add README

README.md

@@ -0,0 +1,826 @@
+---
+datasets:
+- LIUM/tedlium
+language:
+- en
+metrics:
+- wer
+library_name: espnet
+pipeline_tag: automatic-speech-recognition
+---
+
+## ESPnet2 ASR model 
+
+### `espnet/tedlium3`
+
+This model was trained by Dongwei Jiang using tedlium3 recipe in [espnet](https://github.com/espnet/espnet/).
+
+### Demo: How to use in ESPnet2
+
+```bash
+cd espnet
+git checkout ff841366229d539eb74d23ac999cae7c0cc62cad
+pip install -e .
+cd egs2/tedlium3/asr1
+./run.sh --skip_data_prep false --skip_train true --download_model espnet/dongwei_tedlium3_asr_e-branchformer_external_lm
+```
+
+<!-- Generated by scripts/utils/show_asr_result.sh -->
+# RESULTS
+## Environments
+- date: `Tue Apr 11 01:15:36 EDT 2023`
+- python version: `3.8.16 (default, Mar  2 2023, 03:21:46)  [GCC 11.2.0]`
+- espnet version: `espnet 202301`
+- pytorch version: `pytorch 1.8.1`
+- Git hash: `b0cceeac2ecd330e8270789cef945e49058858fa`
+  - Commit date: `Thu Mar 30 08:26:54 2023 -0400`
+
+## exp/asr_train_asr_e_branchformer_size256_mlp1024_e12_mactrue_raw_en_bpe500_sp
+### WER
+
+|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
+|---|---|---|---|---|---|---|---|---|
+|decode_lm_lm_train_lm_en_bpe500_valid.loss.ave_asr_model_valid.acc.ave/test|1155|27500|94.2|2.5|3.3|0.6|6.4|59.2|
+
+### CER
+
+|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
+|---|---|---|---|---|---|---|---|---|
+|decode_lm_lm_train_lm_en_bpe500_valid.loss.ave_asr_model_valid.acc.ave/test|1155|145066|96.8|0.5|2.7|0.6|3.8|59.2|
+
+### TER
+
+|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
+|---|---|---|---|---|---|---|---|---|
+|decode_lm_lm_train_lm_en_bpe500_valid.loss.ave_asr_model_valid.acc.ave/test|1155|54206|95.8|1.6|2.6|0.5|4.7|59.2|
+
+## exp/asr_train_asr_e_branchformer_size256_mlp1024_e12_mactrue_raw_en_bpe500_sp/decode_lm_lm_train_lm_en_bpe500_valid.loss.ave_asr_model_valid.acc.ave
+### WER
+
+|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
+|---|---|---|---|---|---|---|---|---|
+|org/dev|507|17783|93.6|3.1|3.3|0.9|7.3|69.0|
+
+### CER
+
+|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
+|---|---|---|---|---|---|---|---|---|
+|org/dev|507|95429|96.5|0.7|2.8|0.8|4.4|69.0|
+
+### TER
+
+|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
+|---|---|---|---|---|---|---|---|---|
+|org/dev|507|36002|95.4|2.0|2.6|0.8|5.5|69.0|
+
+
+## ASR config
+
+<details><summary>expand</summary>
+
+```
+config: conf/tuning/train_asr_e_branchformer_size256_mlp1024_e12_mactrue.yaml
+print_config: false
+log_level: INFO
+dry_run: false
+iterator_type: sequence
+output_dir: exp/asr_train_asr_e_branchformer_size256_mlp1024_e12_mactrue_raw_en_bpe500_sp
+ngpu: 1
+seed: 2022
+num_workers: 6
+num_att_plot: 3
+dist_backend: nccl
+dist_init_method: env://
+dist_world_size: 4
+dist_rank: 0
+local_rank: 0
+dist_master_addr: localhost
+dist_master_port: 33461
+dist_launcher: null
+multiprocessing_distributed: true
