|
--- |
|
tags: |
|
- espnet |
|
- audio |
|
- classification |
|
language: en |
|
datasets: |
|
- as20k |
|
license: cc-by-4.0 |
|
--- |
|
|
|
## ESPnet2 CLS model |
|
|
|
### `espnet/BEATs-AS20K` |
|
|
|
This model was trained by Shikhar Bharadwaj using as20k recipe in [espnet](https://github.com/espnet/espnet/). |
|
|
|
### Demo: How to use in ESPnet2 |
|
|
|
Follow the [ESPnet installation instructions](https://espnet.github.io/espnet/installation.html) |
|
if you haven't done that already. |
|
|
|
```bash |
|
cd espnet |
|
git checkout 9634114cd3c35e230f4a9dda752e982512517653 |
|
pip install -e . |
|
cd egs2/as20k/cls1 |
|
./run.sh --skip_data_prep false --skip_train true --download_model espnet/BEATs-AS20K |
|
``` |
|
|
|
<!-- Generated by scripts/utils/show_cls_result.sh --> |
|
# RESULTS |
|
## Environments |
|
- date: `Fri Jan 3 23:25:40 EST 2025` |
|
- python version: `3.9.20 (main, Oct 3 2024, 07:27:41) [GCC 11.2.0]` |
|
- espnet version: `espnet 202412` |
|
- pytorch version: `pytorch 2.4.0` |
|
- Git hash: `635b3add116ae68c056f7aa67f64591c9ba7eb3e` |
|
- Commit date: `Thu Jan 2 11:46:32 2025 -0500` |
|
|
|
## cls_beats_iter3p20k.allroll.Bp8p8.20250103.020644 |
|
|Dataset|Metric|Value| |
|
|---|---|---| |
|
./beats_runs/as20k_fulltrain/exp/cls_beats_iter3p20k.allroll.Bp8p8.20250103.020644/cls_eval/score|mean_acc|47.73 |
|
./beats_runs/as20k_fulltrain/exp/cls_beats_iter3p20k.allroll.Bp8p8.20250103.020644/cls_eval/score|mAP|37.46 |
|
./beats_runs/as20k_fulltrain/exp/cls_beats_iter3p20k.allroll.Bp8p8.20250103.020644/cls_eval/score|mean_auc|96.58 |
|
./beats_runs/as20k_fulltrain/exp/cls_beats_iter3p20k.allroll.Bp8p8.20250103.020644/cls_eval/score|n_labels|527.00 |
|
./beats_runs/as20k_fulltrain/exp/cls_beats_iter3p20k.allroll.Bp8p8.20250103.020644/cls_eval/score|n_instances|20123.00 |
|
|
|
## CLS config |
|
|
|
<details><summary>expand</summary> |
|
|
|
``` |
|
config: conf/beats_cls.yaml |
|
print_config: false |
|
log_level: INFO |
|
drop_last_iter: false |
|
dry_run: false |
|
iterator_type: sequence |
|
valid_iterator_type: null |
|
output_dir: ./beats_runs/as20k_fulltrain/exp/cls_beats_iter3p20k.allroll.Bp8p8.20250103.020644 |
|
ngpu: 1 |
|
seed: 0 |
|
num_workers: 2 |
|
num_att_plot: 0 |
|
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: true |
|
sharded_ddp: false |
|
use_deepspeed: false |
|
deepspeed_config: null |
|
cudnn_enabled: true |
|
cudnn_benchmark: false |
|
cudnn_deterministic: true |
|
use_tf32: false |
|
collect_stats: false |
|
write_collected_feats: false |
|
max_epoch: 160 |
|
patience: null |
|
val_scheduler_criterion: |
|
- valid |
|
- loss |
|
early_stopping_criterion: |
|
- valid |
|
- loss |
|
- min |
|
best_model_criterion: |
|
- - valid |
|
- mAP |
|
- max |
|
keep_nbest_models: 1 |
|
nbest_averaging_interval: 0 |
|
grad_clip: 1 |
|
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_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 |
|
use_adapter: false |
|
adapter: lora |
|
save_strategy: all |
|
adapter_conf: {} |
|
pretrain_path: null |
|
init_param: [] |
|
ignore_init_mismatch: false |
|
