upload files

README.md

@@ -0,0 +1 @@
+See <https://github.com/k2-fsa/icefall/pull/479>

data/lang_bpe_500/L.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1b88996f918737fba67fbd29152018b51a537c16ce0718a2b43d5140583224e
+size 19025703

data/lang_bpe_500/LG.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3bb9f021c7aad79d45dc275ba8154a430c4f660a319dcb872cd52500f25553d6
+size 249852195

data/lang_bpe_500/Linv.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbc8b3687a1b8f0811a84106b3b310642566c7b1bc282a929878f9269507a2c6
+size 19025703

data/lang_bpe_500/bpe.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c53433de083c4a6ad12d034550ef22de68cec62c4f58932a7b6b8b2f1e743fa5
+size 244865

data/lang_bpe_500/tokens.txt

@@ -0,0 +1,502 @@
+<blk> 0
+<sos/eos> 1
+<unk> 2
+S 3
+▁THE 4
+▁A 5
+T 6
+▁AND 7
+ED 8
+▁OF 9
+▁TO 10
+E 11
+D 12
+N 13
+ING 14
+▁IN 15
+Y 16
+M 17
+C 18
+▁I 19
+A 20
+P 21
+▁HE 22
+R 23
+O 24
+L 25
+RE 26
+I 27
+U 28
+ER 29
+▁IT 30
+LY 31
+▁THAT 32
+▁WAS 33
+▁ 34
+▁S 35
+AR 36
+▁BE 37
+F 38
+▁C 39
+IN 40
+B 41
+▁FOR 42
+OR 43
+LE 44
+' 45
+▁HIS 46
+▁YOU 47
+AL 48
+▁RE 49
+V 50
+▁B 51
+G 52
+RI 53
+▁E 54
+▁WITH 55
+▁T 56
+▁AS 57
+LL 58
+▁P 59
+▁HER 60
+ST 61
+▁HAD 62
+▁SO 63
+▁F 64
+W 65
+CE 66
+▁IS 67
+ND 68
+▁NOT 69
+TH 70
+▁BUT 71
+EN 72
+▁SHE 73
+▁ON 74
+VE 75
+ON 76
+SE 77
+▁DE 78
+UR 79
+▁G 80
+CH 81
+K 82
+TER 83
+▁AT 84
+IT 85
+▁ME 86
+RO 87
+NE 88
+RA 89
+ES 90
+IL 91
+NG 92
+IC 93
+▁NO 94
+▁HIM 95
+ENT 96
+IR 97
+▁WE 98
+H 99
+▁DO 100
+▁ALL 101
+▁HAVE 102
+LO 103
+▁BY 104
+▁MY 105
+▁MO 106
+▁THIS 107
+LA 108
+▁ST 109
+▁WHICH 110
+▁CON 111
+▁THEY 112
+CK 113
+TE 114
+▁SAID 115
+▁FROM 116
+▁GO 117
+▁WHO 118
+▁TH 119
+▁OR 120
+▁D 121
+▁W 122
+VER 123
+LI 124
+▁SE 125
+▁ONE 126
+▁CA 127
+▁AN 128
+▁LA 129
+▁WERE 130
+EL 131
+▁HA 132
+▁MAN 133
+▁FA 134
+▁EX 135
+AD 136
+▁SU 137
+RY 138
+▁MI 139
+AT 140
+▁BO 141
+▁WHEN 142
+AN 143
+THER 144
+PP 145
+ATION 146
+▁FI 147
+▁WOULD 148
+▁PRO 149
+OW 150
+ET 151
+▁O 152
+▁THERE 153
+▁HO 154
+ION 155
+▁WHAT 156
+▁FE 157
+▁PA 158
+US 159
+MENT 160
+▁MA 161
+UT 162
+▁OUT 163
+▁THEIR 164
+▁IF 165
+▁LI 166
+▁K 167
+▁WILL 168
+▁ARE 169
+ID 170
+▁RO 171
+DE 172
+TION 173
+▁WA 174
+PE 175
+▁UP 176
+▁SP 177
+▁PO 178
+IGHT 179
+▁UN 180
+RU 181
+▁LO 182
+AS 183
+OL 184
+▁LE 185
+▁BEEN 186
+▁SH 187
+▁RA 188
+▁SEE 189
+KE 190
+UL 191
+TED 192
+▁SA 193
+UN 194
+UND 195
+ANT 196
+▁NE 197
+IS 198
+▁THEM 199
+CI 200
+GE 201
+▁COULD 202
+▁DIS 203
+OM 204
+ISH 205
+HE 206
+EST 207
+▁SOME 208
+ENCE 209
+ITY 210
+IVE 211
+▁US 212
+▁MORE 213
+▁EN 214
+ARD 215
+ATE 216
+▁YOUR 217
+▁INTO 218
+▁KNOW 219
+▁CO 220
+ANCE 221
+▁TIME 222
+▁WI 223
+▁YE 224
+AGE 225
+▁NOW 226
+TI 227
+FF 228
+ABLE 229
+▁VERY 230
+▁LIKE 231
+AM 232
+HI 233
+Z 234
+▁OTHER 235
+▁THAN 236
+▁LITTLE 237
+▁DID 238
+▁LOOK 239
+TY 240
+ERS 241
+▁CAN 242
+▁CHA 243
+▁AR 244
+X 245
+FUL 246
+UGH 247
+▁BA 248
+▁DAY 249
