Hugging Face's logo

iliasslasri
/
robust_speech_quantizer

Automatic Speech Recognition
TensorBoard
Joblib
English
speech
discrete-units
quantization
hubert
dinosr
spidr
Model card Files
xet
 Metrics Community
Iliass Lasri commited on
Commit
f76cdb7
·
1 Parent(s): 7af5906

add spidR and DinoSR checkpoints

Browse files
Files changed (32) hide show
  1. DinoSR_original/{config.yaml → .hydra/config.yaml} +2 -2
  2. DinoSR_original/.hydra/hydra.yaml +160 -0
  3. DinoSR_original/.hydra/overrides.yaml +6 -0
  4. DinoSR_original/round_0/E1_best.pt +1 -1
  5. DinoSR_original/round_0/E1_last.pt +1 -1
  6. DinoSR_original/round_0/tensorboard/events.out.tfevents.1774341402.node20.1724484.0 +0 -3
  7. DinoSR_original/round_0/tensorboard/events.out.tfevents.1774378579.node47.1470396.0 +2 -2
  8. DinoSR_original/train.log +264 -0
  9. DinoSR_reproduced/{config.yaml → .hydra/config.yaml} +0 -0
  10. DinoSR_reproduced/.hydra/hydra.yaml +161 -0
  11. DinoSR_reproduced/.hydra/overrides.yaml +7 -0
  12. DinoSR_reproduced/round_0/E1_best.pt +1 -1
  13. DinoSR_reproduced/round_0/E1_last.pt +1 -1
  14. DinoSR_reproduced/round_0/tensorboard/events.out.tfevents.1774341408.node21.246280.0 +0 -3
  15. DinoSR_reproduced/round_0/tensorboard/events.out.tfevents.1774383206.node20.1793573.0 +2 -2
  16. DinoSR_reproduced/train.log +188 -0
  17. SpidR/256/{config.yaml → .hydra/config.yaml} +2 -2
  18. SpidR/256/.hydra/hydra.yaml +160 -0
  19. SpidR/256/.hydra/overrides.yaml +6 -0
  20. SpidR/256/round_0/E1_best.pt +1 -1
  21. SpidR/256/round_0/E1_last.pt +1 -1
  22. SpidR/256/round_0/tensorboard/events.out.tfevents.1774296494.node16.189390.0 +0 -3
  23. SpidR/256/round_0/tensorboard/events.out.tfevents.1774378787.node16.392394.0 +2 -2
  24. SpidR/256/train.log +174 -0
  25. SpidR/256_no_extra_augs/{config.yaml → .hydra/config.yaml} +1 -1
  26. SpidR/256_no_extra_augs/.hydra/hydra.yaml +160 -0
  27. SpidR/256_no_extra_augs/.hydra/overrides.yaml +6 -0
  28. SpidR/256_no_extra_augs/round_0/E1_best.pt +1 -1
  29. SpidR/256_no_extra_augs/round_0/E1_last.pt +1 -1
  30. SpidR/256_no_extra_augs/round_0/tensorboard/events.out.tfevents.1774345454.node47.1425577.0 +0 -3
  31. SpidR/256_no_extra_augs/round_0/tensorboard/events.out.tfevents.1774378527.node20.1786403.0 +2 -2
  32. SpidR/256_no_extra_augs/train.log +255 -0
DinoSR_original/{config.yaml → .hydra/config.yaml} RENAMED
@@ -4,7 +4,7 @@ training:
4
  learning_rate: 0.0001
5
  log_interval: 100
6
  checkpoint_dir: null
7
- resume_from: /home/infres/abrik-22/snlp_project/outputs/dinosr_base_original/2026-03-23/09-35-33/round_0/E1_last.pt
8
  n_iterative_pseudolabeling: 3
9
  lr_scheduler:
10
  _target_: torch.optim.lr_scheduler.CosineAnnealingLR
@@ -69,6 +69,6 @@ model:
69
  name: dinosr_base_original
70
  layer: 5
71
  vocab_size: 256
72
- kind_kmeans: kmeans
73
  quantizer:
74
  hidden_dim: 256
 
4
  learning_rate: 0.0001
5
  log_interval: 100
6
  checkpoint_dir: null
7
+ resume_from: null
8
  n_iterative_pseudolabeling: 3
9
  lr_scheduler:
10
  _target_: torch.optim.lr_scheduler.CosineAnnealingLR
 
69
  name: dinosr_base_original
70
  layer: 5
71
  vocab_size: 256
72
+ kind_kmeans: spidr
73
  quantizer:
74
  hidden_dim: 256
DinoSR_original/.hydra/hydra.yaml ADDED
@@ -0,0 +1,160 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ hydra:
2
+ run:
3
+ dir: outputs/${training.run_name}/${now:%Y-%m-%d}/${now:%H-%M-%S}
4
+ sweep:
5
+ dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S}
6
+ subdir: ${hydra.job.num}
7
+ launcher:
8
+ _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
9
+ sweeper:
10
+ _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
11
+ max_batch_size: null
12
+ params: null
13
+ help:
14
+ app_name: ${hydra.job.name}
15
+ header: '${hydra.help.app_name} is powered by Hydra.
16
+
17
+ '
18
+ footer: 'Powered by Hydra (https://hydra.cc)
19
+
20
+ Use --hydra-help to view Hydra specific help
21
+
22
+ '
23
+ template: '${hydra.help.header}
24
+
25
+ == Configuration groups ==
26
+
27
+ Compose your configuration from those groups (group=option)
28
+
29
+
30
+ $APP_CONFIG_GROUPS
31
+
32
+
33
+ == Config ==
34
+
35
+ Override anything in the config (foo.bar=value)
36
+
37
+
38
+ $CONFIG
39
+
40
+
41
+ ${hydra.help.footer}
42
+
43
+ '
44
+ hydra_help:
45
+ template: 'Hydra (${hydra.runtime.version})
46
+
47
+ See https://hydra.cc for more info.
48
+
49
+
50
+ == Flags ==
51
+
52
+ $FLAGS_HELP
53
+
54
+
55
+ == Configuration groups ==
56
+
57
+ Compose your configuration from those groups (For example, append hydra/job_logging=disabled
58
+ to command line)
59
+
60
+
61
+ $HYDRA_CONFIG_GROUPS
62
+
63
+
64
+ Use ''--cfg hydra'' to Show the Hydra config.
65
+
66
+ '
67
+ hydra_help: ???
68
+ hydra_logging:
69
+ version: 1
70
+ formatters:
71
+ simple:
72
+ format: '[%(asctime)s][HYDRA] %(message)s'
73
+ handlers:
74
+ console:
75
+ class: logging.StreamHandler
76
+ formatter: simple
77
+ stream: ext://sys.stdout
78
+ root:
79
+ level: INFO
80
+ handlers:
81
+ - console
82
+ loggers:
83
+ logging_example:
84
+ level: DEBUG
85
+ disable_existing_loggers: false
86
+ job_logging:
87
+ version: 1
88
+ formatters:
89
+ simple:
90
+ format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
91
+ handlers:
92
+ console:
93
+ class: logging.StreamHandler
94
+ formatter: simple
95
+ stream: ext://sys.stdout
96
+ file:
97
+ class: logging.FileHandler
98
+ formatter: simple
99
+ filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
100
+ root:
101
+ level: INFO
102
+ handlers:
103
+ - console
104
+ - file
105
+ disable_existing_loggers: false
106
+ env: {}
107
+ mode: RUN
108
+ searchpath: []
109
+ callbacks: {}
110
+ output_subdir: .hydra
111
+ overrides:
112
+ hydra:
113
+ - hydra.mode=RUN
114
+ task:
115
+ - training.run_name=dinosr_base_original
116
+ - model.name=dinosr_base_original
117
+ - model.layer=5
118
+ - model.vocab_size=256
119
+ - dataset.augmentations.max_augs=4
120
+ - dataset.augmentations.activate_extra_augs=True
121
+ job:
122
+ name: train
123
+ chdir: null
124
+ override_dirname: dataset.augmentations.activate_extra_augs=True,dataset.augmentations.max_augs=4,model.layer=5,model.name=dinosr_base_original,model.vocab_size=256,training.run_name=dinosr_base_original
125
+ id: ???
126
+ num: ???
127
+ config_name: quantization
128
+ env_set: {}
129
+ env_copy: []
130
+ config:
131
+ override_dirname:
132
+ kv_sep: '='
133
+ item_sep: ','
134
+ exclude_keys: []
135
+ runtime:
136
+ version: 1.3.2
137
+ version_base: '1.3'
138
+ cwd: /home/infres/abrik-22/snlp_project
139
+ config_sources:
140
+ - path: hydra.conf
141
+ schema: pkg
142
+ provider: hydra
143
+ - path: /home/infres/abrik-22/snlp_project/configs
144
+ schema: file
145
+ provider: main
146
+ - path: ''
147
+ schema: structured
148
+ provider: schema
149
+ output_dir: /home/infres/abrik-22/snlp_project/outputs/dinosr_base_original/2026-03-24/19-55-53
150
+ choices:
151
+ hydra/env: default
152
+ hydra/callbacks: null
153
+ hydra/job_logging: default
154
+ hydra/hydra_logging: default
155
+ hydra/hydra_help: default
156
+ hydra/help: default
157
+ hydra/sweeper: basic
158
+ hydra/launcher: basic
159
+ hydra/output: default
160
+ verbose: false
DinoSR_original/.hydra/overrides.yaml ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ - training.run_name=dinosr_base_original
2
+ - model.name=dinosr_base_original
3
+ - model.layer=5
4
+ - model.vocab_size=256
5
+ - dataset.augmentations.max_augs=4
6
+ - dataset.augmentations.activate_extra_augs=True
DinoSR_original/round_0/E1_best.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9a4fe99d8e0905ba324514494cb5d3518ec08cbc472cb380fe4367b419508ee9
3
  size 3961480
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f848ca2cf5ee21122a5c325f53e0ed0c70b8a85d43990643749b029eaa930bec
3
  size 3961480
DinoSR_original/round_0/E1_last.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ab18f3bcf53af73b4e0068450631011f8be79c3b80f9fbb54eb88d24b35e6a26
3
  size 3961480
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc324a875cb76aa5b6c56d0127e20c71eab2ab1c0b35836b9f0f9e7cdc9fe7ce
3
  size 3961480
DinoSR_original/round_0/tensorboard/events.out.tfevents.1774341402.node20.1724484.0 DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:42b81295047bef9b42ae3f165cdf91c3cedc09223d3c8259e2c8ff55d9a54bfd
3
- size 8152
 
 
 
