test

app.py

@@ -32,11 +32,16 @@ def main(audio_fp: str):
     pipeline = PreTrainedPipeline(model_path=HF_HUB_URL, language_model_fp=lm_fp)
 
     # recognize speech
-    pipeline_res = pipeline(inputs=inputs)
-    # text_recognized = pipeline(inputs=inputs)['text'][0]
-    text_recognized = str(pipeline_res)
+    res = pipeline(inputs=inputs)
+    # text_recognized = res['text'][0]
 
-    return text_recognized
+    res['sampling_rate_orig'] = sampling_rate
+    res['init_audio_shape'] = audio.shape
+    res['inputs_shape'] = inputs.shape
+    res['inputs_max'] = np.max(inputs).item()
+    res['inputs_min'] = np.min(inputs).item()
+
+    return str(res)
 
 
 iface = gr.Interface(