+▁ABOUT 250
+TEN 251
+IM 252
+▁ANY 253
+▁PRE 254
+▁OVER 255
+IES 256
+NESS 257
+ME 258
+BLE 259
+▁M 260
+ROW 261
+▁HAS 262
+▁GREAT 263
+▁VI 264
+TA 265
+▁AFTER 266
+PER 267
+▁AGAIN 268
+HO 269
+SH 270
+▁UPON 271
+▁DI 272
+▁HAND 273
+▁COM 274
+IST 275
+TURE 276
+▁STA 277
+▁THEN 278
+▁SHOULD 279
+▁GA 280
+OUS 281
+OUR 282
+▁WELL 283
+▁ONLY 284
+MAN 285
+▁GOOD 286
+▁TWO 287
+▁MAR 288
+▁SAY 289
+▁HU 290
+TING 291
+▁OUR 292
+RESS 293
+▁DOWN 294
+IOUS 295
+▁BEFORE 296
+▁DA 297
+▁NA 298
+QUI 299
+▁MADE 300
+▁EVERY 301
+▁OLD 302
+▁EVEN 303
+IG 304
+▁COME 305
+▁GRA 306
+▁RI 307
+▁LONG 308
+OT 309
+SIDE 310
+WARD 311
+▁FO 312
+▁WHERE 313
+MO 314
+LESS 315
+▁SC 316
+▁MUST 317
+▁NEVER 318
+▁HOW 319
+▁CAME 320
+▁SUCH 321
+▁RU 322
+▁TAKE 323
+▁WO 324
+▁CAR 325
+UM 326
+AK 327
+▁THINK 328
+▁MUCH 329
+▁MISTER 330
+▁MAY 331
+▁JO 332
+▁WAY 333
+▁COMP 334
+▁THOUGHT 335
+▁STO 336
+▁MEN 337
+▁BACK 338
+▁DON 339
+J 340
+▁LET 341
+▁TRA 342
+▁FIRST 343
+▁JUST 344
+▁VA 345
+▁OWN 346
+▁PLA 347
+▁MAKE 348
+ATED 349
+▁HIMSELF 350
+▁WENT 351
+▁PI 352
+GG 353
+RING 354
+▁DU 355
+▁MIGHT 356
+▁PART 357
+▁GIVE 358
+▁IMP 359
+▁BU 360
+▁PER 361
+▁PLACE 362
+▁HOUSE 363
+▁THROUGH 364
+IAN 365
+▁SW 366
+▁UNDER 367
+QUE 368
+▁AWAY 369
+▁LOVE 370
+QUA 371
+▁LIFE 372
+▁GET 373
+▁WITHOUT 374
+▁PASS 375
+▁TURN 376
+IGN 377
+▁HEAD 378
+▁MOST 379
+▁THOSE 380
+▁SHALL 381
+▁EYES 382
+▁COL 383
+▁STILL 384
+▁NIGHT 385
+▁NOTHING 386
+ITION 387
+HA 388
+▁TELL 389
+▁WORK 390
+▁LAST 391
+▁NEW 392
+▁FACE 393
+▁HI 394
+▁WORD 395
+▁FOUND 396
+▁COUNT 397
+▁OB 398
+▁WHILE 399
+▁SHA 400
+▁MEAN 401
+▁SAW 402
+▁PEOPLE 403
+▁FRIEND 404
+▁THREE 405
+▁ROOM 406
+▁SAME 407
+▁THOUGH 408
+▁RIGHT 409
+▁CHILD 410
+▁FATHER 411
+▁ANOTHER 412
+▁HEART 413
+▁WANT 414
+▁TOOK 415
+OOK 416
+▁LIGHT 417
+▁MISSUS 418
+▁OPEN 419
+▁JU 420
+▁ASKED 421
+PORT 422
+▁LEFT 423
+▁JA 424
+▁WORLD 425
+▁HOME 426
+▁WHY 427
+▁ALWAYS 428
+▁ANSWER 429
+▁SEEMED 430
+▁SOMETHING 431
+▁GIRL 432
+▁BECAUSE 433
+▁NAME 434
+▁TOLD 435
+▁NI 436
+▁HIGH 437
+IZE 438
+▁WOMAN 439
+▁FOLLOW 440
+▁RETURN 441
+▁KNEW 442
+▁EACH 443
+▁KIND 444
+▁JE 445
+▁ACT 446
+▁LU 447
+▁CERTAIN 448
+▁YEARS 449
+▁QUITE 450
+▁APPEAR 451
+▁BETTER 452
+▁HALF 453
+▁PRESENT 454
+▁PRINCE 455
+SHIP 456
+▁ALSO 457
+▁BEGAN 458
+▁HAVING 459
+▁ENOUGH 460
+▁PERSON 461
+▁LADY 462
+▁WHITE 463
+▁COURSE 464
+▁VOICE 465
+▁SPEAK 466
+▁POWER 467
+▁MORNING 468
+▁BETWEEN 469
+▁AMONG 470
+▁KEEP 471
+▁WALK 472
+▁MATTER 473
+▁TEA 474
+▁BELIEVE 475
+▁SMALL 476
+▁TALK 477
+▁FELT 478
+▁HORSE 479
+▁MYSELF 480
+▁SIX 481
+▁HOWEVER 482
+▁FULL 483
+▁HERSELF 484
+▁POINT 485
+▁STOOD 486
+▁HUNDRED 487
+▁ALMOST 488
+▁SINCE 489
+▁LARGE 490
+▁LEAVE 491
+▁PERHAPS 492
+▁DARK 493
+▁SUDDEN 494
+▁REPLIED 495
+▁ANYTHING 496
+▁WONDER 497
+▁UNTIL 498
+Q 499
+#0 500
+#1 501

decoding_results/fast_beam_search/log-decode-epoch-35-avg-15-beam-20.0-max-contexts-8-max-states-64-use-averaged-model-2022-08-18-15-34-20

