Update README.md

README.md

@@ -33,6 +33,8 @@ slu = EndToEndSLU.from_hparams("/network/tmp1/ravanelm/slu-direct-fluent-speech-
 slu.decode_file("/network/tmp1/ravanelm/slu-direct-fluent-speech-commands-librispeech-asr/example_fsc.wav")
 >>> '{"action:" "activate"| "object": "lights"| "location": "bedroom"}'
 ```
+The system is trained with recordings sampled at 16kHz (single channel).
+The code will automatically normalize your audio (i.e., resampling + mono channel selection) when calling *decode_file* if needed. Make sure your input tensor is compliant with the expected sampling rate if you use *encode_batch* and *decode_batch*.
 ### Inference on GPU
 To perform inference on the GPU, add  `run_opts={"device":"cuda"}`  when calling the `from_hparams` method.