Update app.py

app.py

@@ -15,7 +15,7 @@ import itertools
 MODEL_TRANSFORMER_BASED = "distilbert-base-uncased"
 MODEL_ONNX_FNAME = "ESG_classifier.onnx"
 MODEL_SENTIMENT_ANALYSIS = "ProsusAI/finbert"
-MODEL_SUMMARY_PEGASUS = "oMateos2020/pegasus-newsroom-cnn_full-adafactor-bs6"
+#MODEL_SUMMARY_PEGASUS = "oMateos2020/pegasus-newsroom-cnn_full-adafactor-bs6"
 
 
 
@@ -40,9 +40,9 @@ def _inference_ner_spancat(text, summary, penalty=0.5, normalise=True, limit_out
    
     return dict(itertools.islice(sorted(comp_raw_text.items(), key=lambda x: x[1], reverse=True), limit_outputs))
 
-def _inference_summary_model_pipeline(text):
-    pipe = pipeline("text2text-generation", model=MODEL_SUMMARY_PEGASUS)
-    return pipe(text,truncation='longest_first')
+#def _inference_summary_model_pipeline(text):
+#    pipe = pipeline("text2text-generation", model=MODEL_SUMMARY_PEGASUS)
+#    return pipe(text,truncation='longest_first')
 
 def _inference_sentiment_model_pipeline(text):
     tokenizer_kwargs = {'padding':True,'truncation':True,'max_length':512}#,'return_tensors':'pt'}
@@ -105,7 +105,7 @@ def _inference_classifier(text):
     # compute ONNX Runtime output prediction
     ort_outs = ort_session.run(None, input_feed=dict(inputs))
 
-    return sigmoid(ort_outs[0])[0]
+    return sigmoid(ort_outs[0])
 
 def inference(input_batch,isurl,use_archive,limit_companies=10):
     input_batch_content = []