@@ -0,0 +1,27 @@
+2022-08-18 15:34:20,540 INFO [decode.py:661] Decoding started
+2022-08-18 15:34:20,540 INFO [decode.py:667] Device: cuda:0
+2022-08-18 15:34:20,544 INFO [decode.py:677] {'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'dim_feedforward': 2048, 'decoder_dim': 512, 'joiner_dim': 512, 'model_warm_step': 3000, 'env_info': {'k2-version': '1.15.1', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '350e449fe5d7b4231f77eb0add764782eed9f5d2', 'k2-git-date': 'Thu May 26 22:56:24 2022', 'lhotse-version': '1.4.0.dev+git.21b5b3f.dirty', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'lstm', 'icefall-git-sha1': '7ee3701-dirty', 'icefall-git-date': 'Sun Aug 14 17:10:39 2022', 'icefall-path': '/ceph-zw/workspace/rnn/icefall', 'k2-path': '/ceph-zw/workspace/share/k2/k2/python/k2/__init__.py', 'lhotse-path': '/ceph-jb/yaozengwei/workspace/rnnt/lhotse/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-0307195509-54c966b95f-rtpfq', 'IP address': '10.177.22.9'}, 'epoch': 35, 'iter': 0, 'avg': 15, 'use_averaged_model': True, 'exp_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'lang_dir': PosixPath('data/lang_bpe_500'), 'decoding_method': 'fast_beam_search', 'beam_size': 4, 'beam': 20.0, 'ngram_lm_scale': 0.01, 'max_contexts': 8, 'max_states': 64, 'context_size': 2, 'max_sym_per_frame': 1, 'num_paths': 200, 'nbest_scale': 0.5, 'num_encoder_layers': 12, 'encoder_dim': 512, 'rnn_hidden_size': 1024, 'aux_layer_period': 0, 'full_libri': True, 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 600, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'input_strategy': 'PrecomputedFeatures', 'res_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/fast_beam_search'), 'suffix': 'epoch-35-avg-15-beam-20.0-max-contexts-8-max-states-64-use-averaged-model', 'blank_id': 0, 'unk_id': 2, 'vocab_size': 500}
+2022-08-18 15:34:20,544 INFO [decode.py:679] About to create model
+2022-08-18 15:34:20,872 INFO [decode.py:746] Calculating the averaged model over epoch range from 20 (excluded) to 35
+2022-08-18 15:34:26,102 INFO [decode.py:782] Number of model parameters: 84689496
+2022-08-18 15:34:26,103 INFO [asr_datamodule.py:444] About to get test-clean cuts
+2022-08-18 15:34:26,104 INFO [asr_datamodule.py:451] About to get test-other cuts
+2022-08-18 15:34:28,208 INFO [decode.py:563] batch 0/?, cuts processed until now is 47
+2022-08-18 15:34:55,417 INFO [decode.py:563] batch 20/?, cuts processed until now is 1656
+2022-08-18 15:35:18,232 INFO [decode.py:581] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/fast_beam_search/recogs-test-clean-beam_20.0_max_contexts_8_max_states_64-epoch-35-avg-15-beam-20.0-max-contexts-8-max-states-64-use-averaged-model.txt
+2022-08-18 15:35:18,301 INFO [utils.py:416] [test-clean-beam_20.0_max_contexts_8_max_states_64] %WER 3.74% [1967 / 52576, 209 ins, 169 del, 1589 sub ]
+2022-08-18 15:35:18,470 INFO [decode.py:594] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/fast_beam_search/errs-test-clean-beam_20.0_max_contexts_8_max_states_64-epoch-35-avg-15-beam-20.0-max-contexts-8-max-states-64-use-averaged-model.txt
+2022-08-18 15:35:18,471 INFO [decode.py:611] 
+For test-clean, WER of different settings are:
+beam_20.0_max_contexts_8_max_states_64	3.74	best for test-clean
+
+2022-08-18 15:35:20,170 INFO [decode.py:563] batch 0/?, cuts processed until now is 57
+2022-08-18 15:35:46,099 INFO [decode.py:563] batch 20/?, cuts processed until now is 1886
+2022-08-18 15:36:07,644 INFO [decode.py:581] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/fast_beam_search/recogs-test-other-beam_20.0_max_contexts_8_max_states_64-epoch-35-avg-15-beam-20.0-max-contexts-8-max-states-64-use-averaged-model.txt
+2022-08-18 15:36:07,715 INFO [utils.py:416] [test-other-beam_20.0_max_contexts_8_max_states_64] %WER 9.59% [5020 / 52343, 499 ins, 516 del, 4005 sub ]
+2022-08-18 15:36:07,890 INFO [decode.py:594] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/fast_beam_search/errs-test-other-beam_20.0_max_contexts_8_max_states_64-epoch-35-avg-15-beam-20.0-max-contexts-8-max-states-64-use-averaged-model.txt
+2022-08-18 15:36:07,891 INFO [decode.py:611] 
+For test-other, WER of different settings are:
+beam_20.0_max_contexts_8_max_states_64	9.59	best for test-other
+
+2022-08-18 15:36:07,891 INFO [decode.py:813] Done!

decoding_results/fast_beam_search/wer-summary-test-clean-beam_20.0_max_contexts_8_max_states_64-epoch-35-avg-15-beam-20.0-max-contexts-8-max-states-64-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+beam_20.0_max_contexts_8_max_states_64	3.74

decoding_results/fast_beam_search/wer-summary-test-other-beam_20.0_max_contexts_8_max_states_64-epoch-35-avg-15-beam-20.0-max-contexts-8-max-states-64-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+beam_20.0_max_contexts_8_max_states_64	9.59

decoding_results/greedy_search/log-decode-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model-2022-08-18-14-34-10