 
DinoSR_original/round_0/tensorboard/events.out.tfevents.1774378579.node47.1470396.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7d9f29d8043f6588abe30d81edba0674d95215b016848a3663939a940707c9c2
3
- size 648070
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e42f75b06c6886becb888416ab86ad20ab902ef8db15ba424e17213bec4be97
3
+ size 674560
DinoSR_original/train.log CHANGED
@@ -295,3 +295,267 @@ model:
295
  [2026-03-25 12:11:00,105][root][INFO] - [Round 0] Epoch 19 | Batch 600 | CTC Loss: 2.7582
296
  [2026-03-25 12:16:07,504][root][INFO] - [Round 0] Epoch 19 | Batch 700 | CTC Loss: 2.2632
297
  [2026-03-25 12:21:23,430][root][INFO] - [Round 0] Epoch 19 | Batch 800 | CTC Loss: 2.6889
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
295
  [2026-03-25 12:11:00,105][root][INFO] - [Round 0] Epoch 19 | Batch 600 | CTC Loss: 2.7582
296
  [2026-03-25 12:16:07,504][root][INFO] - [Round 0] Epoch 19 | Batch 700 | CTC Loss: 2.2632
297
  [2026-03-25 12:21:23,430][root][INFO] - [Round 0] Epoch 19 | Batch 800 | CTC Loss: 2.6889
298
+ [2026-03-25 12:26:34,087][root][INFO] - --- [Round 0] Epoch 19 Complete | Avg CTC Loss: 2.5012 ---
299
+ [2026-03-25 12:26:38,973][root][INFO] - [Round 0] Epoch 20 | Batch 0 | CTC Loss: 2.2525
300
+ [2026-03-25 12:32:11,862][root][INFO] - [Round 0] Epoch 20 | Batch 100 | CTC Loss: 2.3859
301
+ [2026-03-25 12:37:05,244][root][INFO] - [Round 0] Epoch 20 | Batch 200 | CTC Loss: 2.2288
302
+ [2026-03-25 12:42:30,775][root][INFO] - [Round 0] Epoch 20 | Batch 300 | CTC Loss: 2.2854
303
+ [2026-03-25 12:48:08,470][root][INFO] - [Round 0] Epoch 20 | Batch 400 | CTC Loss: 2.6373
304
+ [2026-03-25 12:53:54,488][root][INFO] - [Round 0] Epoch 20 | Batch 500 | CTC Loss: 2.4404
305
+ [2026-03-25 12:59:44,280][root][INFO] - [Round 0] Epoch 20 | Batch 600 | CTC Loss: 2.7357
306
+ [2026-03-25 13:04:38,868][root][INFO] - [Round 0] Epoch 20 | Batch 700 | CTC Loss: 2.1244
307
+ [2026-03-25 13:09:44,444][root][INFO] - [Round 0] Epoch 20 | Batch 800 | CTC Loss: 2.5538
308
+ [2026-03-25 13:14:23,276][root][INFO] - --- [Round 0] Epoch 20 Complete | Avg CTC Loss: 2.4886 ---
309
+ [2026-03-25 13:14:23,319][root][INFO] - New best model saved (loss=2.4886)
310
+ [2026-03-25 13:14:29,360][root][INFO] - [Round 0] Epoch 21 | Batch 0 | CTC Loss: 2.1631
311
+ [2026-03-25 13:20:04,442][root][INFO] - [Round 0] Epoch 21 | Batch 100 | CTC Loss: 2.5535
312
+ [2026-03-25 13:25:24,439][root][INFO] - [Round 0] Epoch 21 | Batch 200 | CTC Loss: 2.6120
313
+ [2026-03-25 13:30:22,221][root][INFO] - [Round 0] Epoch 21 | Batch 300 | CTC Loss: 2.2689
314
+ [2026-03-25 13:35:18,684][root][INFO] - [Round 0] Epoch 21 | Batch 400 | CTC Loss: 2.6706
315
+ [2026-03-25 13:40:27,416][root][INFO] - [Round 0] Epoch 21 | Batch 500 | CTC Loss: 2.3610
316
+ [2026-03-25 13:45:34,334][root][INFO] - [Round 0] Epoch 21 | Batch 600 | CTC Loss: 2.5772
317
+ [2026-03-25 13:51:07,769][root][INFO] - [Round 0] Epoch 21 | Batch 700 | CTC Loss: 2.7072
318
+ [2026-03-25 13:56:30,826][root][INFO] - [Round 0] Epoch 21 | Batch 800 | CTC Loss: 2.3783
319
+ [2026-03-25 14:01:20,060][root][INFO] - --- [Round 0] Epoch 21 Complete | Avg CTC Loss: 2.4831 ---
320
+ [2026-03-25 14:01:20,102][root][INFO] - New best model saved (loss=2.4831)
321
+ [2026-03-25 14:01:54,351][root][INFO] - [Round 0] Epoch 22 | Batch 0 | CTC Loss: 2.4808
322
+ [2026-03-25 14:07:03,628][root][INFO] - [Round 0] Epoch 22 | Batch 100 | CTC Loss: 2.3273
323
+ [2026-03-25 14:12:17,167][root][INFO] - [Round 0] Epoch 22 | Batch 200 | CTC Loss: 2.7796
324
+ [2026-03-25 14:17:17,635][root][INFO] - [Round 0] Epoch 22 | Batch 300 | CTC Loss: 2.5179
325
+ [2026-03-25 14:22:18,300][root][INFO] - [Round 0] Epoch 22 | Batch 400 | CTC Loss: 2.4488
326
+ [2026-03-25 14:27:11,448][root][INFO] - [Round 0] Epoch 22 | Batch 500 | CTC Loss: 2.4148
327
+ [2026-03-25 14:31:55,667][root][INFO] - [Round 0] Epoch 22 | Batch 600 | CTC Loss: 2.4571
328
+ [2026-03-25 14:37:14,065][root][INFO] - [Round 0] Epoch 22 | Batch 700 | CTC Loss: 2.1404
329
+ [2026-03-25 14:42:12,976][root][INFO] - [Round 0] Epoch 22 | Batch 800 | CTC Loss: 2.2999
330
+ [2026-03-25 14:47:04,700][root][INFO] - --- [Round 0] Epoch 22 Complete | Avg CTC Loss: 2.4838 ---
331
+ [2026-03-25 14:47:19,790][root][INFO] - [Round 0] Epoch 23 | Batch 0 | CTC Loss: 2.4588
332
+ [2026-03-25 14:52:25,561][root][INFO] - [Round 0] Epoch 23 | Batch 100 | CTC Loss: 2.1586
333
+ [2026-03-25 14:57:24,996][root][INFO] - [Round 0] Epoch 23 | Batch 200 | CTC Loss: 2.4957
334
+ [2026-03-25 15:02:43,682][root][INFO] - [Round 0] Epoch 23 | Batch 300 | CTC Loss: 2.8344
335
+ [2026-03-25 15:07:35,378][root][INFO] - [Round 0] Epoch 23 | Batch 400 | CTC Loss: 2.7036
336
+ [2026-03-25 15:12:59,105][root][INFO] - [Round 0] Epoch 23 | Batch 500 | CTC Loss: 2.1907
337
+ [2026-03-25 15:18:41,848][root][INFO] - [Round 0] Epoch 23 | Batch 600 | CTC Loss: 2.8787
338
+ [2026-03-25 15:23:53,921][root][INFO] - [Round 0] Epoch 23 | Batch 700 | CTC Loss: 2.8433
339
+ [2026-03-25 15:28:54,407][root][INFO] - [Round 0] Epoch 23 | Batch 800 | CTC Loss: 2.2769
340
+ [2026-03-25 15:33:48,587][root][INFO] - --- [Round 0] Epoch 23 Complete | Avg CTC Loss: 2.4624 ---
341
+ [2026-03-25 15:33:48,630][root][INFO] - New best model saved (loss=2.4624)
342
+ [2026-03-25 15:34:08,387][root][INFO] - [Round 0] Epoch 24 | Batch 0 | CTC Loss: 2.9210
343
+ [2026-03-25 15:39:41,715][root][INFO] - [Round 0] Epoch 24 | Batch 100 | CTC Loss: 2.2812
344
+ [2026-03-25 15:44:38,190][root][INFO] - [Round 0] Epoch 24 | Batch 200 | CTC Loss: 2.5785
345
+ [2026-03-25 15:49:51,027][root][INFO] - [Round 0] Epoch 24 | Batch 300 | CTC Loss: 2.5080
346
+ [2026-03-25 15:55:02,509][root][INFO] - [Round 0] Epoch 24 | Batch 400 | CTC Loss: 2.3760
347
+ [2026-03-25 16:00:29,215][root][INFO] - [Round 0] Epoch 24 | Batch 500 | CTC Loss: 2.3259
348
+ [2026-03-25 16:05:34,605][root][INFO] - [Round 0] Epoch 24 | Batch 600 | CTC Loss: 2.4637
349
+ [2026-03-25 16:10:49,118][root][INFO] - [Round 0] Epoch 24 | Batch 700 | CTC Loss: 2.3912
350
+ [2026-03-25 16:16:03,564][root][INFO] - [Round 0] Epoch 24 | Batch 800 | CTC Loss: 2.3785
351
+ [2026-03-25 16:20:23,103][root][INFO] - --- [Round 0] Epoch 24 Complete | Avg CTC Loss: 2.4689 ---
352
+ [2026-03-25 16:20:49,901][root][INFO] - [Round 0] Epoch 25 | Batch 0 | CTC Loss: 2.3073
353
+ [2026-03-25 16:26:10,544][root][INFO] - [Round 0] Epoch 25 | Batch 100 | CTC Loss: 2.4313
354
+ [2026-03-25 16:30:33,558][root][INFO] - [Round 0] Epoch 25 | Batch 200 | CTC Loss: 3.0086
355
+ [2026-03-25 16:35:48,998][root][INFO] - [Round 0] Epoch 25 | Batch 300 | CTC Loss: 2.7105
356
+ [2026-03-25 16:41:04,634][root][INFO] - [Round 0] Epoch 25 | Batch 400 | CTC Loss: 2.2552
357
+ [2026-03-25 16:45:59,857][root][INFO] - [Round 0] Epoch 25 | Batch 500 | CTC Loss: 2.5288
358
+ [2026-03-25 16:51:23,912][root][INFO] - [Round 0] Epoch 25 | Batch 600 | CTC Loss: 2.5887
359
+ [2026-03-25 16:56:33,286][root][INFO] - [Round 0] Epoch 25 | Batch 700 | CTC Loss: 2.2844
360
+ [2026-03-25 17:01:38,001][root][INFO] - [Round 0] Epoch 25 | Batch 800 | CTC Loss: 2.1999
361
+ [2026-03-25 17:06:16,701][root][INFO] - --- [Round 0] Epoch 25 Complete | Avg CTC Loss: 2.4571 ---
362
+ [2026-03-25 17:06:16,743][root][INFO] - New best model saved (loss=2.4571)
363
+ [2026-03-25 17:06:35,161][root][INFO] - [Round 0] Epoch 26 | Batch 0 | CTC Loss: 2.9013
364
+ [2026-03-25 17:11:57,521][root][INFO] - [Round 0] Epoch 26 | Batch 100 | CTC Loss: 2.4242
365
+ [2026-03-25 17:16:53,678][root][INFO] - [Round 0] Epoch 26 | Batch 200 | CTC Loss: 2.4522
366
+ [2026-03-25 17:22:04,841][root][INFO] - [Round 0] Epoch 26 | Batch 300 | CTC Loss: 2.5616
367
+ [2026-03-25 17:27:22,907][root][INFO] - [Round 0] Epoch 26 | Batch 400 | CTC Loss: 2.3578
368
+ [2026-03-25 17:32:24,814][root][INFO] - [Round 0] Epoch 26 | Batch 500 | CTC Loss: 2.4698
369
+ [2026-03-25 17:37:29,561][root][INFO] - [Round 0] Epoch 26 | Batch 600 | CTC Loss: 2.5040
370
+ [2026-03-25 17:42:37,072][root][INFO] - [Round 0] Epoch 26 | Batch 700 | CTC Loss: 2.9282
371
+ [2026-03-25 17:47:54,465][root][INFO] - [Round 0] Epoch 26 | Batch 800 | CTC Loss: 2.4485
372
+ [2026-03-25 17:52:38,006][root][INFO] - --- [Round 0] Epoch 26 Complete | Avg CTC Loss: 2.4695 ---
373
+ [2026-03-25 17:52:44,951][root][INFO] - [Round 0] Epoch 27 | Batch 0 | CTC Loss: 2.0633
374
+ [2026-03-25 17:58:11,703][root][INFO] - [Round 0] Epoch 27 | Batch 100 | CTC Loss: 2.6759
375
+ [2026-03-25 18:03:30,487][root][INFO] - [Round 0] Epoch 27 | Batch 200 | CTC Loss: 3.1080
376
+ [2026-03-25 18:08:39,056][root][INFO] - [Round 0] Epoch 27 | Batch 300 | CTC Loss: 2.7203
377
+ [2026-03-25 18:14:16,732][root][INFO] - [Round 0] Epoch 27 | Batch 400 | CTC Loss: 2.3894
378
+ [2026-03-25 18:19:46,087][root][INFO] - [Round 0] Epoch 27 | Batch 500 | CTC Loss: 2.4088
379
+ [2026-03-25 18:25:26,401][root][INFO] - [Round 0] Epoch 27 | Batch 600 | CTC Loss: 1.9183
380
+ [2026-03-25 18:30:33,716][root][INFO] - [Round 0] Epoch 27 | Batch 700 | CTC Loss: 2.7079
381
+ [2026-03-25 18:35:32,519][root][INFO] - [Round 0] Epoch 27 | Batch 800 | CTC Loss: 2.6157
382
+ [2026-03-25 18:40:15,497][root][INFO] - --- [Round 0] Epoch 27 Complete | Avg CTC Loss: 2.4621 ---
383
+ [2026-03-25 18:40:26,315][root][INFO] - [Round 0] Epoch 28 | Batch 0 | CTC Loss: 2.4606
384
+ [2026-03-25 18:45:46,203][root][INFO] - [Round 0] Epoch 28 | Batch 100 | CTC Loss: 2.8631
385
+ [2026-03-25 18:50:48,957][root][INFO] - [Round 0] Epoch 28 | Batch 200 | CTC Loss: 2.7509
386
+ [2026-03-25 18:55:31,147][root][INFO] - [Round 0] Epoch 28 | Batch 300 | CTC Loss: 2.2493
387
+ [2026-03-25 19:00:39,324][root][INFO] - [Round 0] Epoch 28 | Batch 400 | CTC Loss: 2.7876
388
+ [2026-03-25 19:05:33,737][root][INFO] - [Round 0] Epoch 28 | Batch 500 | CTC Loss: 2.8257
389
+ [2026-03-25 19:10:21,363][root][INFO] - [Round 0] Epoch 28 | Batch 600 | CTC Loss: 2.3905
390
+ [2026-03-25 19:15:26,942][root][INFO] - [Round 0] Epoch 28 | Batch 700 | CTC Loss: 2.4709
391
+ [2026-03-25 19:20:23,076][root][INFO] - [Round 0] Epoch 28 | Batch 800 | CTC Loss: 2.7204
392
+ [2026-03-25 19:24:49,083][root][INFO] - --- [Round 0] Epoch 28 Complete | Avg CTC Loss: 2.4634 ---
393
+ [2026-03-25 19:25:09,438][root][INFO] - [Round 0] Epoch 29 | Batch 0 | CTC Loss: 2.6251
394
+ [2026-03-25 19:30:26,712][root][INFO] - [Round 0] Epoch 29 | Batch 100 | CTC Loss: 2.2590
395
+ [2026-03-25 19:35:32,447][root][INFO] - [Round 0] Epoch 29 | Batch 200 | CTC Loss: 2.2552
396
+ [2026-03-25 19:40:41,353][root][INFO] - [Round 0] Epoch 29 | Batch 300 | CTC Loss: 2.6949
397
+ [2026-03-25 19:45:42,632][root][INFO] - [Round 0] Epoch 29 | Batch 400 | CTC Loss: 2.2761
398
+ [2026-03-25 19:50:45,291][root][INFO] - [Round 0] Epoch 29 | Batch 500 | CTC Loss: 2.1996
399
+ [2026-03-25 19:56:11,880][root][INFO] - [Round 0] Epoch 29 | Batch 600 | CTC Loss: 2.1528
400
+ [2026-03-25 20:01:07,015][root][INFO] - [Round 0] Epoch 29 | Batch 700 | CTC Loss: 2.1649
401
+ [2026-03-25 20:06:05,097][root][INFO] - [Round 0] Epoch 29 | Batch 800 | CTC Loss: 2.4202
402
+ [2026-03-25 20:10:29,964][root][INFO] - --- [Round 0] Epoch 29 Complete | Avg CTC Loss: 2.4631 ---
403
+ [2026-03-25 20:10:40,805][root][INFO] - [Round 0] Epoch 30 | Batch 0 | CTC Loss: 2.2380
404
+ [2026-03-25 20:15:55,517][root][INFO] - [Round 0] Epoch 30 | Batch 100 | CTC Loss: 2.3376
405
+ [2026-03-25 20:20:57,374][root][INFO] - [Round 0] Epoch 30 | Batch 200 | CTC Loss: 2.1870
406
+ [2026-03-25 20:25:55,686][root][INFO] - [Round 0] Epoch 30 | Batch 300 | CTC Loss: 2.7936
407
+ [2026-03-25 20:31:23,039][root][INFO] - [Round 0] Epoch 30 | Batch 400 | CTC Loss: 2.5356
408
+ [2026-03-25 20:36:27,409][root][INFO] - [Round 0] Epoch 30 | Batch 500 | CTC Loss: 2.5056
409
+ [2026-03-25 20:41:42,816][root][INFO] - [Round 0] Epoch 30 | Batch 600 | CTC Loss: 2.6967
410
+ [2026-03-25 20:46:37,382][root][INFO] - [Round 0] Epoch 30 | Batch 700 | CTC Loss: 2.3304
411
+ [2026-03-25 20:51:47,934][root][INFO] - [Round 0] Epoch 30 | Batch 800 | CTC Loss: 2.6201
412
+ [2026-03-25 20:56:26,626][root][INFO] - --- [Round 0] Epoch 30 Complete | Avg CTC Loss: 2.4409 ---
413
+ [2026-03-25 20:56:26,668][root][INFO] - New best model saved (loss=2.4409)
414
+ [2026-03-25 20:56:43,769][root][INFO] - [Round 0] Epoch 31 | Batch 0 | CTC Loss: 2.4508
415
+ [2026-03-25 21:02:37,829][root][INFO] - [Round 0] Epoch 31 | Batch 100 | CTC Loss: 2.3280
416
+ [2026-03-25 21:07:26,983][root][INFO] - [Round 0] Epoch 31 | Batch 200 | CTC Loss: 2.6942
417
+ [2026-03-25 21:12:31,569][root][INFO] - [Round 0] Epoch 31 | Batch 300 | CTC Loss: 2.5406
418
+ [2026-03-25 21:18:08,709][root][INFO] - [Round 0] Epoch 31 | Batch 400 | CTC Loss: 2.7230
419
+ [2026-03-25 21:23:30,971][root][INFO] - [Round 0] Epoch 31 | Batch 500 | CTC Loss: 2.4610
420
+ [2026-03-25 21:29:25,947][root][INFO] - [Round 0] Epoch 31 | Batch 600 | CTC Loss: 2.8349
421
+ [2026-03-25 21:34:24,392][root][INFO] - [Round 0] Epoch 31 | Batch 700 | CTC Loss: 2.3488
422
+ [2026-03-25 21:39:38,993][root][INFO] - [Round 0] Epoch 31 | Batch 800 | CTC Loss: 2.3053
423
+ [2026-03-25 21:44:44,565][root][INFO] - --- [Round 0] Epoch 31 Complete | Avg CTC Loss: 2.4534 ---
424
+ [2026-03-25 21:44:53,884][root][INFO] - [Round 0] Epoch 32 | Batch 0 | CTC Loss: 2.5507
425
+ [2026-03-25 21:50:21,288][root][INFO] - [Round 0] Epoch 32 | Batch 100 | CTC Loss: 2.4487
426
+ [2026-03-25 21:55:27,752][root][INFO] - [Round 0] Epoch 32 | Batch 200 | CTC Loss: 2.3051
427
+ [2026-03-25 22:00:42,855][root][INFO] - [Round 0] Epoch 32 | Batch 300 | CTC Loss: 2.6066
428
+ [2026-03-25 22:05:53,924][root][INFO] - [Round 0] Epoch 32 | Batch 400 | CTC Loss: 2.4314
429
+ [2026-03-25 22:10:46,453][root][INFO] - [Round 0] Epoch 32 | Batch 500 | CTC Loss: 2.3572
430
+ [2026-03-25 22:15:42,446][root][INFO] - [Round 0] Epoch 32 | Batch 600 | CTC Loss: 2.3172
431
+ [2026-03-25 22:20:45,729][root][INFO] - [Round 0] Epoch 32 | Batch 700 | CTC Loss: 2.5252
432
+ [2026-03-25 22:25:50,295][root][INFO] - [Round 0] Epoch 32 | Batch 800 | CTC Loss: 2.5811
433
+ [2026-03-25 22:30:24,984][root][INFO] - --- [Round 0] Epoch 32 Complete | Avg CTC Loss: 2.4523 ---
434
+ [2026-03-25 22:30:31,521][root][INFO] - [Round 0] Epoch 33 | Batch 0 | CTC Loss: 2.7827
435
+ [2026-03-25 22:35:56,212][root][INFO] - [Round 0] Epoch 33 | Batch 100 | CTC Loss: 2.3689
436
+ [2026-03-25 22:40:36,764][root][INFO] - [Round 0] Epoch 33 | Batch 200 | CTC Loss: 2.4407
437
+ [2026-03-25 22:45:34,270][root][INFO] - [Round 0] Epoch 33 | Batch 300 | CTC Loss: 2.5494
438
+ [2026-03-25 22:51:07,369][root][INFO] - [Round 0] Epoch 33 | Batch 400 | CTC Loss: 2.2735
439
+ [2026-03-25 22:56:04,243][root][INFO] - [Round 0] Epoch 33 | Batch 500 | CTC Loss: 2.3088
440
+ [2026-03-25 23:00:49,369][root][INFO] - [Round 0] Epoch 33 | Batch 600 | CTC Loss: 2.2536
441
+ [2026-03-25 23:05:46,340][root][INFO] - [Round 0] Epoch 33 | Batch 700 | CTC Loss: 2.3202
442
+ [2026-03-25 23:11:10,580][root][INFO] - [Round 0] Epoch 33 | Batch 800 | CTC Loss: 2.4125
443
+ [2026-03-25 23:15:53,575][root][INFO] - --- [Round 0] Epoch 33 Complete | Avg CTC Loss: 2.4372 ---
444
+ [2026-03-25 23:15:53,617][root][INFO] - New best model saved (loss=2.4372)
445
+ [2026-03-25 23:16:12,391][root][INFO] - [Round 0] Epoch 34 | Batch 0 | CTC Loss: 2.5343
446
+ [2026-03-25 23:21:28,985][root][INFO] - [Round 0] Epoch 34 | Batch 100 | CTC Loss: 2.1512
447
+ [2026-03-25 23:26:24,398][root][INFO] - [Round 0] Epoch 34 | Batch 200 | CTC Loss: 2.1940
448
+ [2026-03-25 23:31:21,377][root][INFO] - [Round 0] Epoch 34 | Batch 300 | CTC Loss: 2.5071
449
+ [2026-03-25 23:36:18,446][root][INFO] - [Round 0] Epoch 34 | Batch 400 | CTC Loss: 1.9151
450
+ [2026-03-25 23:41:42,160][root][INFO] - [Round 0] Epoch 34 | Batch 500 | CTC Loss: 2.7630
451
+ [2026-03-25 23:46:47,176][root][INFO] - [Round 0] Epoch 34 | Batch 600 | CTC Loss: 2.5031
452
+ [2026-03-25 23:52:20,080][root][INFO] - [Round 0] Epoch 34 | Batch 700 | CTC Loss: 2.4460
453
+ [2026-03-25 23:57:15,953][root][INFO] - [Round 0] Epoch 34 | Batch 800 | CTC Loss: 2.5298
