Spaces:

Siddhant
/

ESPnet2-SLU

Runtime error

App Files Files Community

ssiidd commited on Nov 13, 2021

Commit

65191fa

1 Parent(s): a79e3a6

Add slurp model files

Browse files

Files changed (8) hide show

.DS_Store +0 -0
README.md +3 -3
app.py +118 -0
audio--1504190171-headset.flac +0 -0
packages.txt +2 -0
requirements.txt +6 -0
slurp/config.yaml +754 -0
slurp/valid.acc.ave_10best.pth +3 -0

.DS_Store ADDED Viewed

Binary file (10.2 kB). View file

README.md CHANGED Viewed

@@ -1,7 +1,7 @@
 ---
-title: ESPnet2 SLU
-emoji: 👀
-colorFrom: pink
 colorTo: green
 sdk: gradio
 app_file: app.py

 ---
+title: ESPnet2 TTS
+emoji: 📈
+colorFrom: green
 colorTo: green
 sdk: gradio
 app_file: app.py

app.py ADDED Viewed

	@@ -0,0 +1,118 @@

+import gradio as gr
+import time
+import torch
+import scipy.io.wavfile
+from espnet2.bin.tts_inference import Text2Speech
+from espnet2.utils.types import str_or_none
+from espnet2.bin.asr_inference import Speech2Text
+# tagen = 'kan-bayashi/ljspeech_vits'
+# vocoder_tagen = "none"
+speech2text = Speech2Text.from_pretrained(
+    asr_train_config="slurp/config.yaml",
+    asr_model_file="slurp/valid.acc.ave_10best.pth",
+    # Decoding parameters are not included in the model file
+    nbest=1
+)
+# Confirm the sampling rate is equal to that of the training corpus.
+# If not, you need to resample the audio data before inputting to speech2text
+speech, rate = soundfile.read("audio--1504190171-headset.flac")
+nbests = speech2text(speech)
+text, *_ = nbests[0]
+print(text)
+exit()
+text2speechen = Text2Speech.from_pretrained(
+    model_tag=str_or_none(tagen),
+    vocoder_tag=str_or_none(vocoder_tagen),
+    device="cpu",
+    # Only for Tacotron 2 & Transformer
+    threshold=0.5,
+    # Only for Tacotron 2
+    minlenratio=0.0,
+    maxlenratio=10.0,
+    use_att_constraint=False,
+    backward_window=1,
+    forward_window=3,
+    # Only for FastSpeech & FastSpeech2 & VITS
+    speed_control_alpha=1.0,
+    # Only for VITS
+    noise_scale=0.333,
+    noise_scale_dur=0.333,
+)
+tagjp = 'kan-bayashi/jsut_full_band_vits_prosody'
+vocoder_tagjp = 'none'
+text2speechjp = Text2Speech.from_pretrained(
+    model_tag=str_or_none(tagjp),
+    vocoder_tag=str_or_none(vocoder_tagjp),
+    device="cpu",
+    # Only for Tacotron 2 & Transformer
+    threshold=0.5,
+    # Only for Tacotron 2
+    minlenratio=0.0,
+    maxlenratio=10.0,
+    use_att_constraint=False,
+    backward_window=1,
+    forward_window=3,
+    # Only for FastSpeech & FastSpeech2 & VITS
+    speed_control_alpha=1.0,
+    # Only for VITS
+    noise_scale=0.333,
+    noise_scale_dur=0.333,
+)
+tagch = 'kan-bayashi/csmsc_full_band_vits'