@@ -0,0 +1,25 @@
+2022-08-18 14:34:10,773 INFO [decode.py:661] Decoding started
+2022-08-18 14:34:10,774 INFO [decode.py:667] Device: cuda:0
+2022-08-18 14:34:10,778 INFO [decode.py:677] {'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'dim_feedforward': 2048, 'decoder_dim': 512, 'joiner_dim': 512, 'model_warm_step': 3000, 'env_info': {'k2-version': '1.15.1', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '350e449fe5d7b4231f77eb0add764782eed9f5d2', 'k2-git-date': 'Thu May 26 22:56:24 2022', 'lhotse-version': '1.4.0.dev+git.21b5b3f.dirty', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'lstm', 'icefall-git-sha1': '7ee3701-dirty', 'icefall-git-date': 'Sun Aug 14 17:10:39 2022', 'icefall-path': '/ceph-zw/workspace/rnn/icefall', 'k2-path': '/ceph-zw/workspace/share/k2/k2/python/k2/__init__.py', 'lhotse-path': '/ceph-jb/yaozengwei/workspace/rnnt/lhotse/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-0307195509-54c966b95f-rtpfq', 'IP address': '10.177.22.9'}, 'epoch': 35, 'iter': 0, 'avg': 15, 'use_averaged_model': True, 'exp_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'lang_dir': PosixPath('data/lang_bpe_500'), 'decoding_method': 'greedy_search', 'beam_size': 4, 'beam': 20.0, 'ngram_lm_scale': 0.01, 'max_contexts': 8, 'max_states': 64, 'context_size': 2, 'max_sym_per_frame': 1, 'num_paths': 200, 'nbest_scale': 0.5, 'num_encoder_layers': 12, 'encoder_dim': 512, 'rnn_hidden_size': 1024, 'aux_layer_period': 0, 'full_libri': True, 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 600, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'input_strategy': 'PrecomputedFeatures', 'res_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/greedy_search'), 'suffix': 'epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model', 'blank_id': 0, 'unk_id': 2, 'vocab_size': 500}
+2022-08-18 14:34:10,778 INFO [decode.py:679] About to create model
+2022-08-18 14:34:11,120 INFO [decode.py:746] Calculating the averaged model over epoch range from 20 (excluded) to 35
+2022-08-18 14:34:28,114 INFO [decode.py:782] Number of model parameters: 84689496
+2022-08-18 14:34:28,114 INFO [asr_datamodule.py:444] About to get test-clean cuts
+2022-08-18 14:34:28,117 INFO [asr_datamodule.py:451] About to get test-other cuts
+2022-08-18 14:34:29,506 INFO [decode.py:563] batch 0/?, cuts processed until now is 47
+2022-08-18 14:34:54,910 INFO [decode.py:581] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/greedy_search/recogs-test-clean-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt
+2022-08-18 14:34:54,979 INFO [utils.py:416] [test-clean-greedy_search] %WER 3.81% [2003 / 52576, 202 ins, 176 del, 1625 sub ]
+2022-08-18 14:34:55,145 INFO [decode.py:594] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/greedy_search/errs-test-clean-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt
+2022-08-18 14:34:55,146 INFO [decode.py:611] 
+For test-clean, WER of different settings are:
+greedy_search	3.81	best for test-clean
+
+2022-08-18 14:34:56,218 INFO [decode.py:563] batch 0/?, cuts processed until now is 57
+2022-08-18 14:35:20,002 INFO [decode.py:581] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/greedy_search/recogs-test-other-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt
+2022-08-18 14:35:20,075 INFO [utils.py:416] [test-other-greedy_search] %WER 9.73% [5094 / 52343, 507 ins, 534 del, 4053 sub ]
+2022-08-18 14:35:20,251 INFO [decode.py:594] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/greedy_search/errs-test-other-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt
+2022-08-18 14:35:20,252 INFO [decode.py:611] 
+For test-other, WER of different settings are:
+greedy_search	9.73	best for test-other
+
+2022-08-18 14:35:20,252 INFO [decode.py:813] Done!

decoding_results/greedy_search/wer-summary-test-clean-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+greedy_search	3.81

decoding_results/greedy_search/wer-summary-test-other-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+greedy_search	9.73

decoding_results/modified_beam_search/log-decode-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model-2022-08-18-15-06-34