454
+ [2026-03-26 00:02:03,899][root][INFO] - --- [Round 0] Epoch 34 Complete | Avg CTC Loss: 2.4466 ---
455
+ [2026-03-26 00:02:20,914][root][INFO] - [Round 0] Epoch 35 | Batch 0 | CTC Loss: 2.8927
456
+ [2026-03-26 00:07:36,798][root][INFO] - [Round 0] Epoch 35 | Batch 100 | CTC Loss: 2.3319
457
+ [2026-03-26 00:12:24,166][root][INFO] - [Round 0] Epoch 35 | Batch 200 | CTC Loss: 2.7091
458
+ [2026-03-26 00:17:12,977][root][INFO] - [Round 0] Epoch 35 | Batch 300 | CTC Loss: 2.3513
459
+ [2026-03-26 00:22:11,939][root][INFO] - [Round 0] Epoch 35 | Batch 400 | CTC Loss: 2.3801
460
+ [2026-03-26 00:27:24,587][root][INFO] - [Round 0] Epoch 35 | Batch 500 | CTC Loss: 2.8462
461
+ [2026-03-26 00:32:38,512][root][INFO] - [Round 0] Epoch 35 | Batch 600 | CTC Loss: 2.1713
462
+ [2026-03-26 00:37:42,882][root][INFO] - [Round 0] Epoch 35 | Batch 700 | CTC Loss: 2.5186
463
+ [2026-03-26 00:42:29,542][root][INFO] - [Round 0] Epoch 35 | Batch 800 | CTC Loss: 2.7120
464
+ [2026-03-26 00:46:55,738][root][INFO] - --- [Round 0] Epoch 35 Complete | Avg CTC Loss: 2.4463 ---
465
+ [2026-03-26 00:47:24,103][root][INFO] - [Round 0] Epoch 36 | Batch 0 | CTC Loss: 2.4176
466
+ [2026-03-26 00:53:02,032][root][INFO] - [Round 0] Epoch 36 | Batch 100 | CTC Loss: 2.5405
467
+ [2026-03-26 00:57:49,950][root][INFO] - [Round 0] Epoch 36 | Batch 200 | CTC Loss: 2.4905
468
+ [2026-03-26 01:02:51,621][root][INFO] - [Round 0] Epoch 36 | Batch 300 | CTC Loss: 2.4027
469
+ [2026-03-26 01:07:32,195][root][INFO] - [Round 0] Epoch 36 | Batch 400 | CTC Loss: 2.8358
470
+ [2026-03-26 01:12:36,485][root][INFO] - [Round 0] Epoch 36 | Batch 500 | CTC Loss: 1.9366
471
+ [2026-03-26 01:17:36,711][root][INFO] - [Round 0] Epoch 36 | Batch 600 | CTC Loss: 2.6594
472
+ [2026-03-26 01:22:32,888][root][INFO] - [Round 0] Epoch 36 | Batch 700 | CTC Loss: 2.6703
473
+ [2026-03-26 01:27:47,764][root][INFO] - [Round 0] Epoch 36 | Batch 800 | CTC Loss: 2.4970
474
+ [2026-03-26 01:32:27,299][root][INFO] - --- [Round 0] Epoch 36 Complete | Avg CTC Loss: 2.4404 ---
475
+ [2026-03-26 01:32:49,984][root][INFO] - [Round 0] Epoch 37 | Batch 0 | CTC Loss: 2.5215
476
+ [2026-03-26 01:37:54,666][root][INFO] - [Round 0] Epoch 37 | Batch 100 | CTC Loss: 2.5731
477
+ [2026-03-26 01:42:44,601][root][INFO] - [Round 0] Epoch 37 | Batch 200 | CTC Loss: 2.6276
478
+ [2026-03-26 01:47:34,530][root][INFO] - [Round 0] Epoch 37 | Batch 300 | CTC Loss: 2.2454
479
+ [2026-03-26 01:52:31,898][root][INFO] - [Round 0] Epoch 37 | Batch 400 | CTC Loss: 2.5885
480
+ [2026-03-26 01:57:28,336][root][INFO] - [Round 0] Epoch 37 | Batch 500 | CTC Loss: 2.4668
481
+ [2026-03-26 02:02:14,205][root][INFO] - [Round 0] Epoch 37 | Batch 600 | CTC Loss: 2.2306
482
+ [2026-03-26 02:06:51,610][root][INFO] - [Round 0] Epoch 37 | Batch 700 | CTC Loss: 2.6288
483
+ [2026-03-26 02:11:54,656][root][INFO] - [Round 0] Epoch 37 | Batch 800 | CTC Loss: 2.4114
484
+ [2026-03-26 02:16:34,493][root][INFO] - --- [Round 0] Epoch 37 Complete | Avg CTC Loss: 2.4460 ---
485
+ [2026-03-26 02:16:44,597][root][INFO] - [Round 0] Epoch 38 | Batch 0 | CTC Loss: 2.4813
486
+ [2026-03-26 02:22:08,719][root][INFO] - [Round 0] Epoch 38 | Batch 100 | CTC Loss: 2.8880
487
+ [2026-03-26 02:27:03,907][root][INFO] - [Round 0] Epoch 38 | Batch 200 | CTC Loss: 2.3666
488
+ [2026-03-26 02:31:58,585][root][INFO] - [Round 0] Epoch 38 | Batch 300 | CTC Loss: 2.2965
489
+ [2026-03-26 02:36:55,132][root][INFO] - [Round 0] Epoch 38 | Batch 400 | CTC Loss: 2.6152
490
+ [2026-03-26 02:41:45,296][root][INFO] - [Round 0] Epoch 38 | Batch 500 | CTC Loss: 2.4657
491
+ [2026-03-26 02:46:39,731][root][INFO] - [Round 0] Epoch 38 | Batch 600 | CTC Loss: 2.5178
492
+ [2026-03-26 02:51:25,128][root][INFO] - [Round 0] Epoch 38 | Batch 700 | CTC Loss: 2.1783
493
+ [2026-03-26 02:56:08,434][root][INFO] - [Round 0] Epoch 38 | Batch 800 | CTC Loss: 2.3087
494
+ [2026-03-26 03:00:40,013][root][INFO] - --- [Round 0] Epoch 38 Complete | Avg CTC Loss: 2.4378 ---
495
+ [2026-03-26 03:01:03,830][root][INFO] - [Round 0] Epoch 39 | Batch 0 | CTC Loss: 2.3169
496
+ [2026-03-26 03:06:07,802][root][INFO] - [Round 0] Epoch 39 | Batch 100 | CTC Loss: 2.1723
497
+ [2026-03-26 03:10:47,867][root][INFO] - [Round 0] Epoch 39 | Batch 200 | CTC Loss: 2.7939
498
+ [2026-03-26 03:15:40,629][root][INFO] - [Round 0] Epoch 39 | Batch 300 | CTC Loss: 2.1614
499
+ [2026-03-26 03:20:27,473][root][INFO] - [Round 0] Epoch 39 | Batch 400 | CTC Loss: 2.3366
500
+ [2026-03-26 03:25:34,240][root][INFO] - [Round 0] Epoch 39 | Batch 500 | CTC Loss: 2.3546
501
+ [2026-03-26 03:30:28,819][root][INFO] - [Round 0] Epoch 39 | Batch 600 | CTC Loss: 2.2486
502
+ [2026-03-26 03:35:31,120][root][INFO] - [Round 0] Epoch 39 | Batch 700 | CTC Loss: 2.2145
503
+ [2026-03-26 03:40:29,523][root][INFO] - [Round 0] Epoch 39 | Batch 800 | CTC Loss: 2.2832
504
+ [2026-03-26 03:44:55,180][root][INFO] - --- [Round 0] Epoch 39 Complete | Avg CTC Loss: 2.4273 ---
505
+ [2026-03-26 03:44:55,222][root][INFO] - New best model saved (loss=2.4273)
506
+ [2026-03-26 03:45:00,007][root][INFO] - [Round 0] Epoch 40 | Batch 0 | CTC Loss: 2.2725
507
+ [2026-03-26 03:50:46,562][root][INFO] - [Round 0] Epoch 40 | Batch 100 | CTC Loss: 2.6368
508
+ [2026-03-26 03:55:47,220][root][INFO] - [Round 0] Epoch 40 | Batch 200 | CTC Loss: 2.7335
509
+ [2026-03-26 04:00:44,679][root][INFO] - [Round 0] Epoch 40 | Batch 300 | CTC Loss: 2.3732
510
+ [2026-03-26 04:05:28,255][root][INFO] - [Round 0] Epoch 40 | Batch 400 | CTC Loss: 2.2853
511
+ [2026-03-26 04:10:28,970][root][INFO] - [Round 0] Epoch 40 | Batch 500 | CTC Loss: 2.4036
512
+ [2026-03-26 04:15:20,355][root][INFO] - [Round 0] Epoch 40 | Batch 600 | CTC Loss: 2.8281
513
+ [2026-03-26 04:20:36,122][root][INFO] - [Round 0] Epoch 40 | Batch 700 | CTC Loss: 2.4658
514
+ [2026-03-26 04:25:37,125][root][INFO] - [Round 0] Epoch 40 | Batch 800 | CTC Loss: 2.5761
515
+ [2026-03-26 04:30:14,477][root][INFO] - --- [Round 0] Epoch 40 Complete | Avg CTC Loss: 2.4350 ---
516
+ [2026-03-26 04:30:22,803][root][INFO] - [Round 0] Epoch 41 | Batch 0 | CTC Loss: 2.2098
517
+ [2026-03-26 04:35:59,703][root][INFO] - [Round 0] Epoch 41 | Batch 100 | CTC Loss: 2.7866
518
+ [2026-03-26 04:40:51,093][root][INFO] - [Round 0] Epoch 41 | Batch 200 | CTC Loss: 2.4754
519
+ [2026-03-26 04:45:36,785][root][INFO] - [Round 0] Epoch 41 | Batch 300 | CTC Loss: 2.9986
520
+ [2026-03-26 04:50:47,265][root][INFO] - [Round 0] Epoch 41 | Batch 400 | CTC Loss: 2.0195
521
+ [2026-03-26 04:55:38,588][root][INFO] - [Round 0] Epoch 41 | Batch 500 | CTC Loss: 2.3196
522
+ [2026-03-26 05:00:45,292][root][INFO] - [Round 0] Epoch 41 | Batch 600 | CTC Loss: 2.3611
523
+ [2026-03-26 05:05:36,573][root][INFO] - [Round 0] Epoch 41 | Batch 700 | CTC Loss: 2.1000
524
+ [2026-03-26 05:10:29,618][root][INFO] - [Round 0] Epoch 41 | Batch 800 | CTC Loss: 2.5980
525
+ [2026-03-26 05:14:55,227][root][INFO] - --- [Round 0] Epoch 41 Complete | Avg CTC Loss: 2.4384 ---
526
+ [2026-03-26 05:15:16,909][root][INFO] - [Round 0] Epoch 42 | Batch 0 | CTC Loss: 2.5810
527
+ [2026-03-26 05:20:29,498][root][INFO] - [Round 0] Epoch 42 | Batch 100 | CTC Loss: 2.4042
528
+ [2026-03-26 05:25:19,662][root][INFO] - [Round 0] Epoch 42 | Batch 200 | CTC Loss: 2.3381
529
+ [2026-03-26 05:30:24,340][root][INFO] - [Round 0] Epoch 42 | Batch 300 | CTC Loss: 2.6696
530
+ [2026-03-26 05:35:25,003][root][INFO] - [Round 0] Epoch 42 | Batch 400 | CTC Loss: 2.0357
531
+ [2026-03-26 05:40:49,554][root][INFO] - [Round 0] Epoch 42 | Batch 500 | CTC Loss: 2.4707
532
+ [2026-03-26 05:45:40,663][root][INFO] - [Round 0] Epoch 42 | Batch 600 | CTC Loss: 2.4414
533
+ [2026-03-26 05:50:51,260][root][INFO] - [Round 0] Epoch 42 | Batch 700 | CTC Loss: 2.2316
534
+ [2026-03-26 05:55:41,861][root][INFO] - [Round 0] Epoch 42 | Batch 800 | CTC Loss: 2.3326
535
+ [2026-03-26 05:59:59,731][root][INFO] - --- [Round 0] Epoch 42 Complete | Avg CTC Loss: 2.4332 ---
536
+ [2026-03-26 06:00:24,487][root][INFO] - [Round 0] Epoch 43 | Batch 0 | CTC Loss: 2.5474
537
+ [2026-03-26 06:05:18,521][root][INFO] - [Round 0] Epoch 43 | Batch 100 | CTC Loss: 2.7438
538
+ [2026-03-26 06:10:14,431][root][INFO] - [Round 0] Epoch 43 | Batch 200 | CTC Loss: 2.5112
539
+ [2026-03-26 06:15:11,737][root][INFO] - [Round 0] Epoch 43 | Batch 300 | CTC Loss: 2.5683
540
+ [2026-03-26 06:20:09,971][root][INFO] - [Round 0] Epoch 43 | Batch 400 | CTC Loss: 2.0980
541
+ [2026-03-26 06:25:25,777][root][INFO] - [Round 0] Epoch 43 | Batch 500 | CTC Loss: 2.4312
542
+ [2026-03-26 06:30:15,431][root][INFO] - [Round 0] Epoch 43 | Batch 600 | CTC Loss: 2.0848
543
+ [2026-03-26 06:35:24,620][root][INFO] - [Round 0] Epoch 43 | Batch 700 | CTC Loss: 2.5933
544
+ [2026-03-26 06:40:48,257][root][INFO] - [Round 0] Epoch 43 | Batch 800 | CTC Loss: 2.6403
545
+ [2026-03-26 06:45:20,589][root][INFO] - --- [Round 0] Epoch 43 Complete | Avg CTC Loss: 2.4310 ---
546
+ [2026-03-26 06:45:38,385][root][INFO] - [Round 0] Epoch 44 | Batch 0 | CTC Loss: 2.3425
547
+ [2026-03-26 06:50:52,664][root][INFO] - [Round 0] Epoch 44 | Batch 100 | CTC Loss: 2.4217
548
+ [2026-03-26 06:55:53,747][root][INFO] - [Round 0] Epoch 44 | Batch 200 | CTC Loss: 2.3033
549
+ [2026-03-26 07:01:00,584][root][INFO] - [Round 0] Epoch 44 | Batch 300 | CTC Loss: 2.3716
550
+ [2026-03-26 07:06:09,638][root][INFO] - [Round 0] Epoch 44 | Batch 400 | CTC Loss: 2.3321
551
+ [2026-03-26 07:11:21,759][root][INFO] - [Round 0] Epoch 44 | Batch 500 | CTC Loss: 2.2344
552
+ [2026-03-26 07:16:30,438][root][INFO] - [Round 0] Epoch 44 | Batch 600 | CTC Loss: 2.2932
553
+ [2026-03-26 07:22:14,105][root][INFO] - [Round 0] Epoch 44 | Batch 700 | CTC Loss: 2.9810
554
+ [2026-03-26 07:27:11,782][root][INFO] - [Round 0] Epoch 44 | Batch 800 | CTC Loss: 2.3272
555
+ [2026-03-26 07:32:01,199][root][INFO] - --- [Round 0] Epoch 44 Complete | Avg CTC Loss: 2.4082 ---
556
+ [2026-03-26 07:32:01,241][root][INFO] - New best model saved (loss=2.4082)
557
+ [2026-03-26 07:32:25,953][root][INFO] - [Round 0] Epoch 45 | Batch 0 | CTC Loss: 2.6504
558
+ [2026-03-26 07:37:40,974][root][INFO] - [Round 0] Epoch 45 | Batch 100 | CTC Loss: 2.4316
559
+ [2026-03-26 07:42:45,786][root][INFO] - [Round 0] Epoch 45 | Batch 200 | CTC Loss: 2.0856
560
+ [2026-03-26 07:47:52,468][root][INFO] - [Round 0] Epoch 45 | Batch 300 | CTC Loss: 2.7049
561
+ [2026-03-26 07:52:50,112][root][INFO] - [Round 0] Epoch 45 | Batch 400 | CTC Loss: 2.4317
DinoSR_reproduced/{config.yaml → .hydra/config.yaml} RENAMED
File without changes
DinoSR_reproduced/.hydra/hydra.yaml ADDED
@@ -0,0 +1,161 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ hydra:
2
+ run:
3
+ dir: outputs/${training.run_name}/${now:%Y-%m-%d}/${now:%H-%M-%S}
4
+ sweep:
5
+ dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S}
6
+ subdir: ${hydra.job.num}
7
+ launcher:
8
+ _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
9
+ sweeper:
10
+ _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
11
+ max_batch_size: null
12
+ params: null
13
+ help:
14
+ app_name: ${hydra.job.name}
15
+ header: '${hydra.help.app_name} is powered by Hydra.
16
+
17
+ '
18
+ footer: 'Powered by Hydra (https://hydra.cc)
19
+
20
+ Use --hydra-help to view Hydra specific help
21
+
22
+ '
23
+ template: '${hydra.help.header}
24
+
25
+ == Configuration groups ==
26
+
27
+ Compose your configuration from those groups (group=option)
28
+
29
+
30
+ $APP_CONFIG_GROUPS
31
+
32
+
33
+ == Config ==
34
+
35
+ Override anything in the config (foo.bar=value)
36
+
37
+
38
+ $CONFIG
39
+
40
+
41
+ ${hydra.help.footer}
42
+
43
+ '
44
+ hydra_help:
45
+ template: 'Hydra (${hydra.runtime.version})
46
+
47
+ See https://hydra.cc for more info.
48
+
49
+
50
+ == Flags ==
51
+
52
+ $FLAGS_HELP
53
+
54
+
55
+ == Configuration groups ==
56
+
57
+ Compose your configuration from those groups (For example, append hydra/job_logging=disabled
58
+ to command line)
59
+
60
+
61
+ $HYDRA_CONFIG_GROUPS
62
+
63
+
64
+ Use ''--cfg hydra'' to Show the Hydra config.
65
+
66
+ '
67
+ hydra_help: ???
68
+ hydra_logging:
69
+ version: 1
70
+ formatters:
71
+ simple:
72
+ format: '[%(asctime)s][HYDRA] %(message)s'
73
+ handlers:
74
+ console:
75
+ class: logging.StreamHandler
76
+ formatter: simple
77
+ stream: ext://sys.stdout
78
+ root:
79
+ level: INFO
80
+ handlers:
81
+ - console
82
+ loggers:
83
+ logging_example:
84
+ level: DEBUG
85
+ disable_existing_loggers: false
86
+ job_logging:
87
+ version: 1
88
+ formatters:
89
+ simple:
90
+ format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
91
+ handlers:
92
+ console:
93
+ class: logging.StreamHandler
94
+ formatter: simple
95
+ stream: ext://sys.stdout
96
+ file:
97
+ class: logging.FileHandler
98
+ formatter: simple
99
+ filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
100
+ root:
101
+ level: INFO
102
+ handlers:
103
+ - console
104
+ - file
105
+ disable_existing_loggers: false
106
+ env: {}
107
+ mode: RUN
108
+ searchpath: []
109
+ callbacks: {}
110
+ output_subdir: .hydra
111
+ overrides:
112
+ hydra:
113
+ - hydra.mode=RUN
114
+ task:
115
+ - training.run_name=dinosr_base_reproduced
116
+ - model.name=dinosr_base_reproduced
117
+ - model.layer=5
118
+ - model.vocab_size=256
119
+ - dataset.augmentations.max_augs=4
120
+ - dataset.augmentations.activate_extra_augs=True
121
+ - training.resume_from=/home/infres/abrik-22/snlp_project/outputs/dinosr_base_reproduced/2026-03-23/09-34-30/round_0/E1_last.pt
122
+ job:
123
+ name: train
124
+ chdir: null
125
+ override_dirname: dataset.augmentations.activate_extra_augs=True,dataset.augmentations.max_augs=4,model.layer=5,model.name=dinosr_base_reproduced,model.vocab_size=256,training.resume_from=/home/infres/abrik-22/snlp_project/outputs/dinosr_base_reproduced/2026-03-23/09-34-30/round_0/E1_last.pt,training.run_name=dinosr_base_reproduced
126
+ id: ???
127
+ num: ???
128
+ config_name: quantization
129
+ env_set: {}
130
+ env_copy: []
131
+ config:
132
+ override_dirname:
133
+ kv_sep: '='
134
+ item_sep: ','
135
+ exclude_keys: []
136
+ runtime:
137
+ version: 1.3.2
138
+ version_base: '1.3'
139
+ cwd: /home/infres/abrik-22/snlp_project
140
+ config_sources:
141
+ - path: hydra.conf
142
+ schema: pkg
143
+ provider: hydra
144
+ - path: /home/infres/abrik-22/snlp_project/configs
145
+ schema: file
146
+ provider: main
147
+ - path: ''
148
+ schema: structured
149
+ provider: schema
150
+ output_dir: /home/infres/abrik-22/snlp_project/outputs/dinosr_base_reproduced/2026-03-24/21-13-09
151
+ choices:
152
+ hydra/env: default
153
+ hydra/callbacks: null
154
+ hydra/job_logging: default
155
+ hydra/hydra_logging: default
156
+ hydra/hydra_help: default
157
+ hydra/help: default
158
+ hydra/sweeper: basic
159
+ hydra/launcher: basic
160
+ hydra/output: default
161
+ verbose: false
DinoSR_reproduced/.hydra/overrides.yaml ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ - training.run_name=dinosr_base_reproduced
2
+ - model.name=dinosr_base_reproduced
3
+ - model.layer=5
4
+ - model.vocab_size=256
5
+ - dataset.augmentations.max_augs=4
6
+ - dataset.augmentations.activate_extra_augs=True
7
+ - training.resume_from=/home/infres/abrik-22/snlp_project/outputs/dinosr_base_reproduced/2026-03-23/09-34-30/round_0/E1_last.pt
DinoSR_reproduced/round_0/E1_best.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4f3e2d36ad5343bfede7820524a4f6b140295c73e1de06a8ceb7d14619392123
3
  size 3961480
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:686b31db381ff796445d327cd988a90c2ac1f897f69cf64ae0953b7746e457ca
3
  size 3961480
DinoSR_reproduced/round_0/E1_last.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2137a99115b5d2f13acda51e5be2c20bb3bc331dfceac3ebc28887ce36e8e74b
3
  size 3961480
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7299a691d5fefeee2135506aebf27c687eace545da954ee01ce97926371b76db
3
  size 3961480
DinoSR_reproduced/round_0/tensorboard/events.out.tfevents.1774341408.node21.246280.0 DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:eb9d9ed16084d176fd4fcecafb3608b5f4970f19faacefff99067021db6528bc
3
- size 7712
 