+unused_parameters: false
+sharded_ddp: false
+cudnn_enabled: true
+cudnn_benchmark: false
+cudnn_deterministic: true
+collect_stats: false
+write_collected_feats: false
+max_epoch: 50
+patience: null
+val_scheduler_criterion:
+- valid
+- loss
+early_stopping_criterion:
+- valid
+- loss
+- min
+best_model_criterion:
+-   - valid
+    - acc
+    - max
+keep_nbest_models: 10
+nbest_averaging_interval: 0
+grad_clip: 5.0
+grad_clip_type: 2.0
+grad_noise: false
+accum_grad: 1
+no_forward_run: false
+resume: true
+train_dtype: float32
+use_amp: true
+log_interval: null
+use_matplotlib: true
+use_tensorboard: true
+create_graph_in_tensorboard: false
+use_wandb: false
+wandb_project: null
+wandb_id: null
+wandb_entity: null
+wandb_name: null
+wandb_model_log_interval: -1
+detect_anomaly: false
+pretrain_path: null
+init_param: []
+ignore_init_mismatch: false
+freeze_param: []
+num_iters_per_epoch: null
+batch_size: 20
+valid_batch_size: null
+batch_bins: 50000000
+valid_batch_bins: null
+train_shape_file:
+- exp/asr_stats_raw_en_bpe500_sp/train/speech_shape
+- exp/asr_stats_raw_en_bpe500_sp/train/text_shape.bpe
+valid_shape_file:
+- exp/asr_stats_raw_en_bpe500_sp/valid/speech_shape
+- exp/asr_stats_raw_en_bpe500_sp/valid/text_shape.bpe
+batch_type: numel
+valid_batch_type: null
+fold_length:
+- 80000
+- 150
+sort_in_batch: descending
+sort_batch: descending
+multiple_iterator: false
+chunk_length: 500
+chunk_shift_ratio: 0.5
+num_cache_chunks: 1024
+chunk_excluded_key_prefixes: []
+train_data_path_and_name_and_type:
+-   - dump/raw/train_sp/wav.scp
+    - speech
+    - kaldi_ark
+-   - dump/raw/train_sp/text
+    - text
+    - text
+valid_data_path_and_name_and_type:
+-   - dump/raw/dev/wav.scp
+    - speech
+    - kaldi_ark
+-   - dump/raw/dev/text
+    - text
+    - text
+allow_variable_data_keys: false
+max_cache_size: 0.0
+max_cache_fd: 32
+valid_max_cache_size: null
+exclude_weight_decay: false
+exclude_weight_decay_conf: {}
+optim: adam
+optim_conf:
+    lr: 0.002
+    weight_decay: 1.0e-06
+scheduler: warmuplr
+scheduler_conf:
+    warmup_steps: 15000
+token_list:
+- <blank>
+- <unk>
+- '[unk]'
+- ▁
+- s
+- ▁the
+- t
+- ▁and
+- e
+- ▁a
+- ▁to
+- d
+- ▁of
+- ''''
+- n
+- ing
+- ▁in
+- ▁that
+- re
+- ▁i
+- c
+- o
+- u
+- ▁we
+- y
+- a
+- ed
+- ▁it
+- ▁you
+- i
+- m
+- ▁is
+- er
+- p
+- g
+- w
+- al
+- ▁this
+- ▁so
+- f
+- le
+- b
+- ar
+- ▁f
+- k
+- ▁c
+- r
+- in
+- or
+- ▁for
+- ▁be
+- ve
+- ▁was
+- te
+- th
+- ▁do
+- es
+- ly
+- ▁they
+- ro
+- ▁are
+- ▁with
+- ▁have
+- an
+- v
+- ch
+- ▁on
+- se
+- lo
+- ▁but
+- en
+- ri
+- li
+- ▁what
+- it
+- ic
+- ▁can
+- l
+- ur
+- ce
+- ent
+- ▁me
+- ▁b
+- ▁ma
+- ▁he
+- ra
+- ▁de
+- ll
+- at
+- ▁about
+- ▁one
+- ▁not
+- ne
+- ▁all
+- ▁my
+- ter
+- el
+- il
+- ▁there
+- 'on'
+- ad
+- ▁mo
+- ol
+- ation
+- nd
+- ▁like
+- ▁people
+- po
+- ▁at
+- ▁us
+- us
+- ▁g
+- ci
+- ▁our
+- h
+- pe
+- ▁as
+- ▁from
+- vi
+- ▁if
+- as
+- ▁ex
+- ▁con
+- ▁an
+- ver
+- ▁out
+- ▁just
+- un
+- ▁see