@@ -0,0 +1,27 @@
+2022-08-18 15:06:34,863 INFO [decode.py:661] Decoding started
+2022-08-18 15:06:34,863 INFO [decode.py:667] Device: cuda:0
+2022-08-18 15:06:34,866 INFO [decode.py:677] {'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'dim_feedforward': 2048, 'decoder_dim': 512, 'joiner_dim': 512, 'model_warm_step': 3000, 'env_info': {'k2-version': '1.15.1', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '350e449fe5d7b4231f77eb0add764782eed9f5d2', 'k2-git-date': 'Thu May 26 22:56:24 2022', 'lhotse-version': '1.4.0.dev+git.21b5b3f.dirty', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'lstm', 'icefall-git-sha1': '7ee3701-dirty', 'icefall-git-date': 'Sun Aug 14 17:10:39 2022', 'icefall-path': '/ceph-zw/workspace/rnn/icefall', 'k2-path': '/ceph-zw/workspace/share/k2/k2/python/k2/__init__.py', 'lhotse-path': '/ceph-jb/yaozengwei/workspace/rnnt/lhotse/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-0307195509-54c966b95f-rtpfq', 'IP address': '10.177.22.9'}, 'epoch': 35, 'iter': 0, 'avg': 15, 'use_averaged_model': True, 'exp_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'lang_dir': PosixPath('data/lang_bpe_500'), 'decoding_method': 'modified_beam_search', 'beam_size': 4, 'beam': 20.0, 'ngram_lm_scale': 0.01, 'max_contexts': 8, 'max_states': 64, 'context_size': 2, 'max_sym_per_frame': 1, 'num_paths': 200, 'nbest_scale': 0.5, 'num_encoder_layers': 12, 'encoder_dim': 512, 'rnn_hidden_size': 1024, 'aux_layer_period': 0, 'full_libri': True, 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 600, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'input_strategy': 'PrecomputedFeatures', 'res_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/modified_beam_search'), 'suffix': 'epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model', 'blank_id': 0, 'unk_id': 2, 'vocab_size': 500}
+2022-08-18 15:06:34,866 INFO [decode.py:679] About to create model
+2022-08-18 15:06:35,200 INFO [decode.py:746] Calculating the averaged model over epoch range from 20 (excluded) to 35
+2022-08-18 15:06:40,540 INFO [decode.py:782] Number of model parameters: 84689496
+2022-08-18 15:06:40,540 INFO [asr_datamodule.py:444] About to get test-clean cuts
+2022-08-18 15:06:40,543 INFO [asr_datamodule.py:451] About to get test-other cuts
+2022-08-18 15:06:44,793 INFO [decode.py:563] batch 0/?, cuts processed until now is 47
+2022-08-18 15:07:55,419 INFO [decode.py:563] batch 20/?, cuts processed until now is 1656
+2022-08-18 15:08:41,172 INFO [decode.py:581] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/modified_beam_search/recogs-test-clean-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt
+2022-08-18 15:08:41,240 INFO [utils.py:416] [test-clean-beam_size_4] %WER 3.64% [1913 / 52576, 204 ins, 154 del, 1555 sub ]
+2022-08-18 15:08:41,409 INFO [decode.py:594] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/modified_beam_search/errs-test-clean-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt
+2022-08-18 15:08:41,410 INFO [decode.py:611] 
+For test-clean, WER of different settings are:
+beam_size_4	3.64	best for test-clean
+
+2022-08-18 15:08:45,255 INFO [decode.py:563] batch 0/?, cuts processed until now is 57
+2022-08-18 15:09:54,333 INFO [decode.py:563] batch 20/?, cuts processed until now is 1886
+2022-08-18 15:10:37,739 INFO [decode.py:581] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/modified_beam_search/recogs-test-other-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt
+2022-08-18 15:10:37,809 INFO [utils.py:416] [test-other-beam_size_4] %WER 9.55% [5001 / 52343, 540 ins, 454 del, 4007 sub ]
+2022-08-18 15:10:37,984 INFO [decode.py:594] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/modified_beam_search/errs-test-other-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt
+2022-08-18 15:10:37,985 INFO [decode.py:611] 
+For test-other, WER of different settings are:
+beam_size_4	9.55	best for test-other
+
+2022-08-18 15:10:37,985 INFO [decode.py:813] Done!

decoding_results/modified_beam_search/wer-summary-test-clean-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+beam_size_4	3.64

decoding_results/modified_beam_search/wer-summary-test-other-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+beam_size_4	9.55

decoding_results/streaming/fast_beam_search/log-streaming-decode-2022-08-18-15-51-58