 
 
 
DinoSR_reproduced/round_0/tensorboard/events.out.tfevents.1774383206.node20.1793573.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:aa2fcb57083475b0a3c9abb0fa2f5ab10c15ed7fd055dcbd23f5ed44e17a1637
3
- size 644698
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1dfc99fcf6e4d198738ef32f6b05e9342d679088e1ef081a08032fa8bfb68d7
3
+ size 663214
DinoSR_reproduced/train.log CHANGED
@@ -200,3 +200,191 @@ model:
200
  [2026-03-25 12:12:58,350][root][INFO] - New best model saved (loss=2.3548)
201
  [2026-03-25 12:13:14,111][root][INFO] - [Round 0] Epoch 11 | Batch 0 | CTC Loss: 2.5230
202
  [2026-03-25 12:21:14,519][root][INFO] - [Round 0] Epoch 11 | Batch 100 | CTC Loss: 2.7838
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
200
  [2026-03-25 12:12:58,350][root][INFO] - New best model saved (loss=2.3548)
201
  [2026-03-25 12:13:14,111][root][INFO] - [Round 0] Epoch 11 | Batch 0 | CTC Loss: 2.5230
202
  [2026-03-25 12:21:14,519][root][INFO] - [Round 0] Epoch 11 | Batch 100 | CTC Loss: 2.7838
203
+ [2026-03-25 12:28:38,868][root][INFO] - [Round 0] Epoch 11 | Batch 200 | CTC Loss: 2.5658
204
+ [2026-03-25 12:35:35,036][root][INFO] - [Round 0] Epoch 11 | Batch 300 | CTC Loss: 2.3451
205
+ [2026-03-25 12:42:34,784][root][INFO] - [Round 0] Epoch 11 | Batch 400 | CTC Loss: 2.4443
206
+ [2026-03-25 12:49:21,193][root][INFO] - [Round 0] Epoch 11 | Batch 500 | CTC Loss: 2.2431
207
+ [2026-03-25 12:55:58,663][root][INFO] - [Round 0] Epoch 11 | Batch 600 | CTC Loss: 2.2264
208
+ [2026-03-25 13:03:11,373][root][INFO] - [Round 0] Epoch 11 | Batch 700 | CTC Loss: 2.4095
209
+ [2026-03-25 13:10:16,900][root][INFO] - [Round 0] Epoch 11 | Batch 800 | CTC Loss: 2.7642
210
+ [2026-03-25 13:17:17,508][root][INFO] - --- [Round 0] Epoch 11 Complete | Avg CTC Loss: 2.3386 ---
211
+ [2026-03-25 13:17:17,525][root][INFO] - New best model saved (loss=2.3386)
212
+ [2026-03-25 13:17:31,063][root][INFO] - [Round 0] Epoch 12 | Batch 0 | CTC Loss: 2.5121
213
+ [2026-03-25 13:26:34,491][root][INFO] - [Round 0] Epoch 12 | Batch 100 | CTC Loss: 2.1756
214
+ [2026-03-25 13:34:23,505][root][INFO] - [Round 0] Epoch 12 | Batch 200 | CTC Loss: 2.5417
215
+ [2026-03-25 13:41:55,803][root][INFO] - [Round 0] Epoch 12 | Batch 300 | CTC Loss: 2.5681
216
+ [2026-03-25 13:49:41,627][root][INFO] - [Round 0] Epoch 12 | Batch 400 | CTC Loss: 2.6153
217
+ [2026-03-25 13:58:08,730][root][INFO] - [Round 0] Epoch 12 | Batch 500 | CTC Loss: 2.2213
218
+ [2026-03-25 14:06:11,101][root][INFO] - [Round 0] Epoch 12 | Batch 600 | CTC Loss: 2.6360
219
+ [2026-03-25 14:14:27,714][root][INFO] - [Round 0] Epoch 12 | Batch 700 | CTC Loss: 3.0342
220
+ [2026-03-25 14:22:58,381][root][INFO] - [Round 0] Epoch 12 | Batch 800 | CTC Loss: 2.2672
221
+ [2026-03-25 14:30:20,960][root][INFO] - --- [Round 0] Epoch 12 Complete | Avg CTC Loss: 2.3372 ---
222
+ [2026-03-25 14:30:20,977][root][INFO] - New best model saved (loss=2.3372)
223
+ [2026-03-25 14:30:34,599][root][INFO] - [Round 0] Epoch 13 | Batch 0 | CTC Loss: 2.3531
224
+ [2026-03-25 14:38:25,818][root][INFO] - [Round 0] Epoch 13 | Batch 100 | CTC Loss: 2.0148
225
+ [2026-03-25 14:45:32,300][root][INFO] - [Round 0] Epoch 13 | Batch 200 | CTC Loss: 2.2333
226
+ [2026-03-25 14:52:47,469][root][INFO] - [Round 0] Epoch 13 | Batch 300 | CTC Loss: 2.0557
227
+ [2026-03-25 15:00:16,425][root][INFO] - [Round 0] Epoch 13 | Batch 400 | CTC Loss: 1.9320
228
+ [2026-03-25 15:08:00,785][root][INFO] - [Round 0] Epoch 13 | Batch 500 | CTC Loss: 2.1419
229
+ [2026-03-25 15:15:31,224][root][INFO] - [Round 0] Epoch 13 | Batch 600 | CTC Loss: 2.4423
230
+ [2026-03-25 15:23:30,217][root][INFO] - [Round 0] Epoch 13 | Batch 700 | CTC Loss: 2.4007
231
+ [2026-03-25 15:31:04,880][root][INFO] - [Round 0] Epoch 13 | Batch 800 | CTC Loss: 2.0369
232
+ [2026-03-25 15:37:58,353][root][INFO] - --- [Round 0] Epoch 13 Complete | Avg CTC Loss: 2.3333 ---
233
+ [2026-03-25 15:37:58,371][root][INFO] - New best model saved (loss=2.3333)
234
+ [2026-03-25 15:38:11,029][root][INFO] - [Round 0] Epoch 14 | Batch 0 | CTC Loss: 2.0526
235
+ [2026-03-25 15:49:17,041][root][INFO] - [Round 0] Epoch 14 | Batch 100 | CTC Loss: 2.2723
236
+ [2026-03-25 15:59:21,259][root][INFO] - [Round 0] Epoch 14 | Batch 200 | CTC Loss: 2.3343
237
+ [2026-03-25 16:09:29,557][root][INFO] - [Round 0] Epoch 14 | Batch 300 | CTC Loss: 2.5579
238
+ [2026-03-25 16:19:07,885][root][INFO] - [Round 0] Epoch 14 | Batch 400 | CTC Loss: 1.9091
239
+ [2026-03-25 16:28:17,120][root][INFO] - [Round 0] Epoch 14 | Batch 500 | CTC Loss: 2.5400
240
+ [2026-03-25 16:37:59,974][root][INFO] - [Round 0] Epoch 14 | Batch 600 | CTC Loss: 2.1476
241
+ [2026-03-25 16:46:53,159][root][INFO] - [Round 0] Epoch 14 | Batch 700 | CTC Loss: 2.0776
242
+ [2026-03-25 16:56:07,799][root][INFO] - [Round 0] Epoch 14 | Batch 800 | CTC Loss: 2.4781
243
+ [2026-03-25 17:04:52,076][root][INFO] - --- [Round 0] Epoch 14 Complete | Avg CTC Loss: 2.3403 ---
244
+ [2026-03-25 17:05:25,818][root][INFO] - [Round 0] Epoch 15 | Batch 0 | CTC Loss: 1.9050
245
+ [2026-03-25 17:13:23,565][root][INFO] - [Round 0] Epoch 15 | Batch 100 | CTC Loss: 2.1512
246
+ [2026-03-25 17:20:40,976][root][INFO] - [Round 0] Epoch 15 | Batch 200 | CTC Loss: 2.3013
247
+ [2026-03-25 17:27:14,368][root][INFO] - [Round 0] Epoch 15 | Batch 300 | CTC Loss: 2.4452
248
+ [2026-03-25 17:34:27,620][root][INFO] - [Round 0] Epoch 15 | Batch 400 | CTC Loss: 2.2118
249
+ [2026-03-25 17:41:34,977][root][INFO] - [Round 0] Epoch 15 | Batch 500 | CTC Loss: 2.3404
250
+ [2026-03-25 17:48:20,982][root][INFO] - [Round 0] Epoch 15 | Batch 600 | CTC Loss: 2.0609
251
+ [2026-03-25 17:55:15,923][root][INFO] - [Round 0] Epoch 15 | Batch 700 | CTC Loss: 2.4634
252
+ [2026-03-25 18:02:13,590][root][INFO] - [Round 0] Epoch 15 | Batch 800 | CTC Loss: 2.2857
253
+ [2026-03-25 18:08:42,434][root][INFO] - --- [Round 0] Epoch 15 Complete | Avg CTC Loss: 2.3164 ---
254
+ [2026-03-25 18:08:42,452][root][INFO] - New best model saved (loss=2.3164)
255
+ [2026-03-25 18:08:50,527][root][INFO] - [Round 0] Epoch 16 | Batch 0 | CTC Loss: 1.8716
256
+ [2026-03-25 18:18:42,328][root][INFO] - [Round 0] Epoch 16 | Batch 100 | CTC Loss: 2.0983
257
+ [2026-03-25 18:28:14,987][root][INFO] - [Round 0] Epoch 16 | Batch 200 | CTC Loss: 2.2175
258
+ [2026-03-25 18:37:46,005][root][INFO] - [Round 0] Epoch 16 | Batch 300 | CTC Loss: 2.6115
259
+ [2026-03-25 18:47:18,670][root][INFO] - [Round 0] Epoch 16 | Batch 400 | CTC Loss: 2.3311
260
+ [2026-03-25 18:56:11,813][root][INFO] - [Round 0] Epoch 16 | Batch 500 | CTC Loss: 2.5753
261
+ [2026-03-25 19:06:15,017][root][INFO] - [Round 0] Epoch 16 | Batch 600 | CTC Loss: 2.2795
262
+ [2026-03-25 19:15:20,161][root][INFO] - [Round 0] Epoch 16 | Batch 700 | CTC Loss: 2.1094
263
+ [2026-03-25 19:24:37,784][root][INFO] - [Round 0] Epoch 16 | Batch 800 | CTC Loss: 2.4100
264
+ [2026-03-25 19:33:13,930][root][INFO] - --- [Round 0] Epoch 16 Complete | Avg CTC Loss: 2.3169 ---
265
+ [2026-03-25 19:33:56,972][root][INFO] - [Round 0] Epoch 17 | Batch 0 | CTC Loss: 2.2481
266
+ [2026-03-25 19:44:23,784][root][INFO] - [Round 0] Epoch 17 | Batch 100 | CTC Loss: 2.2998
267
+ [2026-03-25 19:52:44,477][root][INFO] - [Round 0] Epoch 17 | Batch 200 | CTC Loss: 2.7342
268
+ [2026-03-25 20:00:19,198][root][INFO] - [Round 0] Epoch 17 | Batch 300 | CTC Loss: 2.2698
269
+ [2026-03-25 20:07:10,819][root][INFO] - [Round 0] Epoch 17 | Batch 400 | CTC Loss: 2.3982
270
+ [2026-03-25 20:14:27,339][root][INFO] - [Round 0] Epoch 17 | Batch 500 | CTC Loss: 2.2501
271
+ [2026-03-25 20:22:20,289][root][INFO] - [Round 0] Epoch 17 | Batch 600 | CTC Loss: 2.2645
272
+ [2026-03-25 20:30:30,826][root][INFO] - [Round 0] Epoch 17 | Batch 700 | CTC Loss: 2.4723
273
+ [2026-03-25 20:37:33,655][root][INFO] - [Round 0] Epoch 17 | Batch 800 | CTC Loss: 2.0863
274
+ [2026-03-25 20:45:03,670][root][INFO] - --- [Round 0] Epoch 17 Complete | Avg CTC Loss: 2.3133 ---
275
+ [2026-03-25 20:45:03,689][root][INFO] - New best model saved (loss=2.3133)
276
+ [2026-03-25 20:45:19,268][root][INFO] - [Round 0] Epoch 18 | Batch 0 | CTC Loss: 2.9406
277
+ [2026-03-25 20:53:35,109][root][INFO] - [Round 0] Epoch 18 | Batch 100 | CTC Loss: 2.2487
278
+ [2026-03-25 21:00:40,248][root][INFO] - [Round 0] Epoch 18 | Batch 200 | CTC Loss: 2.3895
279
+ [2026-03-25 21:07:35,111][root][INFO] - [Round 0] Epoch 18 | Batch 300 | CTC Loss: 2.2659
280
+ [2026-03-25 21:14:49,368][root][INFO] - [Round 0] Epoch 18 | Batch 400 | CTC Loss: 2.5412
281
+ [2026-03-25 21:21:37,261][root][INFO] - [Round 0] Epoch 18 | Batch 500 | CTC Loss: 2.2152
282
+ [2026-03-25 21:28:22,743][root][INFO] - [Round 0] Epoch 18 | Batch 600 | CTC Loss: 2.1733
283
+ [2026-03-25 21:35:41,866][root][INFO] - [Round 0] Epoch 18 | Batch 700 | CTC Loss: 2.1822
284
+ [2026-03-25 21:42:34,013][root][INFO] - [Round 0] Epoch 18 | Batch 800 | CTC Loss: 2.1725
285
+ [2026-03-25 21:48:54,206][root][INFO] - --- [Round 0] Epoch 18 Complete | Avg CTC Loss: 2.3167 ---
286
+ [2026-03-25 21:49:09,450][root][INFO] - [Round 0] Epoch 19 | Batch 0 | CTC Loss: 2.2843
287
+ [2026-03-25 21:56:27,343][root][INFO] - [Round 0] Epoch 19 | Batch 100 | CTC Loss: 2.4209
288
+ [2026-03-25 22:03:11,934][root][INFO] - [Round 0] Epoch 19 | Batch 200 | CTC Loss: 2.5386
289
+ [2026-03-25 22:10:14,601][root][INFO] - [Round 0] Epoch 19 | Batch 300 | CTC Loss: 2.4498
290
+ [2026-03-25 22:17:37,694][root][INFO] - [Round 0] Epoch 19 | Batch 400 | CTC Loss: 2.0001
291
+ [2026-03-25 22:24:57,900][root][INFO] - [Round 0] Epoch 19 | Batch 500 | CTC Loss: 2.0217
292
+ [2026-03-25 22:31:36,323][root][INFO] - [Round 0] Epoch 19 | Batch 600 | CTC Loss: 1.9301
293
+ [2026-03-25 22:39:41,713][root][INFO] - [Round 0] Epoch 19 | Batch 700 | CTC Loss: 2.3107
294
+ [2026-03-25 22:49:41,499][root][INFO] - [Round 0] Epoch 19 | Batch 800 | CTC Loss: 2.1291
295
+ [2026-03-25 22:56:35,396][root][INFO] - --- [Round 0] Epoch 19 Complete | Avg CTC Loss: 2.2977 ---
296
+ [2026-03-25 22:56:35,413][root][INFO] - New best model saved (loss=2.2977)
297
+ [2026-03-25 22:56:59,952][root][INFO] - [Round 0] Epoch 20 | Batch 0 | CTC Loss: 2.4359
298
+ [2026-03-25 23:04:32,767][root][INFO] - [Round 0] Epoch 20 | Batch 100 | CTC Loss: 2.2858
299
+ [2026-03-25 23:11:36,806][root][INFO] - [Round 0] Epoch 20 | Batch 200 | CTC Loss: 2.4390
300
+ [2026-03-25 23:18:33,386][root][INFO] - [Round 0] Epoch 20 | Batch 300 | CTC Loss: 2.5137
301
+ [2026-03-25 23:26:36,093][root][INFO] - [Round 0] Epoch 20 | Batch 400 | CTC Loss: 2.3218
302
+ [2026-03-25 23:35:02,288][root][INFO] - [Round 0] Epoch 20 | Batch 500 | CTC Loss: 2.2954
303
+ [2026-03-25 23:42:53,853][root][INFO] - [Round 0] Epoch 20 | Batch 600 | CTC Loss: 2.1926
304
+ [2026-03-25 23:50:29,557][root][INFO] - [Round 0] Epoch 20 | Batch 700 | CTC Loss: 2.6059
305
+ [2026-03-25 23:58:00,878][root][INFO] - [Round 0] Epoch 20 | Batch 800 | CTC Loss: 2.4498
306
+ [2026-03-26 00:04:48,992][root][INFO] - --- [Round 0] Epoch 20 Complete | Avg CTC Loss: 2.2982 ---
307
+ [2026-03-26 00:04:59,488][root][INFO] - [Round 0] Epoch 21 | Batch 0 | CTC Loss: 2.4059
308
+ [2026-03-26 00:12:16,723][root][INFO] - [Round 0] Epoch 21 | Batch 100 | CTC Loss: 2.4177
309
+ [2026-03-26 00:18:45,504][root][INFO] - [Round 0] Epoch 21 | Batch 200 | CTC Loss: 2.0129
310
+ [2026-03-26 00:27:38,578][root][INFO] - [Round 0] Epoch 21 | Batch 300 | CTC Loss: 2.0548
311
+ [2026-03-26 00:36:29,881][root][INFO] - [Round 0] Epoch 21 | Batch 400 | CTC Loss: 2.1417
312
+ [2026-03-26 00:43:59,791][root][INFO] - [Round 0] Epoch 21 | Batch 500 | CTC Loss: 2.1233
313
+ [2026-03-26 00:51:16,897][root][INFO] - [Round 0] Epoch 21 | Batch 600 | CTC Loss: 2.1980
314
+ [2026-03-26 00:57:58,815][root][INFO] - [Round 0] Epoch 21 | Batch 700 | CTC Loss: 2.1509
315
+ [2026-03-26 01:06:36,810][root][INFO] - [Round 0] Epoch 21 | Batch 800 | CTC Loss: 2.4840
316
+ [2026-03-26 01:13:17,430][root][INFO] - --- [Round 0] Epoch 21 Complete | Avg CTC Loss: 2.2922 ---
317
+ [2026-03-26 01:13:17,447][root][INFO] - New best model saved (loss=2.2922)
318
+ [2026-03-26 01:13:28,004][root][INFO] - [Round 0] Epoch 22 | Batch 0 | CTC Loss: 2.5620
319
+ [2026-03-26 01:27:57,368][root][INFO] - [Round 0] Epoch 22 | Batch 100 | CTC Loss: 2.5542
320
+ [2026-03-26 01:38:13,609][root][INFO] - [Round 0] Epoch 22 | Batch 200 | CTC Loss: 2.3064
321
+ [2026-03-26 01:48:42,536][root][INFO] - [Round 0] Epoch 22 | Batch 300 | CTC Loss: 2.3977
322
+ [2026-03-26 01:58:35,989][root][INFO] - [Round 0] Epoch 22 | Batch 400 | CTC Loss: 1.9625
323
+ [2026-03-26 02:05:26,422][root][INFO] - [Round 0] Epoch 22 | Batch 500 | CTC Loss: 1.9687
324
+ [2026-03-26 02:12:06,662][root][INFO] - [Round 0] Epoch 22 | Batch 600 | CTC Loss: 2.3304
325
+ [2026-03-26 02:22:01,030][root][INFO] - [Round 0] Epoch 22 | Batch 700 | CTC Loss: 2.2369
326
+ [2026-03-26 02:33:38,790][root][INFO] - [Round 0] Epoch 22 | Batch 800 | CTC Loss: 2.0651
327
+ [2026-03-26 02:43:03,868][root][INFO] - --- [Round 0] Epoch 22 Complete | Avg CTC Loss: 2.3065 ---
328
+ [2026-03-26 02:43:10,568][root][INFO] - [Round 0] Epoch 23 | Batch 0 | CTC Loss: 2.3154
329
+ [2026-03-26 02:52:06,755][root][INFO] - [Round 0] Epoch 23 | Batch 100 | CTC Loss: 2.0819
330
+ [2026-03-26 02:59:17,900][root][INFO] - [Round 0] Epoch 23 | Batch 200 | CTC Loss: 2.1106
331
+ [2026-03-26 03:06:05,698][root][INFO] - [Round 0] Epoch 23 | Batch 300 | CTC Loss: 2.1881
332
+ [2026-03-26 03:13:39,686][root][INFO] - [Round 0] Epoch 23 | Batch 400 | CTC Loss: 2.6659
333
+ [2026-03-26 03:20:42,365][root][INFO] - [Round 0] Epoch 23 | Batch 500 | CTC Loss: 2.5083
334
+ [2026-03-26 03:27:57,382][root][INFO] - [Round 0] Epoch 23 | Batch 600 | CTC Loss: 2.4104
335
+ [2026-03-26 03:35:04,908][root][INFO] - [Round 0] Epoch 23 | Batch 700 | CTC Loss: 2.3092
336
+ [2026-03-26 03:41:56,607][root][INFO] - [Round 0] Epoch 23 | Batch 800 | CTC Loss: 2.2364
337
+ [2026-03-26 03:48:27,524][root][INFO] - --- [Round 0] Epoch 23 Complete | Avg CTC Loss: 2.2964 ---
338
+ [2026-03-26 03:48:33,043][root][INFO] - [Round 0] Epoch 24 | Batch 0 | CTC Loss: 2.3481
339
+ [2026-03-26 03:56:12,490][root][INFO] - [Round 0] Epoch 24 | Batch 100 | CTC Loss: 2.0853
340
+ [2026-03-26 04:05:19,149][root][INFO] - [Round 0] Epoch 24 | Batch 200 | CTC Loss: 2.3058
341
+ [2026-03-26 04:13:42,789][root][INFO] - [Round 0] Epoch 24 | Batch 300 | CTC Loss: 2.4361
342
+ [2026-03-26 04:21:10,796][root][INFO] - [Round 0] Epoch 24 | Batch 400 | CTC Loss: 2.4720
343
+ [2026-03-26 04:28:32,059][root][INFO] - [Round 0] Epoch 24 | Batch 500 | CTC Loss: 2.2499
344
+ [2026-03-26 04:36:14,540][root][INFO] - [Round 0] Epoch 24 | Batch 600 | CTC Loss: 2.6874
345
+ [2026-03-26 04:45:40,023][root][INFO] - [Round 0] Epoch 24 | Batch 700 | CTC Loss: 2.1193
346
+ [2026-03-26 04:54:07,092][root][INFO] - [Round 0] Epoch 24 | Batch 800 | CTC Loss: 2.4916
347
+ [2026-03-26 05:01:28,853][root][INFO] - --- [Round 0] Epoch 24 Complete | Avg CTC Loss: 2.2909 ---
348
+ [2026-03-26 05:01:28,870][root][INFO] - New best model saved (loss=2.2909)
349
+ [2026-03-26 05:02:06,700][root][INFO] - [Round 0] Epoch 25 | Batch 0 | CTC Loss: 2.1110
350
+ [2026-03-26 05:09:58,361][root][INFO] - [Round 0] Epoch 25 | Batch 100 | CTC Loss: 2.2938
351
+ [2026-03-26 05:18:21,479][root][INFO] - [Round 0] Epoch 25 | Batch 200 | CTC Loss: 2.0954
352
+ [2026-03-26 05:25:08,494][root][INFO] - [Round 0] Epoch 25 | Batch 300 | CTC Loss: 2.1520
353
+ [2026-03-26 05:32:11,579][root][INFO] - [Round 0] Epoch 25 | Batch 400 | CTC Loss: 2.4470
354
+ [2026-03-26 05:41:26,696][root][INFO] - [Round 0] Epoch 25 | Batch 500 | CTC Loss: 2.1030
355
+ [2026-03-26 05:48:47,349][root][INFO] - [Round 0] Epoch 25 | Batch 600 | CTC Loss: 2.4434
356
+ [2026-03-26 05:56:21,891][root][INFO] - [Round 0] Epoch 25 | Batch 700 | CTC Loss: 2.0787
357
+ [2026-03-26 06:03:38,394][root][INFO] - [Round 0] Epoch 25 | Batch 800 | CTC Loss: 2.3191
358
+ [2026-03-26 06:12:33,943][root][INFO] - --- [Round 0] Epoch 25 Complete | Avg CTC Loss: 2.2773 ---
359
+ [2026-03-26 06:12:33,963][root][INFO] - New best model saved (loss=2.2773)
360
+ [2026-03-26 06:12:52,910][root][INFO] - [Round 0] Epoch 26 | Batch 0 | CTC Loss: 2.2689
361
+ [2026-03-26 06:20:49,136][root][INFO] - [Round 0] Epoch 26 | Batch 100 | CTC Loss: 2.5048
362
+ [2026-03-26 06:27:30,852][root][INFO] - [Round 0] Epoch 26 | Batch 200 | CTC Loss: 2.4881
363
+ [2026-03-26 06:34:34,763][root][INFO] - [Round 0] Epoch 26 | Batch 300 | CTC Loss: 2.4310
364
+ [2026-03-26 06:41:10,829][root][INFO] - [Round 0] Epoch 26 | Batch 400 | CTC Loss: 2.3776
365
+ [2026-03-26 06:49:12,972][root][INFO] - [Round 0] Epoch 26 | Batch 500 | CTC Loss: 2.6883
366
+ [2026-03-26 06:57:04,271][root][INFO] - [Round 0] Epoch 26 | Batch 600 | CTC Loss: 2.3896
367
+ [2026-03-26 07:06:09,876][root][INFO] - [Round 0] Epoch 26 | Batch 700 | CTC Loss: 1.6159
368
+ [2026-03-26 07:14:21,131][root][INFO] - [Round 0] Epoch 26 | Batch 800 | CTC Loss: 2.3390
369
+ [2026-03-26 07:21:44,025][root][INFO] - --- [Round 0] Epoch 26 Complete | Avg CTC Loss: 2.2877 ---
370
+ [2026-03-26 07:22:15,822][root][INFO] - [Round 0] Epoch 27 | Batch 0 | CTC Loss: 2.1685
371
+ [2026-03-26 07:29:51,754][root][INFO] - [Round 0] Epoch 27 | Batch 100 | CTC Loss: 2.1251
372
+ [2026-03-26 07:39:58,679][root][INFO] - [Round 0] Epoch 27 | Batch 200 | CTC Loss: 1.8569
373
+ [2026-03-26 07:48:01,288][root][INFO] - [Round 0] Epoch 27 | Batch 300 | CTC Loss: 2.1196
374
+ [2026-03-26 07:55:18,777][root][INFO] - [Round 0] Epoch 27 | Batch 400 | CTC Loss: 2.0875
375
+ [2026-03-26 08:00:38,423][root][INFO] - [Round 0] Epoch 27 | Batch 500 | CTC Loss: 2.2936
376
+ [2026-03-26 08:05:55,214][root][INFO] - [Round 0] Epoch 27 | Batch 600 | CTC Loss: 1.8432
377
+ [2026-03-26 08:11:05,434][root][INFO] - [Round 0] Epoch 27 | Batch 700 | CTC Loss: 2.6728
378
+ [2026-03-26 08:16:12,383][root][INFO] - [Round 0] Epoch 27 | Batch 800 | CTC Loss: 2.5145
379
+ [2026-03-26 08:20:44,407][root][INFO] - --- [Round 0] Epoch 27 Complete | Avg CTC Loss: 2.2926 ---
380
+ [2026-03-26 08:21:09,295][root][INFO] - [Round 0] Epoch 28 | Batch 0 | CTC Loss: 2.1714
381
+ [2026-03-26 08:26:19,345][root][INFO] - [Round 0] Epoch 28 | Batch 100 | CTC Loss: 1.9488
382
+ [2026-03-26 08:31:14,457][root][INFO] - [Round 0] Epoch 28 | Batch 200 | CTC Loss: 2.3098
383
+ [2026-03-26 08:36:33,967][root][INFO] - [Round 0] Epoch 28 | Batch 300 | CTC Loss: 2.5934
384
+ [2026-03-26 08:41:53,827][root][INFO] - [Round 0] Epoch 28 | Batch 400 | CTC Loss: 2.0017
385
+ [2026-03-26 08:47:12,559][root][INFO] - [Round 0] Epoch 28 | Batch 500 | CTC Loss: 1.9758
386
+ [2026-03-26 08:52:18,015][root][INFO] - [Round 0] Epoch 28 | Batch 600 | CTC Loss: 2.3799
387
+ [2026-03-26 08:57:58,820][root][INFO] - [Round 0] Epoch 28 | Batch 700 | CTC Loss: 2.3434
388
+ [2026-03-26 09:03:26,970][root][INFO] - [Round 0] Epoch 28 | Batch 800 | CTC Loss: 2.5496
389
+ [2026-03-26 09:08:33,869][root][INFO] - --- [Round 0] Epoch 28 Complete | Avg CTC Loss: 2.2961 ---
390
+ [2026-03-26 09:08:50,549][root][INFO] - [Round 0] Epoch 29 | Batch 0 | CTC Loss: 2.0147
SpidR/256/{config.yaml → .hydra/config.yaml} RENAMED
@@ -4,7 +4,7 @@ training:
4
  learning_rate: 0.0001
5
  log_interval: 100
6
  checkpoint_dir: null
7
- resume_from: /home/infres/abrik-22/snlp_project/outputs/256/2026-03-22/18-41-33/round_0/E1_last.pt
8
  n_iterative_pseudolabeling: 3
9
  lr_scheduler:
10
  _target_: torch.optim.lr_scheduler.CosineAnnealingLR
@@ -69,6 +69,6 @@ model:
69
  name: spidr_base
70
  layer: 6
71
  vocab_size: 256
72
- kind_kmeans: kmeans
73
  quantizer:
74
  hidden_dim: 256
 