+- la
+- ▁di
+- ▁when
+- ▁now
+- ▁p
+- ha
+- ▁who
+- ck
+- ▁these
+- ▁because
+- ▁or
+- ▁know
+- ion
+- ir
+- ▁co
+- ▁up
+- ▁pa
+- ment
+- ▁think
+- ge
+- ▁how
+- ide
+- ▁by
+- ul
+- ity
+- ▁go
+- ▁get
+- ▁ho
+- ive
+- ▁very
+- ate
+- ng
+- ▁no
+- ▁had
+- ac
+- ▁bo
+- ry
+- ▁more
+- ▁them
+- ▁some
+- mi
+- ▁time
+- ▁your
+- me
+- ▁going
+- op
+- am
+- per
+- et
+- ▁would
+- ru
+- ure
+- ti
+- ist
+- ▁their
+- x
+- ▁were
+- ▁look
+- ▁pro
+- ▁which
+- ▁work
+- tion
+- est
+- ty
+- im
+- z
+- ta
+- ▁want
+- ▁two
+- age
+- ▁really
+- om
+- ma
+- ers
+- ting
+- ▁world
+- co
+- ▁way
+- ▁don
+- wa
+- hi
+- tra
+- ▁la
+- ▁here
+- able
+- lu
+- ▁other
+- mo
+- ies
+- ▁has
+- ▁could
+- j
+- ▁make
+- ally
+- ▁sta
+- ten
+- ▁will
+- ▁un
+- ig
+- ▁where
+- ▁into
+- ke
+- ▁than
+- ▁comp
+- ▁actually
+- tic
+- sh
+- ▁did
+- tor
+- fa
+- ical
+- ▁she
+- ▁years
+- ▁say
+- one
+- ted
+- ▁things
+- ph
+- ▁new
+- ▁pre
+- ▁any
+- ▁thousand
+- ▁been
+- ▁inter
+- ▁his
+- ▁com
+- ▁need
+- nce
+- ▁right
+- ▁take
+- ▁even
+- ▁over
+- ▁start
+- ▁hundred
+- min
+- ▁sp
+- ▁those
+- ▁car
+- ▁then
+- mp
+- ap
+- ▁first
+- les
+- ize
+- ▁every
+- ba
+- ▁something
+- ▁well
+- ard
+- ▁str
+- ▁back
+- und
+- ia
+- pl
+- ki
+- ho
+- ▁call
+- ▁most
+- ▁also
+- bi
+- ▁thing
+- ▁life
+- um
+- ▁said
+- ▁kind
+- ▁lot
+- ▁much
+- va
+- ▁ra
+- ▁little
+- ▁dr
+- ▁got
+- ▁come
+- ful
+- ▁talk
+- ▁part
+- ▁day
+- ant
+- ction
+- ▁happen
+- ▁only
+- ▁many
+- ▁wo
+- pri
+- ▁her
+- ▁br
+- qui
+- ▁mean
+- ▁three
+- iv
+- ▁different
+- ugh
+- ain
+- ▁human
+- ance
+- ▁change
+- ▁let
+- ▁real
+- ▁show
+- ▁good
+- ▁around
+- ▁through
+- ▁jo
+- bu
+- ▁down
+- ight
+- ga
+- ▁why
+- ▁live
+- ff
+- ▁tell
+- ▁put
+- ▁idea
+- port
+- ▁same
+- ▁give
+- ated
+- ish
+- ible
+- ▁though
+- ious
+- ▁problem
+- ▁five
+- par
+- ▁fact
+- ▁cha
+- ition
+- ▁year
+- ▁big
+- ▁plan
+- ▁great
+- ▁find
+- ▁four
+- ▁app
+- ▁after
+- ▁system
+- ▁place
+- ▁em
+- ▁build
+- ▁percent
+- ▁again
+- ▁point
+- ▁learn
+- ▁own
+- ▁long
+- ▁made
+- ▁today
+- ▁nine
+- ities
+- ▁gene
+- ▁six
+- ▁question
+- light
+- ▁should
+- ▁came
+- ▁feel
+- ▁turn
+- ▁person
+- ▁end
+- ▁hu
+- ▁design
+- ▁help
+- ▁brain
+- ▁last
+- ▁create
+- ▁important
+- ▁before
+- ▁high
+- ▁never
+- ▁trans
+- ▁another
+- ▁him
+- ▁eight
+- ▁might
+- ▁understand
+- ▁power
+- ▁better
+- q
+- ▁found
+- ▁play
+- ▁twenty
+- ▁still
+- ▁school
+- ▁each
+- ▁seven
+- ▁together
+- ▁few
+- ▁hand
+- ▁example
+- que
+- ▁next
+- ▁million
+- ▁story
+- ▁women
+- ▁under
+- ▁number
+- ▁course
+- ▁water
+- ▁ago
+- ▁grow
+- ▁between
+- ▁develop
+- ▁america
+- ▁sort
+- ▁technology
+- ▁believe
+- ▁second
+- ▁small
+- ▁maybe
+- ▁become
+- press
+- ▁health
+- ▁space
+- ▁word
+- ▁hard
+- ▁children
+- ▁organ
+- ▁always
+- ▁country
+- ▁reason
+- ▁experience
+- ▁large
+- ▁everything
+- ▁friend
+- ▁project
+- ▁computer
+- ▁fifty
+- ▁money
+- ▁information
+- graph
+- ▁walk
+- ization
+- ▁africa
+- ▁picture
+- ▁process
+- ▁teach
+- ▁enough
+- ▁elect