@@ -0,0 +1,42 @@
+2022-08-18 15:51:58,980 INFO [streaming_decode.py:830] Decoding started
+2022-08-18 15:51:58,981 INFO [streaming_decode.py:836] Device: cuda:0
+2022-08-18 15:51:58,984 INFO [streaming_decode.py:848] {'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'dim_feedforward': 2048, 'decoder_dim': 512, 'joiner_dim': 512, 'model_warm_step': 3000, 'env_info': {'k2-version': '1.15.1', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '350e449fe5d7b4231f77eb0add764782eed9f5d2', 'k2-git-date': 'Thu May 26 22:56:24 2022', 'lhotse-version': '1.4.0.dev+git.21b5b3f.dirty', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'lstm', 'icefall-git-sha1': '7ee3701-dirty', 'icefall-git-date': 'Sun Aug 14 17:10:39 2022', 'icefall-path': '/ceph-zw/workspace/rnn/icefall', 'k2-path': '/ceph-zw/workspace/share/k2/k2/python/k2/__init__.py', 'lhotse-path': '/ceph-jb/yaozengwei/workspace/rnnt/lhotse/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-0307195509-54c966b95f-rtpfq', 'IP address': '10.177.22.9'}, 'epoch': 35, 'iter': 0, 'avg': 15, 'use_averaged_model': True, 'exp_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'decoding_method': 'fast_beam_search', 'beam_size': 4, 'beam': 4.0, 'max_contexts': 4, 'max_states': 8, 'context_size': 2, 'max_sym_per_frame': 1, 'sampling_rate': 16000, 'num_decode_streams': 2000, 'num_encoder_layers': 12, 'encoder_dim': 512, 'rnn_hidden_size': 1024, 'aux_layer_period': 0, 'full_libri': True, 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 200.0, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'input_strategy': 'PrecomputedFeatures', 'res_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/fast_beam_search'), 'suffix': 'epoch-35-avg-15-beam-4.0-max-contexts-4-max-states-8-use-averaged-model', 'blank_id': 0, 'unk_id': 2, 'vocab_size': 500, 'device': device(type='cuda', index=0)}
+2022-08-18 15:51:58,984 INFO [streaming_decode.py:850] About to create model
+2022-08-18 15:51:59,318 INFO [streaming_decode.py:917] Calculating the averaged model over epoch range from 20 (excluded) to 35
+2022-08-18 15:52:04,652 INFO [streaming_decode.py:938] Number of model parameters: 84689496
+2022-08-18 15:52:04,653 INFO [asr_datamodule.py:444] About to get test-clean cuts
+2022-08-18 15:52:04,655 INFO [asr_datamodule.py:451] About to get test-other cuts
+2022-08-18 15:52:04,815 INFO [streaming_decode.py:714] Cuts processed until now is 0.
+2022-08-18 15:52:06,858 INFO [streaming_decode.py:714] Cuts processed until now is 300.
+2022-08-18 15:52:08,951 INFO [streaming_decode.py:714] Cuts processed until now is 600.
+2022-08-18 15:52:11,018 INFO [streaming_decode.py:714] Cuts processed until now is 900.
+2022-08-18 15:52:13,126 INFO [streaming_decode.py:714] Cuts processed until now is 1200.
+2022-08-18 15:52:15,237 INFO [streaming_decode.py:714] Cuts processed until now is 1500.
+2022-08-18 15:52:17,232 INFO [streaming_decode.py:714] Cuts processed until now is 1800.
+2022-08-18 15:53:03,093 INFO [streaming_decode.py:714] Cuts processed until now is 2100.
+2022-08-18 15:53:24,162 INFO [streaming_decode.py:714] Cuts processed until now is 2400.
+2022-08-18 15:56:28,685 INFO [streaming_decode.py:759] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/fast_beam_search/recogs-test-clean-beam_4.0_max_contexts_4_max_states_8-epoch-35-avg-15-beam-4.0-max-contexts-4-max-states-8-use-averaged-model.txt
+2022-08-18 15:56:28,753 INFO [utils.py:416] [test-clean-beam_4.0_max_contexts_4_max_states_8] %WER 3.73% [1963 / 52576, 204 ins, 175 del, 1584 sub ]
+2022-08-18 15:56:28,983 INFO [streaming_decode.py:772] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/fast_beam_search/errs-test-clean-beam_4.0_max_contexts_4_max_states_8-epoch-35-avg-15-beam-4.0-max-contexts-4-max-states-8-use-averaged-model.txt
+2022-08-18 15:56:28,983 INFO [streaming_decode.py:789] 
+For test-clean, WER of different settings are:
+beam_4.0_max_contexts_4_max_states_8	3.73	best for test-clean
+
+2022-08-18 15:56:29,005 INFO [streaming_decode.py:714] Cuts processed until now is 0.
+2022-08-18 15:56:31,680 INFO [streaming_decode.py:714] Cuts processed until now is 300.
+2022-08-18 15:56:34,205 INFO [streaming_decode.py:714] Cuts processed until now is 600.
+2022-08-18 15:56:36,710 INFO [streaming_decode.py:714] Cuts processed until now is 900.
+2022-08-18 15:56:39,127 INFO [streaming_decode.py:714] Cuts processed until now is 1200.
+2022-08-18 15:56:41,551 INFO [streaming_decode.py:714] Cuts processed until now is 1500.
+2022-08-18 15:56:43,969 INFO [streaming_decode.py:714] Cuts processed until now is 1800.
+2022-08-18 15:57:28,379 INFO [streaming_decode.py:714] Cuts processed until now is 2100.
+2022-08-18 15:57:46,312 INFO [streaming_decode.py:714] Cuts processed until now is 2400.
+2022-08-18 15:58:05,676 INFO [streaming_decode.py:714] Cuts processed until now is 2700.
+2022-08-18 16:00:50,495 INFO [streaming_decode.py:759] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/fast_beam_search/recogs-test-other-beam_4.0_max_contexts_4_max_states_8-epoch-35-avg-15-beam-4.0-max-contexts-4-max-states-8-use-averaged-model.txt
+2022-08-18 16:00:50,566 INFO [utils.py:416] [test-other-beam_4.0_max_contexts_4_max_states_8] %WER 9.61% [5030 / 52343, 509 ins, 508 del, 4013 sub ]
+2022-08-18 16:00:50,742 INFO [streaming_decode.py:772] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/fast_beam_search/errs-test-other-beam_4.0_max_contexts_4_max_states_8-epoch-35-avg-15-beam-4.0-max-contexts-4-max-states-8-use-averaged-model.txt
+2022-08-18 16:00:50,743 INFO [streaming_decode.py:789] 
+For test-other, WER of different settings are:
+beam_4.0_max_contexts_4_max_states_8	9.61	best for test-other
+
+2022-08-18 16:00:50,743 INFO [streaming_decode.py:963] Done!

decoding_results/streaming/fast_beam_search/wer-summary-test-clean-beam_4.0_max_contexts_4_max_states_8-epoch-35-avg-15-beam-4.0-max-contexts-4-max-states-8-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+beam_4.0_max_contexts_4_max_states_8	3.73

decoding_results/streaming/fast_beam_search/wer-summary-test-other-beam_4.0_max_contexts_4_max_states_8-epoch-35-avg-15-beam-4.0-max-contexts-4-max-states-8-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+beam_4.0_max_contexts_4_max_states_8	9.61

decoding_results/streaming/greedy_search/log-streaming-decode-2022-08-18-14-38-55