4
  learning_rate: 0.0001
5
  log_interval: 100
6
  checkpoint_dir: null
7
+ resume_from: null
8
  n_iterative_pseudolabeling: 3
9
  lr_scheduler:
10
  _target_: torch.optim.lr_scheduler.CosineAnnealingLR
 
69
  name: spidr_base
70
  layer: 6
71
  vocab_size: 256
72
+ kind_kmeans: spidr
73
  quantizer:
74
  hidden_dim: 256
SpidR/256/.hydra/hydra.yaml ADDED
@@ -0,0 +1,160 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ hydra:
2
+ run:
3
+ dir: outputs/${training.run_name}/${now:%Y-%m-%d}/${now:%H-%M-%S}
4
+ sweep:
5
+ dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S}
6
+ subdir: ${hydra.job.num}
7
+ launcher:
8
+ _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
9
+ sweeper:
10
+ _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
11
+ max_batch_size: null
12
+ params: null
13
+ help:
14
+ app_name: ${hydra.job.name}
15
+ header: '${hydra.help.app_name} is powered by Hydra.
16
+
17
+ '
18
+ footer: 'Powered by Hydra (https://hydra.cc)
19
+
20
+ Use --hydra-help to view Hydra specific help
21
+
22
+ '
23
+ template: '${hydra.help.header}
24
+
25
+ == Configuration groups ==
26
+
27
+ Compose your configuration from those groups (group=option)
28
+
29
+
30
+ $APP_CONFIG_GROUPS
31
+
32
+
33
+ == Config ==
34
+
35
+ Override anything in the config (foo.bar=value)
36
+
37
+
38
+ $CONFIG
39
+
40
+
41
+ ${hydra.help.footer}
42
+
43
+ '
44
+ hydra_help:
45
+ template: 'Hydra (${hydra.runtime.version})
46
+
47
+ See https://hydra.cc for more info.
48
+
49
+
50
+ == Flags ==
51
+
52
+ $FLAGS_HELP
53
+
54
+
55
+ == Configuration groups ==
56
+
57
+ Compose your configuration from those groups (For example, append hydra/job_logging=disabled
58
+ to command line)
59
+
60
+
61
+ $HYDRA_CONFIG_GROUPS
62
+
63
+
64
+ Use ''--cfg hydra'' to Show the Hydra config.
65
+
66
+ '
67
+ hydra_help: ???
68
+ hydra_logging:
69
+ version: 1
70
+ formatters:
71
+ simple:
72
+ format: '[%(asctime)s][HYDRA] %(message)s'
73
+ handlers:
74
+ console:
75
+ class: logging.StreamHandler
76
+ formatter: simple
77
+ stream: ext://sys.stdout
78
+ root:
79
+ level: INFO
80
+ handlers:
81
+ - console
82
+ loggers:
83
+ logging_example:
84
+ level: DEBUG
85
+ disable_existing_loggers: false
86
+ job_logging:
87
+ version: 1
88
+ formatters:
89
+ simple:
90
+ format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
91
+ handlers:
92
+ console:
93
+ class: logging.StreamHandler
94
+ formatter: simple
95
+ stream: ext://sys.stdout
96
+ file:
97
+ class: logging.FileHandler
98
+ formatter: simple
99
+ filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
100
+ root:
101
+ level: INFO
102
+ handlers:
103
+ - console
104
+ - file
105
+ disable_existing_loggers: false
106
+ env: {}
107
+ mode: RUN
108
+ searchpath: []
109
+ callbacks: {}
110
+ output_subdir: .hydra
111
+ overrides:
112
+ hydra:
113
+ - hydra.mode=RUN
114
+ task:
115
+ - training.run_name=spidr
116
+ - model.name=spidr_base
117
+ - model.layer=6
118
+ - model.vocab_size=256
119
+ - dataset.augmentations.max_augs=4
120
+ - dataset.augmentations.activate_extra_augs=True
121
+ job:
122
+ name: train
123
+ chdir: null
124
+ override_dirname: dataset.augmentations.activate_extra_augs=True,dataset.augmentations.max_augs=4,model.layer=6,model.name=spidr_base,model.vocab_size=256,training.run_name=spidr
125
+ id: ???
126
+ num: ???
127
+ config_name: quantization
128
+ env_set: {}
129
+ env_copy: []
130
+ config:
131
+ override_dirname:
132
+ kv_sep: '='
133
+ item_sep: ','
134
+ exclude_keys: []
135
+ runtime:
136
+ version: 1.3.2
137
+ version_base: '1.3'
138
+ cwd: /home/infres/abrik-22/snlp_project
139
+ config_sources:
140
+ - path: hydra.conf
141
+ schema: pkg
142
+ provider: hydra
143
+ - path: /home/infres/abrik-22/snlp_project/configs
144
+ schema: file
145
+ provider: main
146
+ - path: ''
147
+ schema: structured
148
+ provider: schema
149
+ output_dir: /home/infres/abrik-22/snlp_project/outputs/spidr/2026-03-24/19-59-31
150
+ choices:
151
+ hydra/env: default
152
+ hydra/callbacks: null
153
+ hydra/job_logging: default
154
+ hydra/hydra_logging: default
155
+ hydra/hydra_help: default
156
+ hydra/help: default
157
+ hydra/sweeper: basic
158
+ hydra/launcher: basic
159
+ hydra/output: default
160
+ verbose: false
SpidR/256/.hydra/overrides.yaml ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ - training.run_name=spidr
2
+ - model.name=spidr_base
3
+ - model.layer=6
4
+ - model.vocab_size=256
5
+ - dataset.augmentations.max_augs=4
6
+ - dataset.augmentations.activate_extra_augs=True
SpidR/256/round_0/E1_best.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8db7f3ba4fb0c077794dfa0dfc5a65f1cf2b4b11abaf72ec7c3f648f049135df
3
  size 3961480
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:70e91981233395bb163e5f73b41d43a25cb9719cabc75b0c0eeb90b30c091841
3
  size 3961480
SpidR/256/round_0/E1_last.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3b9582b635fc7af37e95f792a67e2e0ebecbd2114e89146ec08734ed73566e65
3
  size 3961480
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4cecfbcf03efeb800012acb6b5996180dfc2ec90610e86d10e6041dbb8a2479
3
  size 3961480
SpidR/256/round_0/tensorboard/events.out.tfevents.1774296494.node16.189390.0 DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:67b5d59c76d66df40fad02b56b65224fef5b3975147de09c5568b59768cc6595
3
- size 15116
 