+vocoder_tagch = "none"
+text2speechch = Text2Speech.from_pretrained(
+    model_tag=str_or_none(tagch),
+    vocoder_tag=str_or_none(vocoder_tagch),
+    device="cpu",
+    # Only for Tacotron 2 & Transformer
+    threshold=0.5,
+    # Only for Tacotron 2
+    minlenratio=0.0,
+    maxlenratio=10.0,
+    use_att_constraint=False,
+    backward_window=1,
+    forward_window=3,
+    # Only for FastSpeech & FastSpeech2 & VITS
+    speed_control_alpha=1.0,
+    # Only for VITS
+    noise_scale=0.333,
+    noise_scale_dur=0.333,
+)
+def inference(text,lang):
+  with torch.no_grad():
+      if lang == "english":
+          wav = text2speechen(text)["wav"]
+          scipy.io.wavfile.write("out.wav",text2speechen.fs , wav.view(-1).cpu().numpy())
+      if lang == "chinese":
+          wav = text2speechch(text)["wav"]
+          scipy.io.wavfile.write("out.wav",text2speechch.fs , wav.view(-1).cpu().numpy())
+      if lang == "japanese":
+          wav = text2speechjp(text)["wav"]
+          scipy.io.wavfile.write("out.wav",text2speechjp.fs , wav.view(-1).cpu().numpy())
+  return  "out.wav"
+title = "ESPnet2-TTS"
+description = "Gradio demo for ESPnet2-TTS: Extending the Edge of TTS Research. To use it, simply add your audio, or click one of the examples to load them. Read more at the links below."
+article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2110.07840' target='_blank'>ESPnet2-TTS: Extending the Edge of TTS Research</a> | <a href='https://github.com/espnet/espnet' target='_blank'>Github Repo</a></p>"
+examples=[['This paper describes ESPnet2-TTS, an end-to-end text-to-speech (E2E-TTS) toolkit. ESPnet2-TTS extends our earlier version, ESPnet-TTS, by adding many new features, including: on-the-fly flexible pre-processing, joint training with neural vocoders, and state-of-the-art TTS models with extensions like full-band E2E text-to-waveform modeling, which simplify the training pipeline and further enhance TTS performance. The unified design of our recipes enables users to quickly reproduce state-of-the-art E2E-TTS results',"english"],['レシピの統一された設計により、ユーザーは最先端のE2E-TTSの結果をすばやく再現できます。また、推論用の統合Pythonインターフェースで事前にトレーニングされたモデルを多数提供し、ユーザーがベースラインサンプルを生成してデモを構築するための迅速な手段を提供します。',"japanese"],['对英语和日语语料库的实验评估表明，我们提供的模型合成了与真实情况相当的话语，达到了最先进的水平',"chinese"]]
+gr.Interface(
+    inference,
+    [gr.inputs.Textbox(label="input text",lines=10),gr.inputs.Radio(choices=["english", "chinese", "japanese"], type="value", default="english", label="language")],
+    gr.outputs.Audio(type="file", label="Output"),
+    title=title,
+    description=description,
+    article=article,
+    enable_queue=True,
+    examples=examples
+    ).launch(debug=True)

