Update app.py

app.py

@@ -6,7 +6,7 @@ import streamlit as st
 
 title = "Fairseq Speech to Speech Translation"
 
-description = "Gradio Demo for fairseq S2S: speech-to-speech translation models. To use it, simply record your audio, or click the example to load. Read more at the links below."
+description = "Gradio Demo for fairseq S2S: speech-to-speech translation models. To use it, simply record your audio. Read more at the links below."
 
 article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2107.05604' target='_blank'>Direct speech-to-speech translation with discrete units</a> | <a href='https://github.com/facebookresearch/fairseq/tree/main/examples/speech_to_speech' target='_blank'>Github Repo</a></p>"
 
@@ -14,25 +14,28 @@ examples = [
   ["enhanced_direct_s2st_units_audios_es-en_set2_source_12478_cv.flac","xm_transformer_s2ut_800m-es-en-st-asr-bt_h1_2022"],
 ]
 
-io1 = gr.Interface.load("huggingface/facebook/xm_transformer_s2ut_800m-es-en-st-asr-bt_h1_2022", api_key=st.secrets["api_key"])
+io1 = gr.Interface.load("huggingface/facebook/xm_transformer_s2ut_en-hk", api_key=st.secrets["api_key"])
+io2 = gr.Interface.load("huggingface/facebook/xm_transformer_s2ut_hk-en", api_key=st.secrets["api_key"])
+io3 = gr.Interface.load("huggingface/facebook/xm_transformer_unity_en-hk", api_key=st.secrets["api_key"])
+io4 = gr.Interface.load("huggingface/facebook/xm_transformer_unity_hk-en", api_key=st.secrets["api_key"])
    
 def inference(audio, model):
-#    if mic is not None and file is None:
-#        audio = mic
-#    elif file is not None and mic is None:
-#        audio = file
-#    else:
-#        return "ERROR: You must and may only select one method, it cannot be empty or select both methods at once."
-    out_audio = io1(audio)   
+    if model == "xm_transformer_s2ut_en-hk":
+        out_audio = io1(text)
+    elif model == "xm_transformer_s2ut_hk-en":
+        out_audio = io2(text)
+    elif model == "xm_transformer_unity_en-hk":
+        out_audio = io3(text)
+    else model == "xm_transformer_unity_en-hk":
+        out_audio = io4(text)
     return out_audio 
 
 
 gr.Interface(
     inference,
-    [gr.inputs.Audio(source="microphone", type="filepath", label="Input"),gr.inputs.Dropdown(choices=["xm_transformer_s2ut_800m-es-en-st-asr-bt_h1_2022"], default="xm_transformer_s2ut_800m-es-en-st-asr-bt_h1_2022",type="value", label="Model")
+    [gr.inputs.Audio(source="microphone", type="filepath", label="Input"),gr.inputs.Dropdown(choices=["xm_transformer_s2ut_en-hk", "xm_transformer_s2ut_hk-en", "xm_transformer_unity_en-hk", "xm_transformer_unity_en-hk"], default="xm_transformer_s2ut_en-hk",type="value", label="Model")
 ],
     gr.outputs.Audio(label="Output"),
     article=article,
     title=title,
-    examples=examples,
     description=description).queue().launch()