@@ -0,0 +1,42 @@
+2022-08-18 14:38:55,067 INFO [streaming_decode.py:829] Decoding started
+2022-08-18 14:38:55,067 INFO [streaming_decode.py:835] Device: cuda:0
+2022-08-18 14:38:55,071 INFO [streaming_decode.py:847] {'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'dim_feedforward': 2048, 'decoder_dim': 512, 'joiner_dim': 512, 'model_warm_step': 3000, 'env_info': {'k2-version': '1.15.1', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '350e449fe5d7b4231f77eb0add764782eed9f5d2', 'k2-git-date': 'Thu May 26 22:56:24 2022', 'lhotse-version': '1.4.0.dev+git.21b5b3f.dirty', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'lstm', 'icefall-git-sha1': '7ee3701-dirty', 'icefall-git-date': 'Sun Aug 14 17:10:39 2022', 'icefall-path': '/ceph-zw/workspace/rnn/icefall', 'k2-path': '/ceph-zw/workspace/share/k2/k2/python/k2/__init__.py', 'lhotse-path': '/ceph-jb/yaozengwei/workspace/rnnt/lhotse/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-0307195509-54c966b95f-rtpfq', 'IP address': '10.177.22.9'}, 'epoch': 35, 'iter': 0, 'avg': 15, 'use_averaged_model': True, 'exp_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'decoding_method': 'greedy_search', 'beam_size': 4, 'beam': 4, 'max_contexts': 4, 'max_states': 8, 'context_size': 2, 'max_sym_per_frame': 1, 'sampling_rate': 16000, 'num_decode_streams': 2000, 'num_encoder_layers': 12, 'encoder_dim': 512, 'rnn_hidden_size': 1024, 'aux_layer_period': 0, 'full_libri': True, 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 200.0, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'input_strategy': 'PrecomputedFeatures', 'res_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/greedy_search'), 'suffix': 'epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model', 'blank_id': 0, 'unk_id': 2, 'vocab_size': 500, 'device': device(type='cuda', index=0)}
+2022-08-18 14:38:55,071 INFO [streaming_decode.py:849] About to create model
+2022-08-18 14:38:55,403 INFO [streaming_decode.py:916] Calculating the averaged model over epoch range from 20 (excluded) to 35
+2022-08-18 14:39:00,722 INFO [streaming_decode.py:937] Number of model parameters: 84689496
+2022-08-18 14:39:00,722 INFO [asr_datamodule.py:444] About to get test-clean cuts
+2022-08-18 14:39:00,724 INFO [asr_datamodule.py:451] About to get test-other cuts
+2022-08-18 14:39:00,897 INFO [streaming_decode.py:713] Cuts processed until now is 0.
+2022-08-18 14:39:02,854 INFO [streaming_decode.py:713] Cuts processed until now is 300.
+2022-08-18 14:39:04,873 INFO [streaming_decode.py:713] Cuts processed until now is 600.
+2022-08-18 14:39:06,852 INFO [streaming_decode.py:713] Cuts processed until now is 900.
+2022-08-18 14:39:08,902 INFO [streaming_decode.py:713] Cuts processed until now is 1200.
+2022-08-18 14:39:10,971 INFO [streaming_decode.py:713] Cuts processed until now is 1500.
+2022-08-18 14:39:12,929 INFO [streaming_decode.py:713] Cuts processed until now is 1800.
+2022-08-18 14:39:20,681 INFO [streaming_decode.py:713] Cuts processed until now is 2100.
+2022-08-18 14:39:24,878 INFO [streaming_decode.py:713] Cuts processed until now is 2400.
+2022-08-18 14:39:46,266 INFO [streaming_decode.py:758] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/greedy_search/recogs-test-clean-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt
+2022-08-18 14:39:46,332 INFO [utils.py:416] [test-clean-greedy_search] %WER 3.78% [1987 / 52576, 203 ins, 181 del, 1603 sub ]
+2022-08-18 14:39:46,499 INFO [streaming_decode.py:771] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/greedy_search/errs-test-clean-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt
+2022-08-18 14:39:46,499 INFO [streaming_decode.py:788] 
+For test-clean, WER of different settings are:
+greedy_search	3.78	best for test-clean
+
+2022-08-18 14:39:46,515 INFO [streaming_decode.py:713] Cuts processed until now is 0.
+2022-08-18 14:39:48,544 INFO [streaming_decode.py:713] Cuts processed until now is 300.
+2022-08-18 14:39:50,483 INFO [streaming_decode.py:713] Cuts processed until now is 600.
+2022-08-18 14:39:52,437 INFO [streaming_decode.py:713] Cuts processed until now is 900.
+2022-08-18 14:39:54,319 INFO [streaming_decode.py:713] Cuts processed until now is 1200.
+2022-08-18 14:39:56,177 INFO [streaming_decode.py:713] Cuts processed until now is 1500.
+2022-08-18 14:39:58,059 INFO [streaming_decode.py:713] Cuts processed until now is 1800.
+2022-08-18 14:40:05,593 INFO [streaming_decode.py:713] Cuts processed until now is 2100.
+2022-08-18 14:40:09,375 INFO [streaming_decode.py:713] Cuts processed until now is 2400.
+2022-08-18 14:40:13,346 INFO [streaming_decode.py:713] Cuts processed until now is 2700.
+2022-08-18 14:40:34,167 INFO [streaming_decode.py:758] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/greedy_search/recogs-test-other-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt
+2022-08-18 14:40:34,235 INFO [utils.py:416] [test-other-greedy_search] %WER 9.79% [5123 / 52343, 512 ins, 550 del, 4061 sub ]
+2022-08-18 14:40:34,413 INFO [streaming_decode.py:771] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/greedy_search/errs-test-other-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt
+2022-08-18 14:40:34,414 INFO [streaming_decode.py:788] 
+For test-other, WER of different settings are:
+greedy_search	9.79	best for test-other
+
+2022-08-18 14:40:34,414 INFO [streaming_decode.py:962] Done!

decoding_results/streaming/greedy_search/wer-summary-test-clean-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+greedy_search	3.78

decoding_results/streaming/greedy_search/wer-summary-test-other-greedy_search-epoch-35-avg-15-context-2-max-sym-per-frame-1-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+greedy_search	9.79

decoding_results/streaming/modified_beam_search/log-streaming-decode-2022-08-18-15-19-45