+- ▁thirty
+- '0'
+- '1'
+- '2'
+- '9'
+- '3'
+- '5'
+- '8'
+- '4'
+- '7'
+- '6'
+- '&'
+- +
+- '#'
+- '@'
+- '*'
+- \
+- ^
+- R
+- _
+- '-'
+- '%'
+- '='
+- $
+- M
+- ā
+- ']'
+- E
+- U
+- A
+- G
+- '['
+- <sos/eos>
+init: null
+input_size: null
+ctc_conf:
+    dropout_rate: 0.0
+    ctc_type: builtin
+    reduce: true
+    ignore_nan_grad: null
+    zero_infinity: true
+joint_net_conf: null
+use_preprocessor: true
+token_type: bpe
+bpemodel: data/en_token_list/bpe_unigram500/bpe.model
+non_linguistic_symbols: null
+cleaner: null
+g2p: null
+speech_volume_normalize: null
+rir_scp: null
+rir_apply_prob: 1.0
+noise_scp: null
+noise_apply_prob: 1.0
+noise_db_range: '13_15'
+short_noise_thres: 0.5
+aux_ctc_tasks: []
+frontend: default
+frontend_conf:
+    n_fft: 512
+    win_length: 400
+    hop_length: 160
+    fs: 16k
+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
+    - 27
+    num_freq_mask: 2
+    apply_time_mask: true
+    time_mask_width_ratio_range:
+    - 0.0
+    - 0.05
+    num_time_mask: 5
+normalize: global_mvn
+normalize_conf:
+    stats_file: exp/asr_stats_raw_en_bpe500_sp/train/feats_stats.npz
+model: espnet
+model_conf:
+    ctc_weight: 0.3
+    lsm_weight: 0.1
+    length_normalized_loss: false
+preencoder: null
+preencoder_conf: {}
+encoder: e_branchformer
+encoder_conf:
+    output_size: 256
+    attention_heads: 4
+    attention_layer_type: rel_selfattn
+    pos_enc_layer_type: rel_pos
+    rel_pos_type: latest
+    cgmlp_linear_units: 1024
+    cgmlp_conv_kernel: 31
+    use_linear_after_conv: false
+    gate_activation: identity
+    num_blocks: 12
+    dropout_rate: 0.1
+    positional_dropout_rate: 0.1
+    attention_dropout_rate: 0.1
+    input_layer: conv2d
+    layer_drop_rate: 0.0
+    linear_units: 1024
+    positionwise_layer_type: linear
+    use_ffn: true
+    macaron_ffn: true
+    merge_conv_kernel: 31
+postencoder: null
+postencoder_conf: {}
+decoder: transformer
+decoder_conf:
+    attention_heads: 4
+    linear_units: 2048
+    num_blocks: 6
+    dropout_rate: 0.1
+    positional_dropout_rate: 0.1
+    self_attention_dropout_rate: 0.1
+    src_attention_dropout_rate: 0.1
+preprocessor: default
+preprocessor_conf: {}
+required:
+- output_dir
+- token_list
+version: '202301'
+distributed: true
+```
+
+</details>
+
+
+
+### Citing ESPnet
+
+```BibTex
+@inproceedings{watanabe2018espnet,
+  author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
+  title={{ESPnet}: End-to-End Speech Processing Toolkit},
+  year={2018},
+  booktitle={Proceedings of Interspeech},
+  pages={2207--2211},
+  doi={10.21437/Interspeech.2018-1456},
+  url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
+}
+
+
+
+
+```
+
+or arXiv:
+
+```bibtex
+@misc{watanabe2018espnet,
+  title={ESPnet: End-to-End Speech Processing Toolkit}, 
+  author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
+  year={2018},
+  eprint={1804.00015},
+  archivePrefix={arXiv},
+  primaryClass={cs.CL}
+}
+```