{ "audio_compression_rate": 8, "audio_downsample_rate": 1, "audio_feat_stride": 1, "dynamic_hd": 36, "feature_extractor_type": "Phi4MultimodalFeatureExtractor", "feature_size": 80, "hop_length": 160, "image_processor_type": "Phi4MMImageProcessor", "n_fft": 512, "padding_side": "right", "padding_value": 0.0, "preemphasis": 0.97, "processor_class": "Phi4MultimodalProcessor", "return_attention_mask": true, "sampling_rate": 16000, "win_length": 400 }