 
 
 
SpidR/256/round_0/tensorboard/events.out.tfevents.1774378787.node16.392394.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8bb8aec4f252f96f4eb4ef6eec826e49fa115b423d20593312e92f0c59f7cc45
3
- size 652474
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:182a9854a6d4700bcb5b5146298e95ab81fb872f699a58758d54f73cc852c00e
3
+ size 669028
SpidR/256/train.log CHANGED
@@ -206,3 +206,177 @@ model:
206
  [2026-03-25 12:03:30,483][root][INFO] - [Round 0] Epoch 11 | Batch 500 | CTC Loss: 3.7918
207
  [2026-03-25 12:13:05,443][root][INFO] - [Round 0] Epoch 11 | Batch 600 | CTC Loss: 3.3099
208
  [2026-03-25 12:22:53,001][root][INFO] - [Round 0] Epoch 11 | Batch 700 | CTC Loss: 2.8892
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
206
  [2026-03-25 12:03:30,483][root][INFO] - [Round 0] Epoch 11 | Batch 500 | CTC Loss: 3.7918
207
  [2026-03-25 12:13:05,443][root][INFO] - [Round 0] Epoch 11 | Batch 600 | CTC Loss: 3.3099
208
  [2026-03-25 12:22:53,001][root][INFO] - [Round 0] Epoch 11 | Batch 700 | CTC Loss: 2.8892
209
+ [2026-03-25 12:32:29,983][root][INFO] - [Round 0] Epoch 11 | Batch 800 | CTC Loss: 4.3561
210
+ [2026-03-25 12:40:30,329][root][INFO] - --- [Round 0] Epoch 11 Complete | Avg CTC Loss: 3.8282 ---
211
+ [2026-03-25 12:40:30,344][root][INFO] - New best model saved (loss=3.8282)
212
+ [2026-03-25 12:40:36,437][root][INFO] - [Round 0] Epoch 12 | Batch 0 | CTC Loss: 4.0538
213
+ [2026-03-25 12:48:42,704][root][INFO] - [Round 0] Epoch 12 | Batch 100 | CTC Loss: 4.1909
214
+ [2026-03-25 12:57:38,816][root][INFO] - [Round 0] Epoch 12 | Batch 200 | CTC Loss: 3.5868
215
+ [2026-03-25 13:06:00,544][root][INFO] - [Round 0] Epoch 12 | Batch 300 | CTC Loss: 3.9711
216
+ [2026-03-25 13:15:11,928][root][INFO] - [Round 0] Epoch 12 | Batch 400 | CTC Loss: 3.8308
217
+ [2026-03-25 13:23:31,329][root][INFO] - [Round 0] Epoch 12 | Batch 500 | CTC Loss: 3.6890
218
+ [2026-03-25 13:32:20,051][root][INFO] - [Round 0] Epoch 12 | Batch 600 | CTC Loss: 3.4924
219
+ [2026-03-25 13:40:49,412][root][INFO] - [Round 0] Epoch 12 | Batch 700 | CTC Loss: 3.4503
220
+ [2026-03-25 13:49:32,793][root][INFO] - [Round 0] Epoch 12 | Batch 800 | CTC Loss: 3.5051
221
+ [2026-03-25 13:57:14,481][root][INFO] - --- [Round 0] Epoch 12 Complete | Avg CTC Loss: 3.8217 ---
222
+ [2026-03-25 13:57:14,497][root][INFO] - New best model saved (loss=3.8217)
223
+ [2026-03-25 13:57:45,706][root][INFO] - [Round 0] Epoch 13 | Batch 0 | CTC Loss: 4.2638
224
+ [2026-03-25 14:06:29,437][root][INFO] - [Round 0] Epoch 13 | Batch 100 | CTC Loss: 4.2236
225
+ [2026-03-25 14:15:00,626][root][INFO] - [Round 0] Epoch 13 | Batch 200 | CTC Loss: 3.5774
226
+ [2026-03-25 14:23:31,230][root][INFO] - [Round 0] Epoch 13 | Batch 300 | CTC Loss: 3.3103
227
+ [2026-03-25 14:32:05,490][root][INFO] - [Round 0] Epoch 13 | Batch 400 | CTC Loss: 3.4096
228
+ [2026-03-25 14:40:35,612][root][INFO] - [Round 0] Epoch 13 | Batch 500 | CTC Loss: 3.9014
229
+ [2026-03-25 14:49:19,175][root][INFO] - [Round 0] Epoch 13 | Batch 600 | CTC Loss: 3.9071
230
+ [2026-03-25 14:57:29,574][root][INFO] - [Round 0] Epoch 13 | Batch 700 | CTC Loss: 3.2664
231
+ [2026-03-25 15:06:12,554][root][INFO] - [Round 0] Epoch 13 | Batch 800 | CTC Loss: 3.9340
232
+ [2026-03-25 15:13:29,107][root][INFO] - --- [Round 0] Epoch 13 Complete | Avg CTC Loss: 3.7767 ---
233
+ [2026-03-25 15:13:29,132][root][INFO] - New best model saved (loss=3.7767)
234
+ [2026-03-25 15:13:33,330][root][INFO] - [Round 0] Epoch 14 | Batch 0 | CTC Loss: 4.1724
235
+ [2026-03-25 15:22:31,191][root][INFO] - [Round 0] Epoch 14 | Batch 100 | CTC Loss: 4.0324
236
+ [2026-03-25 15:31:46,443][root][INFO] - [Round 0] Epoch 14 | Batch 200 | CTC Loss: 3.8171
237
+ [2026-03-25 15:40:44,275][root][INFO] - [Round 0] Epoch 14 | Batch 300 | CTC Loss: 3.7898
238
+ [2026-03-25 15:49:34,546][root][INFO] - [Round 0] Epoch 14 | Batch 400 | CTC Loss: 4.0469
239
+ [2026-03-25 15:58:05,871][root][INFO] - [Round 0] Epoch 14 | Batch 500 | CTC Loss: 3.7921
240
+ [2026-03-25 16:07:04,765][root][INFO] - [Round 0] Epoch 14 | Batch 600 | CTC Loss: 3.9607
241
+ [2026-03-25 16:15:46,815][root][INFO] - [Round 0] Epoch 14 | Batch 700 | CTC Loss: 3.5509
242
+ [2026-03-25 16:25:29,298][root][INFO] - [Round 0] Epoch 14 | Batch 800 | CTC Loss: 3.9956
243
+ [2026-03-25 16:33:21,568][root][INFO] - --- [Round 0] Epoch 14 Complete | Avg CTC Loss: 3.7700 ---
244
+ [2026-03-25 16:33:21,583][root][INFO] - New best model saved (loss=3.7700)
245
+ [2026-03-25 16:33:39,373][root][INFO] - [Round 0] Epoch 15 | Batch 0 | CTC Loss: 3.6319
246
+ [2026-03-25 16:42:38,956][root][INFO] - [Round 0] Epoch 15 | Batch 100 | CTC Loss: 3.9354
247
+ [2026-03-25 16:51:00,361][root][INFO] - [Round 0] Epoch 15 | Batch 200 | CTC Loss: 4.3327
248
+ [2026-03-25 16:59:57,229][root][INFO] - [Round 0] Epoch 15 | Batch 300 | CTC Loss: 4.2747
249
+ [2026-03-25 17:08:13,749][root][INFO] - [Round 0] Epoch 15 | Batch 400 | CTC Loss: 3.7257
250
+ [2026-03-25 17:17:13,593][root][INFO] - [Round 0] Epoch 15 | Batch 500 | CTC Loss: 3.2945
251
+ [2026-03-25 17:25:48,183][root][INFO] - [Round 0] Epoch 15 | Batch 600 | CTC Loss: 3.3559
252
+ [2026-03-25 17:34:44,601][root][INFO] - [Round 0] Epoch 15 | Batch 700 | CTC Loss: 3.3591
253
+ [2026-03-25 17:43:43,015][root][INFO] - [Round 0] Epoch 15 | Batch 800 | CTC Loss: 4.1683
254
+ [2026-03-25 17:51:52,536][root][INFO] - --- [Round 0] Epoch 15 Complete | Avg CTC Loss: 3.7355 ---
255
+ [2026-03-25 17:51:52,551][root][INFO] - New best model saved (loss=3.7355)
256
+ [2026-03-25 17:52:21,253][root][INFO] - [Round 0] Epoch 16 | Batch 0 | CTC Loss: 3.7749
257
+ [2026-03-25 17:59:59,189][root][INFO] - [Round 0] Epoch 16 | Batch 100 | CTC Loss: 3.6293
258
+ [2026-03-25 18:07:45,907][root][INFO] - [Round 0] Epoch 16 | Batch 200 | CTC Loss: 3.8308
259
+ [2026-03-25 18:16:32,494][root][INFO] - [Round 0] Epoch 16 | Batch 300 | CTC Loss: 4.3118
260
+ [2026-03-25 18:24:58,831][root][INFO] - [Round 0] Epoch 16 | Batch 400 | CTC Loss: 4.5075
261
+ [2026-03-25 18:33:13,839][root][INFO] - [Round 0] Epoch 16 | Batch 500 | CTC Loss: 3.5422
262
+ [2026-03-25 18:41:21,295][root][INFO] - [Round 0] Epoch 16 | Batch 600 | CTC Loss: 3.7810
263
+ [2026-03-25 18:49:25,623][root][INFO] - [Round 0] Epoch 16 | Batch 700 | CTC Loss: 3.2383
264
+ [2026-03-25 18:58:23,718][root][INFO] - [Round 0] Epoch 16 | Batch 800 | CTC Loss: 3.3407
265
+ [2026-03-25 19:06:51,950][root][INFO] - --- [Round 0] Epoch 16 Complete | Avg CTC Loss: 3.7275 ---
266
+ [2026-03-25 19:06:51,965][root][INFO] - New best model saved (loss=3.7275)
267
+ [2026-03-25 19:07:12,186][root][INFO] - [Round 0] Epoch 17 | Batch 0 | CTC Loss: 3.3890
268
+ [2026-03-25 19:15:22,588][root][INFO] - [Round 0] Epoch 17 | Batch 100 | CTC Loss: 3.7778
269
+ [2026-03-25 19:23:04,757][root][INFO] - [Round 0] Epoch 17 | Batch 200 | CTC Loss: 3.5934
270
+ [2026-03-25 19:31:10,655][root][INFO] - [Round 0] Epoch 17 | Batch 300 | CTC Loss: 3.7792
271
+ [2026-03-25 19:38:47,306][root][INFO] - [Round 0] Epoch 17 | Batch 400 | CTC Loss: 4.0482
272
+ [2026-03-25 19:47:38,188][root][INFO] - [Round 0] Epoch 17 | Batch 500 | CTC Loss: 3.7235
273
+ [2026-03-25 19:57:01,099][root][INFO] - [Round 0] Epoch 17 | Batch 600 | CTC Loss: 3.9368
274
+ [2026-03-25 20:05:56,538][root][INFO] - [Round 0] Epoch 17 | Batch 700 | CTC Loss: 4.5842
275
+ [2026-03-25 20:13:52,577][root][INFO] - [Round 0] Epoch 17 | Batch 800 | CTC Loss: 3.4304
276
+ [2026-03-25 20:20:38,481][root][INFO] - --- [Round 0] Epoch 17 Complete | Avg CTC Loss: 3.7092 ---
277
+ [2026-03-25 20:20:38,496][root][INFO] - New best model saved (loss=3.7092)
278
+ [2026-03-25 20:20:56,450][root][INFO] - [Round 0] Epoch 18 | Batch 0 | CTC Loss: 4.0175
279
+ [2026-03-25 20:28:50,512][root][INFO] - [Round 0] Epoch 18 | Batch 100 | CTC Loss: 4.2467
280
+ [2026-03-25 20:36:54,084][root][INFO] - [Round 0] Epoch 18 | Batch 200 | CTC Loss: 3.7234
281
+ [2026-03-25 20:44:57,821][root][INFO] - [Round 0] Epoch 18 | Batch 300 | CTC Loss: 4.4347
282
+ [2026-03-25 20:52:55,465][root][INFO] - [Round 0] Epoch 18 | Batch 400 | CTC Loss: 4.0542
283
+ [2026-03-25 21:00:43,561][root][INFO] - [Round 0] Epoch 18 | Batch 500 | CTC Loss: 3.5506
284
+ [2026-03-25 21:10:15,484][root][INFO] - [Round 0] Epoch 18 | Batch 600 | CTC Loss: 3.2290
285
+ [2026-03-25 21:18:07,742][root][INFO] - [Round 0] Epoch 18 | Batch 700 | CTC Loss: 3.9812
286
+ [2026-03-25 21:25:00,029][root][INFO] - [Round 0] Epoch 18 | Batch 800 | CTC Loss: 3.5093
287
+ [2026-03-25 21:31:26,838][root][INFO] - --- [Round 0] Epoch 18 Complete | Avg CTC Loss: 3.7092 ---
288
+ [2026-03-25 21:31:38,894][root][INFO] - [Round 0] Epoch 19 | Batch 0 | CTC Loss: 3.0095
289
+ [2026-03-25 21:38:50,069][root][INFO] - [Round 0] Epoch 19 | Batch 100 | CTC Loss: 3.2577
290
+ [2026-03-25 21:45:00,035][root][INFO] - [Round 0] Epoch 19 | Batch 200 | CTC Loss: 4.6424
291
+ [2026-03-25 21:52:05,779][root][INFO] - [Round 0] Epoch 19 | Batch 300 | CTC Loss: 3.7431
292
+ [2026-03-25 21:59:02,487][root][INFO] - [Round 0] Epoch 19 | Batch 400 | CTC Loss: 3.3792
293
+ [2026-03-25 22:06:05,192][root][INFO] - [Round 0] Epoch 19 | Batch 500 | CTC Loss: 3.8418
294
+ [2026-03-25 22:12:43,183][root][INFO] - [Round 0] Epoch 19 | Batch 600 | CTC Loss: 4.0602
295
+ [2026-03-25 22:19:52,600][root][INFO] - [Round 0] Epoch 19 | Batch 700 | CTC Loss: 4.0497
296
+ [2026-03-25 22:26:59,840][root][INFO] - [Round 0] Epoch 19 | Batch 800 | CTC Loss: 3.8004
297
+ [2026-03-25 22:33:33,808][root][INFO] - --- [Round 0] Epoch 19 Complete | Avg CTC Loss: 3.7023 ---
298
+ [2026-03-25 22:33:33,828][root][INFO] - New best model saved (loss=3.7023)
299
+ [2026-03-25 22:33:54,557][root][INFO] - [Round 0] Epoch 20 | Batch 0 | CTC Loss: 4.5230
300
+ [2026-03-25 22:40:36,931][root][INFO] - [Round 0] Epoch 20 | Batch 100 | CTC Loss: 3.4090
301
+ [2026-03-25 22:47:22,853][root][INFO] - [Round 0] Epoch 20 | Batch 200 | CTC Loss: 3.5369
302
+ [2026-03-25 22:54:02,230][root][INFO] - [Round 0] Epoch 20 | Batch 300 | CTC Loss: 4.3570
303
+ [2026-03-25 23:01:25,360][root][INFO] - [Round 0] Epoch 20 | Batch 400 | CTC Loss: 3.3017
304
+ [2026-03-25 23:09:06,992][root][INFO] - [Round 0] Epoch 20 | Batch 500 | CTC Loss: 3.2585
305
+ [2026-03-25 23:19:49,242][root][INFO] - [Round 0] Epoch 20 | Batch 600 | CTC Loss: 3.8777
306
+ [2026-03-25 23:32:12,582][root][INFO] - [Round 0] Epoch 20 | Batch 700 | CTC Loss: 3.6071
307
+ [2026-03-25 23:42:12,716][root][INFO] - [Round 0] Epoch 20 | Batch 800 | CTC Loss: 3.9407
308
+ [2026-03-25 23:50:24,136][root][INFO] - --- [Round 0] Epoch 20 Complete | Avg CTC Loss: 3.6854 ---
309
+ [2026-03-25 23:50:24,222][root][INFO] - New best model saved (loss=3.6854)
310
+ [2026-03-25 23:50:51,576][root][INFO] - [Round 0] Epoch 21 | Batch 0 | CTC Loss: 3.4586
311
+ [2026-03-25 23:59:54,146][root][INFO] - [Round 0] Epoch 21 | Batch 100 | CTC Loss: 3.4522
312
+ [2026-03-26 00:08:04,728][root][INFO] - [Round 0] Epoch 21 | Batch 200 | CTC Loss: 3.6354
313
+ [2026-03-26 00:16:17,780][root][INFO] - [Round 0] Epoch 21 | Batch 300 | CTC Loss: 3.5993
314
+ [2026-03-26 00:24:37,150][root][INFO] - [Round 0] Epoch 21 | Batch 400 | CTC Loss: 3.6679
315
+ [2026-03-26 00:33:43,384][root][INFO] - [Round 0] Epoch 21 | Batch 500 | CTC Loss: 3.9266
316
+ [2026-03-26 00:43:15,153][root][INFO] - [Round 0] Epoch 21 | Batch 600 | CTC Loss: 3.7381
317
+ [2026-03-26 00:51:38,442][root][INFO] - [Round 0] Epoch 21 | Batch 700 | CTC Loss: 3.6290
318
+ [2026-03-26 01:00:04,732][root][INFO] - [Round 0] Epoch 21 | Batch 800 | CTC Loss: 3.3096
319
+ [2026-03-26 01:08:14,135][root][INFO] - --- [Round 0] Epoch 21 Complete | Avg CTC Loss: 3.6834 ---
320
+ [2026-03-26 01:08:14,198][root][INFO] - New best model saved (loss=3.6834)
321
+ [2026-03-26 01:08:43,828][root][INFO] - [Round 0] Epoch 22 | Batch 0 | CTC Loss: 3.6669
322
+ [2026-03-26 01:18:07,142][root][INFO] - [Round 0] Epoch 22 | Batch 100 | CTC Loss: 3.9773
323
+ [2026-03-26 01:27:10,117][root][INFO] - [Round 0] Epoch 22 | Batch 200 | CTC Loss: 3.2496
324
+ [2026-03-26 01:36:58,215][root][INFO] - [Round 0] Epoch 22 | Batch 300 | CTC Loss: 3.4904
325
+ [2026-03-26 01:46:40,113][root][INFO] - [Round 0] Epoch 22 | Batch 400 | CTC Loss: 3.8353
326
+ [2026-03-26 01:56:16,766][root][INFO] - [Round 0] Epoch 22 | Batch 500 | CTC Loss: 3.6435
327
+ [2026-03-26 02:06:05,895][root][INFO] - [Round 0] Epoch 22 | Batch 600 | CTC Loss: 3.9816
328
+ [2026-03-26 02:16:41,584][root][INFO] - [Round 0] Epoch 22 | Batch 700 | CTC Loss: 3.8355
329
+ [2026-03-26 02:26:06,566][root][INFO] - [Round 0] Epoch 22 | Batch 800 | CTC Loss: 3.1930
330
+ [2026-03-26 02:34:42,022][root][INFO] - --- [Round 0] Epoch 22 Complete | Avg CTC Loss: 3.6776 ---
331
+ [2026-03-26 02:34:42,092][root][INFO] - New best model saved (loss=3.6776)
332
+ [2026-03-26 02:35:05,060][root][INFO] - [Round 0] Epoch 23 | Batch 0 | CTC Loss: 3.7572
333
+ [2026-03-26 02:43:51,208][root][INFO] - [Round 0] Epoch 23 | Batch 100 | CTC Loss: 3.4665
334
+ [2026-03-26 02:52:41,689][root][INFO] - [Round 0] Epoch 23 | Batch 200 | CTC Loss: 3.7709
335
+ [2026-03-26 03:02:26,774][root][INFO] - [Round 0] Epoch 23 | Batch 300 | CTC Loss: 3.5425
336
+ [2026-03-26 03:11:58,145][root][INFO] - [Round 0] Epoch 23 | Batch 400 | CTC Loss: 3.8046
337
+ [2026-03-26 03:21:17,785][root][INFO] - [Round 0] Epoch 23 | Batch 500 | CTC Loss: 3.2514
338
+ [2026-03-26 03:29:59,504][root][INFO] - [Round 0] Epoch 23 | Batch 600 | CTC Loss: 3.6488
339
+ [2026-03-26 03:38:02,343][root][INFO] - [Round 0] Epoch 23 | Batch 700 | CTC Loss: 3.0197
340
+ [2026-03-26 03:46:44,993][root][INFO] - [Round 0] Epoch 23 | Batch 800 | CTC Loss: 3.3234
341
+ [2026-03-26 03:54:45,754][root][INFO] - --- [Round 0] Epoch 23 Complete | Avg CTC Loss: 3.6616 ---
342
+ [2026-03-26 03:54:45,806][root][INFO] - New best model saved (loss=3.6616)
343
+ [2026-03-26 03:55:03,915][root][INFO] - [Round 0] Epoch 24 | Batch 0 | CTC Loss: 3.7447
344
+ [2026-03-26 04:03:17,014][root][INFO] - [Round 0] Epoch 24 | Batch 100 | CTC Loss: 4.3861
345
+ [2026-03-26 04:11:58,270][root][INFO] - [Round 0] Epoch 24 | Batch 200 | CTC Loss: 3.7630
346
+ [2026-03-26 04:20:08,664][root][INFO] - [Round 0] Epoch 24 | Batch 300 | CTC Loss: 3.3895
347
+ [2026-03-26 04:28:35,280][root][INFO] - [Round 0] Epoch 24 | Batch 400 | CTC Loss: 3.7187
348
+ [2026-03-26 04:36:46,361][root][INFO] - [Round 0] Epoch 24 | Batch 500 | CTC Loss: 4.0656
349
+ [2026-03-26 04:45:36,483][root][INFO] - [Round 0] Epoch 24 | Batch 600 | CTC Loss: 4.0420
350
+ [2026-03-26 04:54:28,526][root][INFO] - [Round 0] Epoch 24 | Batch 700 | CTC Loss: 3.7114
351
+ [2026-03-26 05:03:39,658][root][INFO] - [Round 0] Epoch 24 | Batch 800 | CTC Loss: 3.2641
352
+ [2026-03-26 05:12:05,811][root][INFO] - --- [Round 0] Epoch 24 Complete | Avg CTC Loss: 3.6671 ---
353
+ [2026-03-26 05:12:27,088][root][INFO] - [Round 0] Epoch 25 | Batch 0 | CTC Loss: 3.7479
354
+ [2026-03-26 05:21:08,139][root][INFO] - [Round 0] Epoch 25 | Batch 100 | CTC Loss: 3.2445
355
+ [2026-03-26 05:29:39,643][root][INFO] - [Round 0] Epoch 25 | Batch 200 | CTC Loss: 3.6328
356
+ [2026-03-26 05:38:32,555][root][INFO] - [Round 0] Epoch 25 | Batch 300 | CTC Loss: 4.4605
357
+ [2026-03-26 05:46:34,274][root][INFO] - [Round 0] Epoch 25 | Batch 400 | CTC Loss: 3.7377
358
+ [2026-03-26 05:54:12,936][root][INFO] - [Round 0] Epoch 25 | Batch 500 | CTC Loss: 3.3931
359
+ [2026-03-26 06:00:23,917][root][INFO] - [Round 0] Epoch 25 | Batch 600 | CTC Loss: 3.5351
360
+ [2026-03-26 06:06:45,507][root][INFO] - [Round 0] Epoch 25 | Batch 700 | CTC Loss: 3.3477
361
+ [2026-03-26 06:12:52,679][root][INFO] - [Round 0] Epoch 25 | Batch 800 | CTC Loss: 3.1154
362
+ [2026-03-26 06:18:15,843][root][INFO] - --- [Round 0] Epoch 25 Complete | Avg CTC Loss: 3.6528 ---
363
+ [2026-03-26 06:18:15,859][root][INFO] - New best model saved (loss=3.6528)
364
+ [2026-03-26 06:18:36,088][root][INFO] - [Round 0] Epoch 26 | Batch 0 | CTC Loss: 3.7934
365
+ [2026-03-26 06:25:12,019][root][INFO] - [Round 0] Epoch 26 | Batch 100 | CTC Loss: 3.1090
366
+ [2026-03-26 06:31:23,965][root][INFO] - [Round 0] Epoch 26 | Batch 200 | CTC Loss: 2.9507
367
+ [2026-03-26 06:37:26,058][root][INFO] - [Round 0] Epoch 26 | Batch 300 | CTC Loss: 3.5294
368
+ [2026-03-26 06:43:52,498][root][INFO] - [Round 0] Epoch 26 | Batch 400 | CTC Loss: 3.2548
369
+ [2026-03-26 06:49:56,125][root][INFO] - [Round 0] Epoch 26 | Batch 500 | CTC Loss: 4.1084
370
+ [2026-03-26 06:56:27,102][root][INFO] - [Round 0] Epoch 26 | Batch 600 | CTC Loss: 4.9640
371
+ [2026-03-26 07:02:57,755][root][INFO] - [Round 0] Epoch 26 | Batch 700 | CTC Loss: 3.6453
372
+ [2026-03-26 07:09:13,978][root][INFO] - [Round 0] Epoch 26 | Batch 800 | CTC Loss: 3.3120
373
+ [2026-03-26 07:14:31,670][root][INFO] - --- [Round 0] Epoch 26 Complete | Avg CTC Loss: 3.6287 ---
374
+ [2026-03-26 07:14:31,686][root][INFO] - New best model saved (loss=3.6287)
375
+ [2026-03-26 07:14:48,880][root][INFO] - [Round 0] Epoch 27 | Batch 0 | CTC Loss: 3.2820
376
+ [2026-03-26 07:21:27,373][root][INFO] - [Round 0] Epoch 27 | Batch 100 | CTC Loss: 3.4363
377
+ [2026-03-26 07:27:44,409][root][INFO] - [Round 0] Epoch 27 | Batch 200 | CTC Loss: 3.9326
378
+ [2026-03-26 07:33:42,234][root][INFO] - [Round 0] Epoch 27 | Batch 300 | CTC Loss: 3.3009
379
+ [2026-03-26 07:40:09,428][root][INFO] - [Round 0] Epoch 27 | Batch 400 | CTC Loss: 3.0934
380
+ [2026-03-26 07:46:26,774][root][INFO] - [Round 0] Epoch 27 | Batch 500 | CTC Loss: 3.2058
381
+ [2026-03-26 07:52:20,184][root][INFO] - [Round 0] Epoch 27 | Batch 600 | CTC Loss: 3.3033
382
+ [2026-03-26 07:58:25,245][root][INFO] - [Round 0] Epoch 27 | Batch 700 | CTC Loss: 4.0912
SpidR/256_no_extra_augs/{config.yaml → .hydra/config.yaml} RENAMED
@@ -4,7 +4,7 @@ training:
4
  learning_rate: 0.0001
5
  log_interval: 100
6
  checkpoint_dir: null
7
- resume_from: /home/infres/abrik-22/snlp_project/outputs/spidr_base/2026-03-23/22-15-04/round_0/E1_last.pt
8
  n_iterative_pseudolabeling: 3
9
  lr_scheduler:
10
  _target_: torch.optim.lr_scheduler.CosineAnnealingLR
 