audio--1504190171-headset.flac ADDED Viewed

Binary file (40.9 kB). View file

packages.txt ADDED Viewed

	@@ -0,0 +1,2 @@


1	+ cmake
2	+ libsndfile1

requirements.txt ADDED Viewed

	@@ -0,0 +1,6 @@

+espnet==0.10.3
+pyopenjtalk==0.1.5
+parallel_wavegan==0.5.3
+espnet_model_zoo
+scipy
+torch

slurp/config.yaml ADDED Viewed

	@@ -0,0 +1,754 @@

+config: conf/tuning/train_asr_conformer.yaml
+print_config: false
+log_level: INFO
+dry_run: false
+iterator_type: sequence
+output_dir: exp/asr_train_asr_conformer_raw_en_word
+ngpu: 1
+seed: 0
+num_workers: 1
+num_att_plot: 3
+dist_backend: nccl
+dist_init_method: env://
+dist_world_size: null
+dist_rank: null
+local_rank: 0
+dist_master_addr: null
+dist_master_port: null
+dist_launcher: null
+multiprocessing_distributed: false
+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
+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: false
+log_interval: null
+use_tensorboard: true
+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: 1000000
+valid_batch_bins: null
+train_shape_file:
+- exp/asr_stats_raw_en_word/train/speech_shape
+- exp/asr_stats_raw_en_word/train/text_shape.word
+valid_shape_file:
+- exp/asr_stats_raw_en_word/valid/speech_shape
+- exp/asr_stats_raw_en_word/valid/text_shape.word
+batch_type: folded
+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
+train_data_path_and_name_and_type:
+-   - dump/raw/train/wav.scp
+    - speech
+    - sound
+-   - dump/raw/train/text
+    - text
+    - text
+valid_data_path_and_name_and_type:
+-   - dump/raw/devel/wav.scp
+    - speech
+    - sound
+-   - dump/raw/devel/text
+    - text
+    - text
+allow_variable_data_keys: false
+max_cache_size: 0.0
+max_cache_fd: 32
+valid_max_cache_size: null
+optim: adam
+optim_conf:
+    lr: 0.0002
+scheduler: warmuplr
+scheduler_conf:
+    warmup_steps: 25000
+token_list:
+- <blank>
+- <unk>
+- ▁the
+- s
+- ▁to
+- ▁i
+- ▁me
+- ▁you
+- ▁what
+- ▁a
+- ▁is
+- ▁my
+- ▁please
+- a
+- ''''
+- y
+- ▁in
+- ing
+- ▁s
+- e
+- ▁for
+- i
+- ▁on
+- d
+- t
+- o
+- u
+- er
+- p
+- ▁of
+- es
+- re
+- l
+- ▁it
+- ▁p
+- le
+- ▁f
+- ▁m
+- ▁email
+- ▁d
+- m
+- ▁c
+- st
+- r
+- n
+- ar
+- ▁h
+- b
+- ▁that
+- c
+- ▁this
+- h
+- an
+- email_query
+- ▁play
+- ▁re
+- ▁b
+- ▁do
+- ▁can
+- at
+- ▁have
+- g
+- ▁from
+- ▁and
+- en
+- email_sendemail
+- ▁olly
+- 'on'
+- ▁new
+- it
+- qa_factoid
+- calendar_set
+- ▁any
+- or
+- ▁g
+- ▁how
+- ▁t
+- ▁tell
+- ch
+- ▁not
+- ▁about
+- ▁at
+- ate
+- general_negate
+- f
+- ▁today
+- ▁e
+- ed
+- ▁list
+- ▁r
+- in
+- k
+- ic
+- social_post
+- ▁are
+- play_music
+- general_quirky
+- ▁l
+- al
+- v
+- ent
+- ▁n
+- ▁be
+- ▁an
+- ▁st
+- et
+- ▁am
+- general_praise
+- ▁time
+- weather_query
+- ▁up
+- ▁check
+- calendar_query
+- ▁w
+- om
+- ur
+- ▁send
+- ▁with
+- ly
+- w
+- general_explain
+- ad
+- ▁th