@@ -0,0 +1,42 @@
+2022-08-18 15:19:45,672 INFO [streaming_decode.py:829] Decoding started
+2022-08-18 15:19:45,672 INFO [streaming_decode.py:835] Device: cuda:0
+2022-08-18 15:19:45,676 INFO [streaming_decode.py:847] {'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'dim_feedforward': 2048, 'decoder_dim': 512, 'joiner_dim': 512, 'model_warm_step': 3000, 'env_info': {'k2-version': '1.15.1', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '350e449fe5d7b4231f77eb0add764782eed9f5d2', 'k2-git-date': 'Thu May 26 22:56:24 2022', 'lhotse-version': '1.4.0.dev+git.21b5b3f.dirty', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'lstm', 'icefall-git-sha1': '7ee3701-dirty', 'icefall-git-date': 'Sun Aug 14 17:10:39 2022', 'icefall-path': '/ceph-zw/workspace/rnn/icefall', 'k2-path': '/ceph-zw/workspace/share/k2/k2/python/k2/__init__.py', 'lhotse-path': '/ceph-jb/yaozengwei/workspace/rnnt/lhotse/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-0307195509-54c966b95f-rtpfq', 'IP address': '10.177.22.9'}, 'epoch': 35, 'iter': 0, 'avg': 15, 'use_averaged_model': True, 'exp_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'decoding_method': 'modified_beam_search', 'beam_size': 4, 'beam': 4, 'max_contexts': 4, 'max_states': 8, 'context_size': 2, 'max_sym_per_frame': 1, 'sampling_rate': 16000, 'num_decode_streams': 2000, 'num_encoder_layers': 12, 'encoder_dim': 512, 'rnn_hidden_size': 1024, 'aux_layer_period': 0, 'full_libri': True, 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 200.0, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'input_strategy': 'PrecomputedFeatures', 'res_dir': PosixPath('lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/modified_beam_search'), 'suffix': 'epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model', 'blank_id': 0, 'unk_id': 2, 'vocab_size': 500, 'device': device(type='cuda', index=0)}
+2022-08-18 15:19:45,676 INFO [streaming_decode.py:849] About to create model
+2022-08-18 15:19:46,011 INFO [streaming_decode.py:916] Calculating the averaged model over epoch range from 20 (excluded) to 35
+2022-08-18 15:19:51,336 INFO [streaming_decode.py:937] Number of model parameters: 84689496
+2022-08-18 15:19:51,336 INFO [asr_datamodule.py:444] About to get test-clean cuts
+2022-08-18 15:19:51,338 INFO [asr_datamodule.py:451] About to get test-other cuts
+2022-08-18 15:19:51,503 INFO [streaming_decode.py:713] Cuts processed until now is 0.
+2022-08-18 15:19:53,501 INFO [streaming_decode.py:713] Cuts processed until now is 300.
+2022-08-18 15:19:55,555 INFO [streaming_decode.py:713] Cuts processed until now is 600.
+2022-08-18 15:19:57,568 INFO [streaming_decode.py:713] Cuts processed until now is 900.
+2022-08-18 15:19:59,634 INFO [streaming_decode.py:713] Cuts processed until now is 1200.
+2022-08-18 15:20:01,729 INFO [streaming_decode.py:713] Cuts processed until now is 1500.
+2022-08-18 15:20:03,720 INFO [streaming_decode.py:713] Cuts processed until now is 1800.
+2022-08-18 15:20:34,786 INFO [streaming_decode.py:713] Cuts processed until now is 2100.
+2022-08-18 15:20:48,128 INFO [streaming_decode.py:713] Cuts processed until now is 2400.
+2022-08-18 15:22:16,727 INFO [streaming_decode.py:758] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/modified_beam_search/recogs-test-clean-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt
+2022-08-18 15:22:16,844 INFO [utils.py:416] [test-clean-beam_size_4] %WER 3.65% [1917 / 52576, 207 ins, 152 del, 1558 sub ]
+2022-08-18 15:22:17,014 INFO [streaming_decode.py:771] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/modified_beam_search/errs-test-clean-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt
+2022-08-18 15:22:17,014 INFO [streaming_decode.py:788] 
+For test-clean, WER of different settings are:
+beam_size_4	3.65	best for test-clean
+
+2022-08-18 15:22:17,032 INFO [streaming_decode.py:713] Cuts processed until now is 0.
+2022-08-18 15:22:19,063 INFO [streaming_decode.py:713] Cuts processed until now is 300.
+2022-08-18 15:22:21,035 INFO [streaming_decode.py:713] Cuts processed until now is 600.
+2022-08-18 15:22:23,029 INFO [streaming_decode.py:713] Cuts processed until now is 900.
+2022-08-18 15:22:24,950 INFO [streaming_decode.py:713] Cuts processed until now is 1200.
+2022-08-18 15:22:26,846 INFO [streaming_decode.py:713] Cuts processed until now is 1500.
+2022-08-18 15:22:28,762 INFO [streaming_decode.py:713] Cuts processed until now is 1800.
+2022-08-18 15:22:58,952 INFO [streaming_decode.py:713] Cuts processed until now is 2100.
+2022-08-18 15:23:10,422 INFO [streaming_decode.py:713] Cuts processed until now is 2400.
+2022-08-18 15:23:22,486 INFO [streaming_decode.py:713] Cuts processed until now is 2700.
+2022-08-18 15:24:41,878 INFO [streaming_decode.py:758] The transcripts are stored in lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/modified_beam_search/recogs-test-other-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt
+2022-08-18 15:24:41,947 INFO [utils.py:416] [test-other-beam_size_4] %WER 9.51% [4976 / 52343, 526 ins, 476 del, 3974 sub ]
+2022-08-18 15:24:42,127 INFO [streaming_decode.py:771] Wrote detailed error stats to lstm_transducer_stateless/exp-full-lr-epochs-10-layer-12-rnn-hidden-1024-no-pad/streaming/modified_beam_search/errs-test-other-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt
+2022-08-18 15:24:42,128 INFO [streaming_decode.py:788] 
+For test-other, WER of different settings are:
+beam_size_4	9.51	best for test-other
+
+2022-08-18 15:24:42,128 INFO [streaming_decode.py:962] Done!

decoding_results/streaming/modified_beam_search/wer-summary-test-clean-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+beam_size_4	3.65

decoding_results/streaming/modified_beam_search/wer-summary-test-other-beam_size_4-epoch-35-avg-15-modified_beam_search-beam-size-4-use-averaged-model.txt

@@ -0,0 +1,2 @@
+settings	WER
+beam_size_4	9.51

exp/decode.sh

@@ -0,0 +1,14 @@
+export CUDA_VISIBLE_DEVICES=0
+
+./lstm_transducer_stateless/decode.py \
+  --epoch 35 \
+  --avg 15 \
+  --exp-dir ./lstm_transducer_stateless/exp \
+  --max-duration 600 \
+  --num-encoder-layers 12 \
+  --encoder-dim 512 \
+  --rnn-hidden-size 1024 \
+  --decoding-method greedy_search \
+  --use-averaged-model True
+
+