4
  learning_rate: 0.0001
5
  log_interval: 100
6
  checkpoint_dir: null
7
+ resume_from: null
8
  n_iterative_pseudolabeling: 3
9
  lr_scheduler:
10
  _target_: torch.optim.lr_scheduler.CosineAnnealingLR
SpidR/256_no_extra_augs/.hydra/hydra.yaml ADDED
@@ -0,0 +1,160 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ hydra:
2
+ run:
3
+ dir: outputs/${training.run_name}/${now:%Y-%m-%d}/${now:%H-%M-%S}
4
+ sweep:
5
+ dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S}
6
+ subdir: ${hydra.job.num}
7
+ launcher:
8
+ _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
9
+ sweeper:
10
+ _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
11
+ max_batch_size: null
12
+ params: null
13
+ help:
14
+ app_name: ${hydra.job.name}
15
+ header: '${hydra.help.app_name} is powered by Hydra.
16
+
17
+ '
18
+ footer: 'Powered by Hydra (https://hydra.cc)
19
+
20
+ Use --hydra-help to view Hydra specific help
21
+
22
+ '
23
+ template: '${hydra.help.header}
24
+
25
+ == Configuration groups ==
26
+
27
+ Compose your configuration from those groups (group=option)
28
+
29
+
30
+ $APP_CONFIG_GROUPS
31
+
32
+
33
+ == Config ==
34
+
35
+ Override anything in the config (foo.bar=value)
36
+
37
+
38
+ $CONFIG
39
+
40
+
41
+ ${hydra.help.footer}
42
+
43
+ '
44
+ hydra_help:
45
+ template: 'Hydra (${hydra.runtime.version})
46
+
47
+ See https://hydra.cc for more info.
48
+
49
+
50
+ == Flags ==
51
+
52
+ $FLAGS_HELP
53
+
54
+
55
+ == Configuration groups ==
56
+
57
+ Compose your configuration from those groups (For example, append hydra/job_logging=disabled
58
+ to command line)
59
+
60
+
61
+ $HYDRA_CONFIG_GROUPS
62
+
63
+
64
+ Use ''--cfg hydra'' to Show the Hydra config.
65
+
66
+ '
67
+ hydra_help: ???
68
+ hydra_logging:
69
+ version: 1
70
+ formatters:
71
+ simple:
72
+ format: '[%(asctime)s][HYDRA] %(message)s'
73
+ handlers:
74
+ console:
75
+ class: logging.StreamHandler
76
+ formatter: simple
77
+ stream: ext://sys.stdout
78
+ root:
79
+ level: INFO
80
+ handlers:
81
+ - console
82
+ loggers:
83
+ logging_example:
84
+ level: DEBUG
85
+ disable_existing_loggers: false
86
+ job_logging:
87
+ version: 1
88
+ formatters:
89
+ simple:
90
+ format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
91
+ handlers:
92
+ console:
93
+ class: logging.StreamHandler
94
+ formatter: simple
95
+ stream: ext://sys.stdout
96
+ file:
97
+ class: logging.FileHandler
98
+ formatter: simple
99
+ filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
100
+ root:
101
+ level: INFO
102
+ handlers:
103
+ - console
104
+ - file
105
+ disable_existing_loggers: false
106
+ env: {}
107
+ mode: RUN
108
+ searchpath: []
109
+ callbacks: {}
110
+ output_subdir: .hydra
111
+ overrides:
112
+ hydra:
113
+ - hydra.mode=RUN
114
+ task:
115
+ - training.run_name=spidr_base
116
+ - model.name=spidr_base
117
+ - model.layer=6
118
+ - model.vocab_size=256
119
+ - dataset.augmentations.max_augs=1
120
+ - dataset.augmentations.activate_extra_augs=False
121
+ job:
122
+ name: train
123
+ chdir: null
124
+ override_dirname: dataset.augmentations.activate_extra_augs=False,dataset.augmentations.max_augs=1,model.layer=6,model.name=spidr_base,model.vocab_size=256,training.run_name=spidr_base
125
+ id: ???
126
+ num: ???
127
+ config_name: quantization
128
+ env_set: {}
129
+ env_copy: []
130
+ config:
131
+ override_dirname:
132
+ kv_sep: '='
133
+ item_sep: ','
134
+ exclude_keys: []
135
+ runtime:
136
+ version: 1.3.2
137
+ version_base: '1.3'
138
+ cwd: /home/infres/abrik-22/snlp_project
139
+ config_sources:
140
+ - path: hydra.conf
141
+ schema: pkg
142
+ provider: hydra
143
+ - path: /home/infres/abrik-22/snlp_project/configs
144
+ schema: file
145
+ provider: main
146
+ - path: ''
147
+ schema: structured
148
+ provider: schema
149
+ output_dir: /home/infres/abrik-22/snlp_project/outputs/spidr_base/2026-03-24/19-55-13
150
+ choices:
151
+ hydra/env: default
152
+ hydra/callbacks: null
153
+ hydra/job_logging: default
154
+ hydra/hydra_logging: default
155
+ hydra/hydra_help: default
156
+ hydra/help: default
157
+ hydra/sweeper: basic
158
+ hydra/launcher: basic
159
+ hydra/output: default
160
+ verbose: false
SpidR/256_no_extra_augs/.hydra/overrides.yaml ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ - training.run_name=spidr_base
2
+ - model.name=spidr_base
3
+ - model.layer=6
4
+ - model.vocab_size=256
5
+ - dataset.augmentations.max_augs=1
6
+ - dataset.augmentations.activate_extra_augs=False
SpidR/256_no_extra_augs/round_0/E1_best.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a8f6281df692e3f0aa563baeb0d6126019d75f1ad66bfaff7a5c9bd82845d0be
3
  size 3961480
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3afa7ff4caf2d85b6d8eec5ee56ece5b3ae5ceb68a438037a116c3412c03534
3
  size 3961480
SpidR/256_no_extra_augs/round_0/E1_last.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:98c0419a53fa2c7d5dac9b5c4253da86666b119f007bfdf11f779d9cbae05739
3
  size 3961480
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c96ac546534843ed766784626ed594ea8006beee103cf6de502bf29dfcf34dba
3
  size 3961480
SpidR/256_no_extra_augs/round_0/tensorboard/events.out.tfevents.1774345454.node47.1425577.0 DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:9b24e73db74ad8005cca1f249be0c046d577920fe06b0db057cd2617e7c95c04
3
- size 9060
 
 
 
