Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.
Python asraudiochineseemotion-recognitionfunasrmcp-servermultilingual-asropenai-compatible-apiparaformerpunctuationpytorchreal-time-asrspeaker-diarizationspeech-recognitionspeech-to-textstreaming-asrtranscriptionvllmvoice-activity-detectionwhisper-alternative
update rwkv template
维石8b033794343e
diff --git a/funasr/models/rwkv_bat/template.yaml b/funasr/models/rwkv_bat/template.yamlnew file mode 100644index 000000000..dad84bf38--- /dev/null+++ b/funasr/models/rwkv_bat/template.yaml@@ -0,0 +1,64 @@+# network architecture+model: Transducer+model_conf:+ auxiliary_ctc_weight: 0.0++# encoder+encoder: RWKVEncoder+encoder_conf:+ kernel: 3+ subsampling_factor: 4+ output_size: 512+ num_blocks: 18+ time_reduction_factor: 2+ att_dropout_rate: 0.1+ ffn_dropout_rate: 0.1+ dropout_rate: 0.1++# decoder (prediction network)+decoder: rnnt_decoder+decoder_conf:+ embed_size: 512+ hidden_size: 512+ embed_dropout_rate: 0.1+ dropout_rate: 0.1+ use_embed_mask: false++# joint network+joint_network: joint_network+joint_network_conf:+ joint_space_size: 512++frontend: WavFrontend+frontend_conf:+ fs: 16000+ window: hamming+ n_mels: 80+ frame_length: 25+ frame_shift: 10+ lfr_m: 1+ lfr_n: 1+ upsacle_samples: true++specaug: SpecAugLFR+specaug_conf:+ apply_time_warp: false+ time_warp_window: 5+ time_warp_mode: bicubic+ apply_freq_mask: true+ freq_mask_width_range:+ - 0+ - 30+ lfr_rate: 6+ num_freq_mask: 1+ apply_time_mask: true+ time_mask_width_range:+ - 0+ - 12+ num_time_mask: 1++tokenizer: CharTokenizer+tokenizer_conf:+ unk_symbol: <unk>+ split_with_space: true+\ No newline at end of file