autocast():
             prediction = self.pipeline(inputs, generate_kwargs={"forced_decoder_ids": self.forced_decoder_ids}, max_new_tokens=255)["text"]
-        return {"prediction": prediction}

         inputs = data.pop("inputs", data)
         with torch.cuda.amp.autocast():
             prediction = self.pipeline(inputs, generate_kwargs={"forced_decoder_ids": self.forced_decoder_ids}, max_new_tokens=255)["text"]
+        return prediction