FUNASR训练
parent: 6149a248 | 补丁 | 提交 | ignore whitespace
嘉渊
2023-05-23 8bf745a55e459f0b91d744184ba9fce66f0e6fa7 
update repo
16个文件已修改
56 ■■■■■ 已修改文件
docs/academic_recipe/asr_recipe.md 2 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell/conformer/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell/data2vec_paraformer_finetune/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell/data2vec_transformer_finetune/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell/paraformer/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell/paraformerbert/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell/transformer/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell/transformer/utils/compute_cmvn.sh 2 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell2/conformer/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell2/data2vec_pretrain/run.sh 2 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell2/paraformer/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell2/paraformerbert/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell2/transformer/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/aishell2/transformer/utils/compute_cmvn.sh 2 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/librispeech/conformer/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史
egs/librispeech_100h/conformer/run.sh 4 ●●●●● 补丁 | 查看 | 原始文档 | blame | 历史 
 docs/academic_recipe/asr_recipe.md


@@ -47,7 +47,7 @@


These two files both have two columns, while the first column is wav ids and the second column is the corresponding wav paths/label tokens.




### Stage 1: Feature and CMVN Generation


This stage computes CMVN based on `train` dataset, which is used in the following stages. Users can set `nj` to control the number of jobs for computing CMVN. The generated CMVN file is saved as `$feats_dir/data/train/cmvn/cmvn.mvn`.


This stage computes CMVN based on `train` dataset, which is used in the following stages. Users can set `nj` to control the number of jobs for computing CMVN. The generated CMVN file is saved as `$feats_dir/data/train/cmvn/am.mvn`.




### Stage 2: Dictionary Preparation


This stage processes the dictionary, which is used as a mapping between label characters and integer indices during ASR training. The processed dictionary file is saved as `$feats_dir/data/$lang_toekn_list/$token_type/tokens.txt`. An example of `tokens.txt` is as follows:




 egs/aishell/conformer/run.sh


@@ -136,7 +136,7 @@


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --resume true \


                --output_dir ${exp_dir}/exp/${model_dir} \


@@ -186,7 +186,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/aishell/data2vec_paraformer_finetune/run.sh


@@ -141,7 +141,7 @@


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text" \


                --init_param ${init_param} \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --resume true \


                --output_dir ${exp_dir}/exp/${model_dir} \


                --config $asr_config \


@@ -190,7 +190,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \




 egs/aishell/data2vec_transformer_finetune/run.sh


@@ -141,7 +141,7 @@


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text" \


                --init_param ${init_param} \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --resume true \


                --output_dir ${exp_dir}/exp/${model_dir} \


@@ -191,7 +191,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/aishell/paraformer/run.sh


@@ -136,7 +136,7 @@


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --resume true \


                --output_dir ${exp_dir}/exp/${model_dir} \


@@ -186,7 +186,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/aishell/paraformerbert/run.sh


@@ -147,7 +147,7 @@


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text,embeds.scp" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --resume true \


                --output_dir ${exp_dir}/exp/${model_dir} \


@@ -197,7 +197,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/aishell/transformer/run.sh


@@ -136,7 +136,7 @@


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --resume true \


                --output_dir ${exp_dir}/exp/${model_dir} \


@@ -186,7 +186,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/aishell/transformer/utils/compute_cmvn.sh


@@ -29,6 +29,6 @@




python utils/combine_cmvn_file.py --dim ${feats_dim} --cmvn_dir $split_dir --nj $nj --output_dir ${fbankdir}/cmvn




python utils/cmvn_converter.py --cmvn_json ${fbankdir}/cmvn/cmvn.json --am_mvn ${fbankdir}/cmvn/cmvn.mvn


python utils/cmvn_converter.py --cmvn_json ${fbankdir}/cmvn/cmvn.json --am_mvn ${fbankdir}/cmvn/am.mvn




echo "$0: Succeeded compute global cmvn"




 egs/aishell2/conformer/run.sh


@@ -138,7 +138,7 @@


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --dataset_type $dataset_type \


                --resume true \


@@ -189,7 +189,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/aishell2/data2vec_pretrain/run.sh


@@ -109,7 +109,7 @@


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --data_file_names "wav.scp" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --dataset_type $dataset_type \


                --resume true \




 egs/aishell2/paraformer/run.sh


@@ -138,7 +138,7 @@


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --dataset_type $dataset_type \


                --resume true \


@@ -189,7 +189,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/aishell2/paraformerbert/run.sh


@@ -148,7 +148,7 @@


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text,embeds.scp" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --dataset_type $dataset_type \


                --resume true \


@@ -199,7 +199,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/aishell2/transformer/run.sh


@@ -138,7 +138,7 @@


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --data_file_names "wav.scp,text" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --dataset_type $dataset_type \


                --resume true \


@@ -189,7 +189,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/aishell2/transformer/utils/compute_cmvn.sh


@@ -29,6 +29,6 @@




python utils/combine_cmvn_file.py --dim ${feats_dim} --cmvn_dir $split_dir --nj $nj --output_dir ${fbankdir}/cmvn




python utils/cmvn_converter.py --cmvn_json ${fbankdir}/cmvn/cmvn.json --am_mvn ${fbankdir}/cmvn/cmvn.mvn


python utils/cmvn_converter.py --cmvn_json ${fbankdir}/cmvn/cmvn.json --am_mvn ${fbankdir}/cmvn/am.mvn




echo "$0: Succeeded compute global cmvn"




 egs/librispeech/conformer/run.sh


@@ -150,7 +150,7 @@


                --data_dir ${feats_dir}/data \


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --resume true \


                --output_dir ${exp_dir}/exp/${model_dir} \


@@ -201,7 +201,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \




 egs/librispeech_100h/conformer/run.sh


@@ -146,7 +146,7 @@


                --data_dir ${feats_dir}/data \


                --train_set ${train_set} \


                --valid_set ${valid_set} \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --speed_perturb ${speed_perturb} \


                --resume true \


                --output_dir ${exp_dir}/exp/${model_dir} \


@@ -197,7 +197,7 @@


                --njob ${njob} \


                --gpuid_list ${gpuid_list} \


                --data_path_and_name_and_type "${_data}/${scp},speech,${type}" \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/cmvn.mvn \


                --cmvn_file ${feats_dir}/data/${train_set}/cmvn/am.mvn \


                --key_file "${_logdir}"/keys.JOB.scp \


                --asr_train_config "${asr_exp}"/config.yaml \


                --asr_model_file "${asr_exp}"/"${inference_asr_model}" \