freeze_param: [] |
|
num_iters_per_epoch: null |
|
batch_size: 80 |
|
valid_batch_size: 1200 |
|
batch_bins: 1000000 |
|
valid_batch_bins: null |
|
category_sample_size: 10 |
|
train_shape_file: |
|
- ./beats_runs/as20k_fulltrain/exp/cls_stats_16k/train/speech_shape |
|
- ./beats_runs/as20k_fulltrain/exp/cls_stats_16k/train/label_shape |
|
valid_shape_file: |
|
- ./beats_runs/as20k_fulltrain/exp/cls_stats_16k/valid/speech_shape |
|
- ./beats_runs/as20k_fulltrain/exp/cls_stats_16k/valid/label_shape |
|
batch_type: folded |
|
valid_batch_type: null |
|
fold_length: |
|
- 160000 |
|
- 600 |
|
sort_in_batch: descending |
|
shuffle_within_batch: false |
|
sort_batch: descending |
|
multiple_iterator: false |
|
chunk_length: 500 |
|
chunk_shift_ratio: 0.5 |
|
num_cache_chunks: 1024 |
|
chunk_excluded_key_prefixes: [] |
|
chunk_default_fs: null |
|
chunk_max_abs_length: null |
|
chunk_discard_short_samples: true |
|
train_data_path_and_name_and_type: |
|
- - ./beats_runs/as20k_fulltrain/dump/train/wav.scp |
|
- speech |
|
- sound |
|
- - ./beats_runs/as20k_fulltrain/dump/train/text |
|
- label |
|
- text |
|
valid_data_path_and_name_and_type: |
|
- - ./beats_runs/as20k_fulltrain/dump/val/wav.scp |
|
- speech |
|
- sound |
|
- - ./beats_runs/as20k_fulltrain/dump/val/text |
|
- label |
|
- text |
|
multi_task_dataset: false |
|
allow_variable_data_keys: false |
|
max_cache_size: 0.0 |
|
max_cache_fd: 32 |
|
allow_multi_rates: false |
|
valid_max_cache_size: null |
|
exclude_weight_decay: false |
|
exclude_weight_decay_conf: {} |
|
optim: adamw |
|
optim_conf: |
|
lr: 3.0e-05 |
|
weight_decay: 0.01 |
|
betas: |
|
- 0.9 |
|
- 0.98 |
|
scheduler: cosineannealingwarmuprestarts |
|
scheduler_conf: |
|
first_cycle_steps: 95000 |
|
warmup_steps: 8000 |
|
max_lr: 3.0e-05 |
|
min_lr: 5.0e-06 |
|
token_list: |
|
- Music |
|
- Speech |
|
- Vehicle |
|
- Inside,_small_room |
|
- Animal |
|
- Musical_instrument |
|
- Singing |
|
- Domestic_animals,_pets |
|
- Guitar |
|
- Plucked_string_instrument |
|
- Water |
|
- Car |
|
- Dog |
|
- Percussion |
|
- Wind_instrument,_woodwind_instrument |
|
- Outside,_urban_or_manmade |
|
- Outside,_rural_or_natural |
|
- Boat,_Water_vehicle |
|
- Brass_instrument |
|
- Fowl |
|
- Drum |
|
- Siren |
|
- Engine |
|
- Bird |
|
- Insect |
|
- Gunshot,_gunfire |
|
- Wood |
|
- Rail_transport |
|
- Train |
|
- Wind |
|
- Inside,_large_room_or_hall |
|
- Railroad_car,_train_wagon |
|
- Child_speech,_kid_speaking |
|
- Crowd |
|
- Rub |
|
- Keyboard_(musical) |
|
- Wind_noise_(microphone) |
|
- Pizzicato |
|
- Emergency_vehicle |
|
- Bird_vocalization,_bird_call,_bird_song |
|
- Livestock,_farm_animals,_working_animals |
|
- Cat |
|
- Organ |
|
- Fly,_housefly |
|
- Mechanisms |
|
- Bowed_string_instrument |
|
- Rain |
|
- Laughter |
|
- Aircraft |
|
- Electronic_music |
|
- Effects_unit |
|
- Hum |
|
- Tools |
|
- Drum_kit |
|
- Snare_drum |
|
- Hiss |
|
- Piano |
|
- Water_tap,_faucet |
|
- Rimshot |
|
- Bass_drum |
|
- Chicken,_rooster |
|
- Marimba,_xylophone |
|
- Horse |
|
- Song |
|
- Quack |
|
- Power_tool |
|
- Heart_sounds,_heartbeat |
