update repo

egs/wenetspeech/conformer/conf/decode_asr_transformer_5beam.yaml

@@ -0,0 +1,6 @@
+beam_size: 5
+penalty: 0.0
+maxlenratio: 0.0
+minlenratio: 0.0
+ctc_weight: 0.5
+lm_weight: 0.7

egs/wenetspeech/conformer/conf/train_asr_conformer.yaml

@@ -0,0 +1,104 @@
+# network architecture
+# encoder related
+encoder: conformer
+encoder_conf:
+    output_size: 512    # dimension of attention
+    attention_heads: 8
+    linear_units: 2048  # the number of units of position-wise feed forward
+    num_blocks: 12      # the number of encoder blocks
+    dropout_rate: 0.1
+    positional_dropout_rate: 0.1
+    attention_dropout_rate: 0.0
+    input_layer: conv2d # encoder architecture type
+    normalize_before: true
+    rel_pos_type: latest
+    pos_enc_layer_type: rel_pos
+    selfattention_layer_type: rel_selfattn
+    activation_type: swish
+    macaron_style: true
+    use_cnn_module: true
+    cnn_module_kernel: 15
+
+# decoder related
+decoder: transformer
+decoder_conf:
+    attention_heads: 8
+    linear_units: 2048
+    num_blocks: 6
+    dropout_rate: 0.1
+    positional_dropout_rate: 0.1
+    self_attention_dropout_rate: 0.0
+    src_attention_dropout_rate: 0.0
+
+# CTC realted
+ctc_conf:
+    ignore_nan_grad: true
+
+# frontend related
+frontend: wav_frontend
+frontend_conf:
+    fs: 16000
+    window: hamming
+    n_mels: 80
+    frame_length: 25
+    frame_shift: 10
+    lfr_m: 1
+    lfr_n: 1
+
+# hybrid CTC/attention
+model_conf:
+    ctc_weight: 0.3
+    lsm_weight: 0.1     # label smoothing option
+    length_normalized_loss: false
+
+# optimization related
+accum_grad: 4
+grad_clip: 5
+patience: none
+max_epoch: 30
+val_scheduler_criterion:
+    - valid
+    - acc
+best_model_criterion:
+-   - valid
+    - acc
+    - max
+keep_nbest_models: 10
+
+optim: adam
+optim_conf:
+   lr: 0.0015
+scheduler: warmuplr
+scheduler_conf:
+   warmup_steps: 30000
+
+specaug: specaug
+specaug_conf:
+    apply_time_warp: true
+    time_warp_window: 5
+    time_warp_mode: bicubic
+    apply_freq_mask: true
+    freq_mask_width_range:
+    - 0
+    - 30
+    num_freq_mask: 2
+    apply_time_mask: true
+    time_mask_width_range:
+    - 0
+    - 40
+    num_time_mask: 2
+
+dataset_conf:
+    data_names: speech,text
+    data_types: sound,text
+    shuffle: True
+    shuffle_conf:
+        shuffle_size: 2048
+        sort_size: 500
+    batch_conf:
+        batch_type: token
+        batch_size: 32000
+    num_workers: 8
+
+log_interval: 50
+normalize: None

egs/wenetspeech/conformer/path.sh

@@ -0,0 +1,5 @@
+export FUNASR_DIR=$PWD/../../..
+
+# NOTE(kan-bayashi): Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
+export PYTHONIOENCODING=UTF-8
+export PATH=$FUNASR_DIR/funasr/bin:$PATH

egs/wenetspeech/conformer/run.sh

@@ -0,0 +1,13 @@
+#!/usr/bin/env bash
+
+. ./path.sh || exit 1;
+
+# machines configuration
+CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7"
+gpu_num=8
+count=1
+gpu_inference=true  # Whether to perform gpu decoding, set false for cpu decoding
+# for gpu decoding, inference_nj=ngpu*njob; for cpu decoding, inference_nj=njob
+njob=5
+train_cmd=utils/run.pl
+infer_cmd=utils/run.pl

egs/wenetspeech/conformer/utils

@@ -0,0 +1 @@
+../../aishell/transformer/utils