+- news_query
+- ▁one
+- ▁emails
+- day
+- ▁sh
+- ce
+- ▁last
+- ve
+- ▁he
+- z
+- ▁ch
+- ▁will
+- ▁set
+- ▁would
+- ▁was
+- x
+- general_repeat
+- ▁add
+- ou
+- ▁again
+- ▁ex
+- is
+- ct
+- general_affirm
+- general_confirm
+- ▁song
+- ▁next
+- ▁j
+- ▁meeting
+- um
+- ation
+- ▁turn
+- ▁did
+- if
+- ▁alarm
+- am
+- ▁like
+- datetime_query
+- ter
+- ▁remind
+- ▁o
+- qa_definition
+- ▁said
+- ▁calendar
+- ll
+- se
+- ers
+- th
+- ▁get
+- our
+- ▁need
+- ▁all
+- ot
+- ▁want
+- ▁off
+- and
+- ▁right
+- ▁de
+- ▁tr
+- ut
+- general_dontcare
+- ▁
+- ▁week
+- as
+- ▁tweet
+- ight
+- ir
+- ▁your
+- ▁event
+- ▁news
+- ▁se
+- ay
+- ion
+- ▁com
+- ▁there
+- ▁ye
+- ▁weather
+- un
+- ▁confirm
+- ld
+- calendar_remove
+- ▁y
+- ▁lights
+- ▁more
+- ▁v
+- play_radio
+- ▁does
+- ▁po
+- ▁now
+- id
+- email_querycontact
+- ▁show
+- ▁could
+- ery
+- op
+- ▁day
+- ▁pm
+- ▁music
+- ▁tomorrow
+- ▁train
+- ▁u
+- ine
+- ▁or
+- ange
+- qa_currency
+- ice
+- ▁contact
+- ▁just
+- ▁jo
+- ▁think
+- qa_stock
+- end
+- ss
+- ber
+- ▁tw
+- ▁command
+- ▁make
+- ▁no
+- ▁mo
+- pe
+- ▁find
+- general_commandstop
+- ▁when
+- social_query
+- ▁so
+- ong
+- ▁co
+- ant
+- ow
+- ▁much
+- ▁where
+- ul
+- ue
+- ri
+- ap
+- ▁start
+- ▁mar
+- ▁by
+- one
+- ▁know
+- ▁wor
+- oo
+- ▁give
+- ▁let
+- ▁events
+- der
+- ▁ro
+- ▁pr
+- ▁pl
+- play_podcasts
+- art
+- us
+- ▁work
+- ▁current
+- ol
+- cooking_recipe
+- nt
+- ▁correct
+- transport_query
+- ia
+- ▁stock
+- ▁br
+- ive
+- ▁app
+- ▁two
+- ▁latest
+- lists_query
+- ▁some
+- recommendation_events
+- ab
+- ▁go
+- ▁but
+- ook
+- ke
+- alarm_set
+- play_audiobook
+- ▁k
+- ▁response
+- ▁wr
+- cast
+- ▁open
+- ▁cle
+- ▁done
+- ▁got
+- ▁ca
+- ite
+- ase
+- ▁thank
+- iv
+- ah
+- ag
+- ▁answer
+- ie
+- ▁five
+- ▁book
+- ist
+- ▁rec
+- ore
+- ▁john
+- ment
+- ▁appreci
+- ▁fri
+- ack
+- ▁remove
+- ated
+- ock
+- ree
+- j
+- ▁good
+- ▁many
+- orn
+- fe
+- ▁radio
+- ▁we
+- int
+- ▁facebook
+- ▁cl
+- ▁sev
+- ▁schedule
+- ard
+- ▁per
+- ▁li
+- ▁going
+- nd
+- ain
+- recommendation_locations
+- ▁post
+- lists_createoradd
+- ff
+- ▁su
+- red
+- iot_hue_lightoff
+- lists_remove
+- ▁ar
+- een
+- ▁say
+- ro
+- ▁volume
+- ▁le
+- ▁reply
+- ▁complaint
+- ▁out
+- ▁delete
+- ▁ne
+- ame
+- ▁detail
+- ▁if
+- im
+- ▁happ
+- orr
+- ich
+- em
+- ▁ev
+- ction
+- ▁dollar
+- ▁as
+- alarm_query
+- audio_volume_mute
+- ac
+- music_query
+- ▁mon
+- ther
+- ▁thanks
+- cel
+- ▁who
+- ave
+- ▁service
+- ▁mail
+- ty
+- ▁hear
+- de
+- ▁si
+- ▁wh
+- ood
+- ell
+- ▁con
+- ▁once
+- ound
+- ▁don
+- ▁loc
+- ▁light
+- ▁birthday
+- ▁inf
+- ort
+- ffe
+- ▁playlist
+- el
+- ening
+- ▁us
+- ▁un
+- ▁has
+- own
+- ▁inc
+- ai
+- ▁speak
+- age
+- ▁mess
+- ast
+- ci
+- ver
+- ▁ten
+- ▁underst
+- ▁pro
+- ▁q
+- enty
+- ▁ticket
+- gh
+- audio_volume_up
+- ▁take
+- ▁bo
+- ally
+- ome
+- transport_ticket