|
- Goose |
|
- Hammond_organ |
|
- Rock_music |
|
- Ocean |
|
- Mains_hum |
|
- Thunder |
|
- Chime |
|
- Electronic_dance_music |
|
- Typing |
|
- Sink_(filling_or_washing) |
|
- Raindrop |
|
- Cello |
|
- Electric_guitar |
|
- Cheering |
|
- Church_bell |
|
- Christian_music |
|
- Drum_roll |
|
- Trombone |
|
- Glockenspiel |
|
- Trumpet |
|
- Cymbal |
|
- Tabla |
|
- Clickety-clack |
|
- Cricket |
|
- Steam_whistle |
|
- Explosion |
|
- Saxophone |
|
- Thunderstorm |
|
- Pop_music |
|
- Zither |
|
- Applause |
|
- Choir |
|
- Whack,_thwack |
|
- Clarinet |
|
- Camera |
|
- Electric_piano |
|
- Independent_music |
|
- Fire |
|
- Frog |
|
- Jet_engine |
|
- Music_of_Asia |
|
- Ding |
|
- Waves,_surf |
|
- Cattle,_bovinae |
|
- Turkey |
|
- Television |
|
- Coo |
|
- Scratching_(performance_technique) |
|
- Flute |
|
- Liquid |
|
- Harp |
|
- Progressive_rock |
|
- Happy_music |
|
- Steel_guitar,_slide_guitar |
|
- Whoosh,_swoosh,_swish |
|
- Boom |
|
- Breathing |
|
- Electronic_organ |
|
- Environmental_noise |
|
- Distortion |
|
- Alarm_clock |
|
- Fixed-wing_aircraft,_airplane |
|
- Violin,_fiddle |
|
- Whistling |
|
- Accordion |
|
- Disco |
|
- Pump_(liquid) |
|
- Waterfall |
|
- Beep,_bleep |
|
- Blues |
|
- Grunge |
|
- Hip_hop_music |
|
- Whistle |
|
- Fusillade |
|
- Splash,_splatter |
|
- Gush |
|
- Toothbrush |
|
- Knock |
|
- Gargling |
|
- Snoring |
|
- Hammer |
|
- Gobble |
|
- Walk,_footsteps |
|
- Jackhammer |
|
- Filing_(rasp) |
|
- Snort |
|
- Narration,_monologue |
|
- Tire_squeal |
|
- Fire_alarm |
|
- Squeal |
|
- Meow |
|
- Caterwaul |
|
- Cutlery,_silverware |
|
- Mantra |
|
- Opera |
|
- Classical_music |
|
- Theremin |
|
- Burst,_pop |
|
- Drip |
|
- Tick |
|
- Children_shouting |
|
- Creak |
|
- Hiccup |
|
- Pigeon,_dove |
|
- Bicycle_bell |
|
- Baby_cry,_infant_cry |
|
- Duck |
|
- Fireworks |
|
- Tambourine |
|
- Rodents,_rats,_mice |
|
- Buzzer |
|
- Splinter |
|
- Writing |
|
- Goat |
|
- Sheep |
|
- Heavy_metal |
|
- Ska |
|
- Neigh,_whinny |
|
- Sizzle |
|
- Rowboat,_canoe,_kayak |
|
- Wood_block |
|
- Clang |
|
- Door |
|
- Female_singing |
|
- Stream |
|
- Chant |
|
- Vocal_music |
|
- Yodeling |
|
- Bee,_wasp,_etc. |
|
- Air_brake |
|
- Whir |
|
- Bird_flight,_flapping_wings |
|
- French_horn |
|
- Telephone_dialing,_DTMF |
|
- Squeak |
|
- Sitar |
|
- Smoke_detector,_smoke_alarm |
|
- Tick-tock |
|
- Gurgling |
|
- Bellow |
|
- Harmonic |
|
- Male_singing |
|
- Giggle |
|
- Bark |
|
- Vibration |
|
- Drill |
|
- Skidding |
|
- Scratch |
|
- Drawer_open_or_close |
|
- Chop |
|
- Drum_machine |
|
- Squish |
|
- Toilet_flush |
|
- Fart |
|
- Basketball_bounce |
|
- Electronic_tuner |
|
- Singing_bowl |
|
- Squawk |
|
- Conversation |
|
- Reggae |
|
- Funny_music |
|
- Scrape |
|
- Sewing_machine |
|
- Tender_music |
|
- Swing_music |
|
- Dishes,_pots,_and_pans |
|
- Sampler |
|
- Synthesizer |
|
- Clapping |
|
- Hubbub,_speech_noise,_speech_babble |
|
- Engine_knocking |
|
- Canidae,_dogs,_wolves |
|
- Chainsaw |
|
- Pour |
|
- Croak |
|
- Chewing,_mastication |
|
- Cowbell |
|
- Propeller,_airscrew |
|
- Didgeridoo |