+- ind
+- iot_hue_lightchange
+- pp
+- iot_coffee
+- ▁res
+- plain
+- io
+- lar
+- takeaway_query
+- ge
+- takeaway_order
+- email_addcontact
+- play_game
+- ak
+- ▁fa
+- transport_traffic
+- music_likeness
+- ▁rep
+- act
+- ust
+- transport_taxi
+- iot_hue_lightdim
+- ▁mu
+- ▁ti
+- ick
+- ▁ha
+- ould
+- general_joke
+- '1'
+- qa_maths
+- ▁lo
+- iot_cleaning
+- q
+- ake
+- ill
+- her
+- iot_hue_lightup
+- pl
+- '2'
+- alarm_remove
+- orrect
+- ▁cont
+- mail
+- out
+- audio_volume_down
+- book
+- ail
+- recommendation_movies
+- ck
+- ▁man
+- ▁mus
+- ▁che
+- me
+- ume
+- ▁answ
+- datetime_convert
+- ▁late
+- iot_wemo_on
+- ▁twe
+- music_settings
+- iot_wemo_off
+- orre
+- ith
+- ▁tom
+- ▁fr
+- ere
+- ▁ad
+- xt
+- ▁ab
+- ank
+- general_greet
+- now
+- ▁meet
+- ▁curre
+- ▁respon
+- ▁ag
+- ght
+- audio_volume_other
+- ink
+- ▁spe
+- iot_hue_lighton
+- ▁rem
+- lly
+- '?'
+- urn
+- ▁op
+- ▁complain
+- ▁comm
+- let
+- music_dislikeness
+- ove
+- ▁sch
+- ather
+- ▁rad
+- edule
+- ▁under
+- icket
+- lease
+- ▁bir
+- erv
+- ▁birth
+- ▁face
+- ▁cur
+- sw
+- ▁serv
+- ek
+- aid
+- '9'
+- ▁vol
+- edu
+- '5'
+- cooking_query
+- lete
+- ▁joh
+- ▁det
+- firm
+- nder
+- '0'
+- irm
+- '8'
+- '&'
+- _
+- list
+- pon
+- qa_query
+- '7'
+- '3'
+- '-'
+- reci
+- ▁doll
+- <sos/eos>
+init: null
+input_size: null
+ctc_conf:
+    dropout_rate: 0.0
+    ctc_type: builtin
+    reduce: true
+    ignore_nan_grad: true
+model_conf:
+    ctc_weight: 0.3
+    lsm_weight: 0.1
+    length_normalized_loss: false
+    extract_feats_in_collect_stats: false
+use_preprocessor: true
+token_type: word
+bpemodel: null
+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'
+frontend: default
+frontend_conf:
+    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
+    - 30
+    num_freq_mask: 2
+    apply_time_mask: true
+    time_mask_width_range:
+    - 0
+    - 40
+    num_time_mask: 2
+normalize: utterance_mvn
+normalize_conf: {}
+preencoder: null
+preencoder_conf: {}
+encoder: conformer
+encoder_conf:
+    output_size: 512
+    attention_heads: 8
+    linear_units: 2048
+    num_blocks: 12
+    dropout_rate: 0.1
+    positional_dropout_rate: 0.1
+    attention_dropout_rate: 0.1
+    input_layer: conv2d
+    normalize_before: true
+    macaron_style: true
+    pos_enc_layer_type: rel_pos
+    selfattention_layer_type: rel_selfattn
+    activation_type: swish
+    use_cnn_module: true
+    cnn_module_kernel: 31
+postencoder: null
+postencoder_conf: {}
+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: 0.1
+    src_attention_dropout_rate: 0.1
+required:
+- output_dir
+- token_list
+version: 0.10.3a2
+distributed: false

slurp/valid.acc.ave_10best.pth ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:0b2144a9a26a71d820fd5923124e33cf8be4accf7e731e3c72526805350806d9
+size 437699867