|
- Ringtone |
|
- Rattle_(instrument) |
|
- Artillery_fire |
|
- Cash_register |
|
- Crack |
|
- Growling |
|
- Mosquito |
|
- Carnatic_music |
|
- Honk |
|
- Howl |
|
- Cacophony |
|
- Gospel_music |
|
- Firecracker |
|
- Strum |
|
- Motorboat,_speedboat |
|
- Clock |
|
- Dance_music |
|
- Microwave_oven |
|
- Country |
|
- Bluegrass |
|
- Rattle |
|
- Mallet_percussion |
|
- Computer_keyboard |
|
- Bass_guitar |
|
- Electric_shaver,_electric_razor |
|
- Sawing |
|
- Owl |
|
- Whip |
|
- White_noise |
|
- Chirp_tone |
|
- Boiling |
|
- Ship |
|
- Mouse |
|
- Breaking |
|
- Silence |
|
- Throat_clearing |
|
- Bleat |
|
- Salsa_music |
|
- Patter |
|
- Vibraphone |
|
- Flap |
|
- Typewriter |
|
- Change_ringing_(campanology) |
|
- Trickle,_dribble |
|
- Video_game_music |
|
- Glass |
|
- Dial_tone |
|
- Radio |
|
- Bell |
|
- Moo |
|
- Heart_murmur |
|
- Clatter |
|
- Sniff |
|
- Double_bass |
|
- Background_music |
|
- Lawn_mower |
|
- Printer |
|
- House_music |
|
- Tearing |
|
- Angry_music |
|
- Male_speech,_man_speaking |
|
- Wild_animals |
|
- Cupboard_open_or_close |
|
- Harpsichord |
|
- Light_engine_(high_frequency) |
|
- Child_singing |
|
- Zipper_(clothing) |
|
- Jazz |
|
- Belly_laugh |
|
- Roar |
|
- Motor_vehicle_(road) |
|
- Crowing,_cock-a-doodle-doo |
|
- Cluck |
|
- Sad_music |
|
- Hi-hat |
|
- Cough |
|
- Stomach_rumble |
|
- Alarm |
|
- String_section |
|
- Sonar |
|
- Keys_jangling |
|
- Synthetic_singing |
|
- Rapping |
|
- Sidetone |
|
- Orchestra |
|
- Throbbing |
|
- Whale_vocalization |
|
- Thunk |
|
- Children_playing |
|
- Snake |
|
- Chink,_clink |
|
- Chirp,_tweet |
|
- Boing |
|
- Shuffle |
|
- Pulse |
|
- Punk_rock |
|
- Crow |
|
- Caw |
|
- Static |
|
- Clicking |
|
- Snicker |
|
- Whispering |
|
- Pink_noise |
|
- Crushing |
|
- Wedding_music |
|
- Crumpling,_crinkling |
|
- Crackle |
|
- Whoop |
|
- Electric_toothbrush |
|
- Train_wheels_squealing |
|
- Yell |
|
- Wind_chime |
|
- Frying_(food) |
|
- Christmas_music |
|
- Fill_(with_liquid) |
|
- Reverberation |
|
- Beatboxing |
|
- Harmonica |
|
- Banjo |
|
- Sliding_door |
|
- Groan |
|
- Bagpipes |
|
- Spray |
|
- Stir |
|
- Acoustic_guitar |
|
- Tap |
|
- Chorus_effect |
|
- Noise |
|
- Crunch |
|
- Biting |
|
- Aircraft_engine |
|
- Busy_signal |
|
- Bang |
|
- Techno |
|
- Tuning_fork |
|
- Tapping_(guitar_technique) |
|
- Pig |
|
- Maraca |
|
- Vacuum_cleaner |
|
- Mandolin |
|
- Electronica |
|
- Theme_music |
|
- Yip |
|
- A_capella |
|
- Rustle |
|
- Chatter |
|
- Traditional_music |
|
- Soul_music |
|
- Rustling_leaves |
|
- Afrobeat |
|
- Hoot |
|
- Slosh |
|
- Roaring_cats_(lions,_tigers) |
|
- Chopping_(food) |
|
- Heavy_engine_(low_frequency) |
|
- Sine_wave |
|
- Speech_synthesizer |
|
- Middle_Eastern_music |
|
- Music_of_Latin_America |
|
- Arrow |
|
- Timpani |
|
- Eruption |
|
- Shofar |
|
- Jingle_bell |
|
- Humming |
|
- Sanding |
|
- Female_speech,_woman_speaking |
|
- Gong |
|
- Rain_on_surface |
|
- Pant |
|
- Dubstep |
|
- Clip-clop |
|
- Finger_snapping |
|
- Blender |
|
- Drum_and_bass |
|
- Bouncing |
|
- Vehicle_horn,_car_horn,_honking |
|
- Slam |
|
- Idling |
|
- Rhythm_and_blues |
|
- Race_car,_auto_racing |
|
- Single-lens_reflex_camera |
|
- Smash,_crash |
|
- Purr |
|
- Shatter |
|
- Steelpan |
|
- Whimper_(dog) |
|
- Power_windows,_electric_windows |
|
- Battle_cry |
|
- Scary_music |
|
- Hands |
|
- Echo |
|
- Truck |
|
- Buzz |
|
- Mechanical_fan |
|
- Plop |
|
- Run |
|
- Gasp |
|
- Psychedelic_rock |
|
- Grunt |
|
- Helicopter |
|
- Dental_drill,_dentist's_drill |
|
- Babbling |
|
- Zing |
|
- Oink |
|
- Soundtrack_music |
|
- Ambulance_(siren) |
|
- Exciting_music |
|
- Telephone |
|
- Jingle_(music) |
|
- Tubular_bells |
|
- Burping,_eructation |
|
- Baby_laughter |
|
- Ping |
|
- Bow-wow |
|
- Foghorn |
|
- Machine_gun |
|
- Ukulele |
|
- Telephone_bell_ringing |
|
- Pulleys |
|
- Gears |
|
- Sigh |
|
- Coin_(dropping) |
|
- Music_of_Africa |
|
- Scissors |
|
- Inside,_public_space |
|
- Trance_music |
|
- Roll |
|
- Thump,_thud |
|
- Air_conditioning |
|
- Ding-dong |
|
- Ratchet,_pawl |
|
- Hair_dryer |
|
- Shout |
|
- Ambient_music |
|
- Music_for_children |
|
- Toot |
|
- Bathtub_(filling_or_washing) |
|
- Slap,_smack |
|
- Chuckle,_chortle |
|
- Traffic_noise,_roadway_noise |
|
- Bicycle |
|
- Whimper |
|
- Doorbell |
|
- Wheeze |
|
- Sailboat,_sailing_ship |
|
- Cap_gun |
|
- Wail,_moan |
|
- Rock_and_roll |
|
- Jingle,_tinkle |
|
- Fire_engine,_fire_truck_(siren) |
|
- Funk |
|
- Lullaby |
|
- Field_recording |
|
- Skateboard |
|
- Steam |
|
- Rumble |
|
- Medium_engine_(mid_frequency) |
|
- Sound_effect |
|
- Flamenco |
|
- Shuffling_cards |
|
- Subway,_metro,_underground |
|
- Police_car_(siren) |
|
- Folk_music |
|
- Crying,_sobbing |
|
- New-age_music |
|
- Ice_cream_truck,_ice_cream_van |
|
- Music_of_Bollywood |
|
- Accelerating,_revving,_vroom |
|
- Screaming |
|
- Motorcycle |
|
- Engine_starting |
|
- Train_whistle |
|
- Car_passing_by |
|
- Bus |
|
- Sneeze |
|
- Train_horn |
|
- Air_horn,_truck_horn |
|
- Civil_defense_siren |
|
- Car_alarm |
|
- Reversing_beeps |
|
- <unk> |
|
token_type: word |
|
init: xavier_normal |
|
input_size: 1 |
|
use_preprocessor: true |
|
frontend: null |
|
frontend_conf: {} |
|
specaug: null |
|
specaug_conf: {} |
|
normalize: null |
|
normalize_conf: {} |
|
preencoder: null |
|
preencoder_conf: {} |
|
encoder: beats |
|
encoder_conf: |
|
beats_ckpt_path: /compute/babel-13-33/sbharad2/models/BEATs/BEATs_iter3_plus_AS20K.pt |
|
beats_config: |
|
layer_wise_gradient_decay_ratio: 0.3 |
|
encoder_layerdrop: 0.1 |
|
dropout: 0.0 |
|
use_weighted_representation: false |
|
specaug_config: |
|
apply_time_warp: true |
|
apply_freq_mask: false |
|
apply_time_mask: true |
|
time_mask_width_ratio_range: |
|
- 0 |
|
- 0.06 |
|
num_time_mask: 1 |
|
roll_augment: true |
|
roll_interval: 1 |
|
decoder: linear |
|
decoder_conf: {} |
|
model: espnet |
|
model_conf: |
|
classification_type: multi-label |
|
mixup_augmentation: true |
|
lsm_weight: 0.0 |
|
required: |
|
- output_dir |
|
- token_list |
|
version: '202412' |
|
distributed: false |
|
``` |
|
|
|
</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} |
|
} |
|
``` |
|
|