 
SpidR/256_no_extra_augs/round_0/tensorboard/events.out.tfevents.1774378527.node20.1786403.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:af82284c6ff59ed1eae97055355c01457ac0adb81b86e8e606604628cb9b9d93
3
- size 282972
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd0b2d07d70d1276ef59dca708a34dacdd8d6ecbd0256b8725f7513699e111ca
3
+ size 308692
SpidR/256_no_extra_augs/train.log CHANGED
@@ -290,3 +290,258 @@ model:
290
  [2026-03-25 12:15:24,747][root][INFO] - [Round 0] Epoch 19 | Batch 100 | CTC Loss: 3.9526
291
  [2026-03-25 12:20:08,846][root][INFO] - [Round 0] Epoch 19 | Batch 200 | CTC Loss: 3.6656
292
  [2026-03-25 12:24:35,201][root][INFO] - [Round 0] Epoch 19 | Batch 300 | CTC Loss: 3.8360
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
290
  [2026-03-25 12:15:24,747][root][INFO] - [Round 0] Epoch 19 | Batch 100 | CTC Loss: 3.9526
291
  [2026-03-25 12:20:08,846][root][INFO] - [Round 0] Epoch 19 | Batch 200 | CTC Loss: 3.6656
292
  [2026-03-25 12:24:35,201][root][INFO] - [Round 0] Epoch 19 | Batch 300 | CTC Loss: 3.8360
293
+ [2026-03-25 12:29:41,233][root][INFO] - [Round 0] Epoch 19 | Batch 400 | CTC Loss: 2.8035
294
+ [2026-03-25 12:34:42,077][root][INFO] - [Round 0] Epoch 19 | Batch 500 | CTC Loss: 4.3012
295
+ [2026-03-25 12:39:51,478][root][INFO] - [Round 0] Epoch 19 | Batch 600 | CTC Loss: 3.8623
296
+ [2026-03-25 12:45:00,590][root][INFO] - [Round 0] Epoch 19 | Batch 700 | CTC Loss: 4.0740
297
+ [2026-03-25 12:50:30,317][root][INFO] - [Round 0] Epoch 19 | Batch 800 | CTC Loss: 3.4692
298
+ [2026-03-25 12:55:13,089][root][INFO] - --- [Round 0] Epoch 19 Complete | Avg CTC Loss: 3.6497 ---
299
+ [2026-03-25 12:55:29,717][root][INFO] - [Round 0] Epoch 20 | Batch 0 | CTC Loss: 3.8753
300
+ [2026-03-25 13:01:21,552][root][INFO] - [Round 0] Epoch 20 | Batch 100 | CTC Loss: 3.7793
301
+ [2026-03-25 13:06:11,834][root][INFO] - [Round 0] Epoch 20 | Batch 200 | CTC Loss: 4.0607
302
+ [2026-03-25 13:10:36,071][root][INFO] - [Round 0] Epoch 20 | Batch 300 | CTC Loss: 3.3336
303
+ [2026-03-25 13:15:16,344][root][INFO] - [Round 0] Epoch 20 | Batch 400 | CTC Loss: 3.8760
304
+ [2026-03-25 13:20:53,198][root][INFO] - [Round 0] Epoch 20 | Batch 500 | CTC Loss: 4.1954
305
+ [2026-03-25 13:26:49,363][root][INFO] - [Round 0] Epoch 20 | Batch 600 | CTC Loss: 3.8851
306
+ [2026-03-25 13:33:12,383][root][INFO] - [Round 0] Epoch 20 | Batch 700 | CTC Loss: 3.7731
307
+ [2026-03-25 13:39:36,833][root][INFO] - [Round 0] Epoch 20 | Batch 800 | CTC Loss: 3.4848
308
+ [2026-03-25 13:45:28,534][root][INFO] - --- [Round 0] Epoch 20 Complete | Avg CTC Loss: 3.6499 ---
309
+ [2026-03-25 13:45:56,309][root][INFO] - [Round 0] Epoch 21 | Batch 0 | CTC Loss: 3.5699
310
+ [2026-03-25 13:53:04,174][root][INFO] - [Round 0] Epoch 21 | Batch 100 | CTC Loss: 3.9475
311
+ [2026-03-25 13:59:25,642][root][INFO] - [Round 0] Epoch 21 | Batch 200 | CTC Loss: 3.5928
312
+ [2026-03-25 14:05:18,016][root][INFO] - [Round 0] Epoch 21 | Batch 300 | CTC Loss: 3.2913
313
+ [2026-03-25 14:11:16,156][root][INFO] - [Round 0] Epoch 21 | Batch 400 | CTC Loss: 3.3938
314
+ [2026-03-25 14:17:57,713][root][INFO] - [Round 0] Epoch 21 | Batch 500 | CTC Loss: 3.8090
315
+ [2026-03-25 14:24:23,222][root][INFO] - [Round 0] Epoch 21 | Batch 600 | CTC Loss: 3.5404
316
+ [2026-03-25 14:30:56,150][root][INFO] - [Round 0] Epoch 21 | Batch 700 | CTC Loss: 3.5861
317
+ [2026-03-25 14:36:37,790][root][INFO] - [Round 0] Epoch 21 | Batch 800 | CTC Loss: 4.0167
318
+ [2026-03-25 14:41:30,786][root][INFO] - --- [Round 0] Epoch 21 Complete | Avg CTC Loss: 3.6328 ---
319
+ [2026-03-25 14:41:30,802][root][INFO] - New best model saved (loss=3.6328)
320
+ [2026-03-25 14:41:55,437][root][INFO] - [Round 0] Epoch 22 | Batch 0 | CTC Loss: 3.6411
321
+ [2026-03-25 14:48:00,729][root][INFO] - [Round 0] Epoch 22 | Batch 100 | CTC Loss: 3.8181
322
+ [2026-03-25 14:53:45,253][root][INFO] - [Round 0] Epoch 22 | Batch 200 | CTC Loss: 3.3228
323
+ [2026-03-25 14:59:35,142][root][INFO] - [Round 0] Epoch 22 | Batch 300 | CTC Loss: 3.4523
324
+ [2026-03-25 15:05:28,807][root][INFO] - [Round 0] Epoch 22 | Batch 400 | CTC Loss: 3.7408
325
+ [2026-03-25 15:11:22,384][root][INFO] - [Round 0] Epoch 22 | Batch 500 | CTC Loss: 4.1492
326
+ [2026-03-25 15:16:53,345][root][INFO] - [Round 0] Epoch 22 | Batch 600 | CTC Loss: 3.3775
327
+ [2026-03-25 15:22:46,954][root][INFO] - [Round 0] Epoch 22 | Batch 700 | CTC Loss: 3.2791
328
+ [2026-03-25 15:28:51,653][root][INFO] - [Round 0] Epoch 22 | Batch 800 | CTC Loss: 3.7694
329
+ [2026-03-25 15:34:08,233][root][INFO] - --- [Round 0] Epoch 22 Complete | Avg CTC Loss: 3.6272 ---
330
+ [2026-03-25 15:34:08,247][root][INFO] - New best model saved (loss=3.6272)
331
+ [2026-03-25 15:34:28,461][root][INFO] - [Round 0] Epoch 23 | Batch 0 | CTC Loss: 3.7218
332
+ [2026-03-25 15:40:08,657][root][INFO] - [Round 0] Epoch 23 | Batch 100 | CTC Loss: 3.3474
333
+ [2026-03-25 15:44:29,842][root][INFO] - [Round 0] Epoch 23 | Batch 200 | CTC Loss: 3.4187
334
+ [2026-03-25 15:49:11,470][root][INFO] - [Round 0] Epoch 23 | Batch 300 | CTC Loss: 3.6899
335
+ [2026-03-25 15:53:37,826][root][INFO] - [Round 0] Epoch 23 | Batch 400 | CTC Loss: 3.1727
336
+ [2026-03-25 15:58:03,390][root][INFO] - [Round 0] Epoch 23 | Batch 500 | CTC Loss: 3.8516
337
+ [2026-03-25 16:02:32,528][root][INFO] - [Round 0] Epoch 23 | Batch 600 | CTC Loss: 3.1551
338
+ [2026-03-25 16:06:56,641][root][INFO] - [Round 0] Epoch 23 | Batch 700 | CTC Loss: 4.0029
339
+ [2026-03-25 16:11:26,106][root][INFO] - [Round 0] Epoch 23 | Batch 800 | CTC Loss: 3.8292
340
+ [2026-03-25 16:15:27,196][root][INFO] - --- [Round 0] Epoch 23 Complete | Avg CTC Loss: 3.5968 ---
341
+ [2026-03-25 16:15:27,211][root][INFO] - New best model saved (loss=3.5968)
342
+ [2026-03-25 16:15:53,348][root][INFO] - [Round 0] Epoch 24 | Batch 0 | CTC Loss: 4.5441
343
+ [2026-03-25 16:20:40,027][root][INFO] - [Round 0] Epoch 24 | Batch 100 | CTC Loss: 3.6386
344
+ [2026-03-25 16:25:01,815][root][INFO] - [Round 0] Epoch 24 | Batch 200 | CTC Loss: 3.6145
345
+ [2026-03-25 16:29:26,153][root][INFO] - [Round 0] Epoch 24 | Batch 300 | CTC Loss: 3.2107
346
+ [2026-03-25 16:34:05,331][root][INFO] - [Round 0] Epoch 24 | Batch 400 | CTC Loss: 3.6807
347
+ [2026-03-25 16:38:40,744][root][INFO] - [Round 0] Epoch 24 | Batch 500 | CTC Loss: 3.4387
348
+ [2026-03-25 16:43:03,482][root][INFO] - [Round 0] Epoch 24 | Batch 600 | CTC Loss: 3.5205
349
+ [2026-03-25 16:47:32,960][root][INFO] - [Round 0] Epoch 24 | Batch 700 | CTC Loss: 3.1282
350
+ [2026-03-25 16:52:14,687][root][INFO] - [Round 0] Epoch 24 | Batch 800 | CTC Loss: 4.1578
351
+ [2026-03-25 16:56:22,206][root][INFO] - --- [Round 0] Epoch 24 Complete | Avg CTC Loss: 3.6184 ---
352
+ [2026-03-25 16:56:31,146][root][INFO] - [Round 0] Epoch 25 | Batch 0 | CTC Loss: 4.3838
353
+ [2026-03-25 17:01:30,024][root][INFO] - [Round 0] Epoch 25 | Batch 100 | CTC Loss: 4.4904
354
+ [2026-03-25 17:06:18,808][root][INFO] - [Round 0] Epoch 25 | Batch 200 | CTC Loss: 3.2599
355
+ [2026-03-25 17:11:51,321][root][INFO] - [Round 0] Epoch 25 | Batch 300 | CTC Loss: 4.2021
356
+ [2026-03-25 17:17:17,574][root][INFO] - [Round 0] Epoch 25 | Batch 400 | CTC Loss: 3.9542
357
+ [2026-03-25 17:22:55,725][root][INFO] - [Round 0] Epoch 25 | Batch 500 | CTC Loss: 3.8178
358
+ [2026-03-25 17:28:18,444][root][INFO] - [Round 0] Epoch 25 | Batch 600 | CTC Loss: 3.3892
359
+ [2026-03-25 17:33:57,963][root][INFO] - [Round 0] Epoch 25 | Batch 700 | CTC Loss: 3.5181
360
+ [2026-03-25 17:39:17,542][root][INFO] - [Round 0] Epoch 25 | Batch 800 | CTC Loss: 3.6709
361
+ [2026-03-25 17:44:14,057][root][INFO] - --- [Round 0] Epoch 25 Complete | Avg CTC Loss: 3.6069 ---
362
+ [2026-03-25 17:44:33,573][root][INFO] - [Round 0] Epoch 26 | Batch 0 | CTC Loss: 3.2466
363
+ [2026-03-25 17:50:21,473][root][INFO] - [Round 0] Epoch 26 | Batch 100 | CTC Loss: 3.4415
364
+ [2026-03-25 17:55:29,010][root][INFO] - [Round 0] Epoch 26 | Batch 200 | CTC Loss: 3.4575
365
+ [2026-03-25 18:01:02,256][root][INFO] - [Round 0] Epoch 26 | Batch 300 | CTC Loss: 4.1024
366
+ [2026-03-25 18:06:35,986][root][INFO] - [Round 0] Epoch 26 | Batch 400 | CTC Loss: 3.5115
367
+ [2026-03-25 18:11:32,240][root][INFO] - [Round 0] Epoch 26 | Batch 500 | CTC Loss: 3.7145
368
+ [2026-03-25 18:15:52,332][root][INFO] - [Round 0] Epoch 26 | Batch 600 | CTC Loss: 3.7434
369
+ [2026-03-25 18:20:22,508][root][INFO] - [Round 0] Epoch 26 | Batch 700 | CTC Loss: 3.4672
370
+ [2026-03-25 18:24:52,528][root][INFO] - [Round 0] Epoch 26 | Batch 800 | CTC Loss: 3.7331
371
+ [2026-03-25 18:28:58,328][root][INFO] - --- [Round 0] Epoch 26 Complete | Avg CTC Loss: 3.6009 ---
372
+ [2026-03-25 18:29:23,082][root][INFO] - [Round 0] Epoch 27 | Batch 0 | CTC Loss: 3.6779
373
+ [2026-03-25 18:33:55,236][root][INFO] - [Round 0] Epoch 27 | Batch 100 | CTC Loss: 3.9144
374
+ [2026-03-25 18:38:14,274][root][INFO] - [Round 0] Epoch 27 | Batch 200 | CTC Loss: 3.3428
375
+ [2026-03-25 18:42:38,863][root][INFO] - [Round 0] Epoch 27 | Batch 300 | CTC Loss: 3.8594
376
+ [2026-03-25 18:47:07,649][root][INFO] - [Round 0] Epoch 27 | Batch 400 | CTC Loss: 3.4887
377
+ [2026-03-25 18:51:32,197][root][INFO] - [Round 0] Epoch 27 | Batch 500 | CTC Loss: 3.0052
378
+ [2026-03-25 18:56:03,752][root][INFO] - [Round 0] Epoch 27 | Batch 600 | CTC Loss: 3.5498
379
+ [2026-03-25 19:00:25,527][root][INFO] - [Round 0] Epoch 27 | Batch 700 | CTC Loss: 3.3226
380
+ [2026-03-25 19:04:44,480][root][INFO] - [Round 0] Epoch 27 | Batch 800 | CTC Loss: 2.9365
381
+ [2026-03-25 19:08:41,072][root][INFO] - --- [Round 0] Epoch 27 Complete | Avg CTC Loss: 3.5995 ---
382
+ [2026-03-25 19:08:55,545][root][INFO] - [Round 0] Epoch 28 | Batch 0 | CTC Loss: 3.2900
383
+ [2026-03-25 19:13:39,885][root][INFO] - [Round 0] Epoch 28 | Batch 100 | CTC Loss: 3.3864
384
+ [2026-03-25 19:18:01,759][root][INFO] - [Round 0] Epoch 28 | Batch 200 | CTC Loss: 3.7722
385
+ [2026-03-25 19:22:28,830][root][INFO] - [Round 0] Epoch 28 | Batch 300 | CTC Loss: 3.5791
386
+ [2026-03-25 19:26:58,105][root][INFO] - [Round 0] Epoch 28 | Batch 400 | CTC Loss: 3.6093
387
+ [2026-03-25 19:31:27,768][root][INFO] - [Round 0] Epoch 28 | Batch 500 | CTC Loss: 3.1078
388
+ [2026-03-25 19:35:56,774][root][INFO] - [Round 0] Epoch 28 | Batch 600 | CTC Loss: 3.7419
389
+ [2026-03-25 19:40:23,484][root][INFO] - [Round 0] Epoch 28 | Batch 700 | CTC Loss: 3.2298
390
+ [2026-03-25 19:45:26,322][root][INFO] - [Round 0] Epoch 28 | Batch 800 | CTC Loss: 3.7927
391
+ [2026-03-25 19:51:22,743][root][INFO] - --- [Round 0] Epoch 28 Complete | Avg CTC Loss: 3.5689 ---
392
+ [2026-03-25 19:51:22,758][root][INFO] - New best model saved (loss=3.5689)
393
+ [2026-03-25 19:52:03,766][root][INFO] - [Round 0] Epoch 29 | Batch 0 | CTC Loss: 4.0982
394
+ [2026-03-25 19:58:13,789][root][INFO] - [Round 0] Epoch 29 | Batch 100 | CTC Loss: 3.1790
395
+ [2026-03-25 20:03:31,062][root][INFO] - [Round 0] Epoch 29 | Batch 200 | CTC Loss: 3.7950
396
+ [2026-03-25 20:08:45,380][root][INFO] - [Round 0] Epoch 29 | Batch 300 | CTC Loss: 3.5095
397
+ [2026-03-25 20:14:08,771][root][INFO] - [Round 0] Epoch 29 | Batch 400 | CTC Loss: 3.2626
398
+ [2026-03-25 20:19:48,654][root][INFO] - [Round 0] Epoch 29 | Batch 500 | CTC Loss: 3.7421
399
+ [2026-03-25 20:25:26,393][root][INFO] - [Round 0] Epoch 29 | Batch 600 | CTC Loss: 4.1946
400
+ [2026-03-25 20:30:33,846][root][INFO] - [Round 0] Epoch 29 | Batch 700 | CTC Loss: 4.1888
401
+ [2026-03-25 20:35:57,755][root][INFO] - [Round 0] Epoch 29 | Batch 800 | CTC Loss: 3.8665
402
+ [2026-03-25 20:40:39,733][root][INFO] - --- [Round 0] Epoch 29 Complete | Avg CTC Loss: 3.5980 ---
403
+ [2026-03-25 20:40:58,496][root][INFO] - [Round 0] Epoch 30 | Batch 0 | CTC Loss: 2.8874
404
+ [2026-03-25 20:47:05,056][root][INFO] - [Round 0] Epoch 30 | Batch 100 | CTC Loss: 3.3631
405
+ [2026-03-25 20:52:23,083][root][INFO] - [Round 0] Epoch 30 | Batch 200 | CTC Loss: 3.5804
406
+ [2026-03-25 20:58:00,363][root][INFO] - [Round 0] Epoch 30 | Batch 300 | CTC Loss: 3.8400
407
+ [2026-03-25 21:03:28,048][root][INFO] - [Round 0] Epoch 30 | Batch 400 | CTC Loss: 3.6998
408
+ [2026-03-25 21:09:01,100][root][INFO] - [Round 0] Epoch 30 | Batch 500 | CTC Loss: 3.5309
409
+ [2026-03-25 21:14:07,038][root][INFO] - [Round 0] Epoch 30 | Batch 600 | CTC Loss: 3.5656
410
+ [2026-03-25 21:19:42,255][root][INFO] - [Round 0] Epoch 30 | Batch 700 | CTC Loss: 3.0675
411
+ [2026-03-25 21:24:54,046][root][INFO] - [Round 0] Epoch 30 | Batch 800 | CTC Loss: 4.2648
412
+ [2026-03-25 21:29:48,993][root][INFO] - --- [Round 0] Epoch 30 Complete | Avg CTC Loss: 3.5892 ---
413
+ [2026-03-25 21:30:03,003][root][INFO] - [Round 0] Epoch 31 | Batch 0 | CTC Loss: 3.9041
414
+ [2026-03-25 21:35:54,511][root][INFO] - [Round 0] Epoch 31 | Batch 100 | CTC Loss: 3.3211
415
+ [2026-03-25 21:41:19,207][root][INFO] - [Round 0] Epoch 31 | Batch 200 | CTC Loss: 4.1335
416
+ [2026-03-25 21:46:24,685][root][INFO] - [Round 0] Epoch 31 | Batch 300 | CTC Loss: 3.2268
417
+ [2026-03-25 21:52:13,360][root][INFO] - [Round 0] Epoch 31 | Batch 400 | CTC Loss: 3.7726
418
+ [2026-03-25 21:57:26,835][root][INFO] - [Round 0] Epoch 31 | Batch 500 | CTC Loss: 3.3182
419
+ [2026-03-25 22:02:32,270][root][INFO] - [Round 0] Epoch 31 | Batch 600 | CTC Loss: 3.5794
420
+ [2026-03-25 22:07:56,420][root][INFO] - [Round 0] Epoch 31 | Batch 700 | CTC Loss: 4.2252
421
+ [2026-03-25 22:12:56,632][root][INFO] - [Round 0] Epoch 31 | Batch 800 | CTC Loss: 2.9715
422
+ [2026-03-25 22:17:36,304][root][INFO] - --- [Round 0] Epoch 31 Complete | Avg CTC Loss: 3.5779 ---
423
+ [2026-03-25 22:17:57,647][root][INFO] - [Round 0] Epoch 32 | Batch 0 | CTC Loss: 3.9335
424
+ [2026-03-25 22:23:44,396][root][INFO] - [Round 0] Epoch 32 | Batch 100 | CTC Loss: 3.3803
425
+ [2026-03-25 22:28:43,064][root][INFO] - [Round 0] Epoch 32 | Batch 200 | CTC Loss: 3.7007
426
+ [2026-03-25 22:34:00,083][root][INFO] - [Round 0] Epoch 32 | Batch 300 | CTC Loss: 3.6780
427
+ [2026-03-25 22:39:20,452][root][INFO] - [Round 0] Epoch 32 | Batch 400 | CTC Loss: 3.8699
428
+ [2026-03-25 22:44:47,763][root][INFO] - [Round 0] Epoch 32 | Batch 500 | CTC Loss: 3.8989
429
+ [2026-03-25 22:49:49,517][root][INFO] - [Round 0] Epoch 32 | Batch 600 | CTC Loss: 3.5343
430
+ [2026-03-25 22:55:17,407][root][INFO] - [Round 0] Epoch 32 | Batch 700 | CTC Loss: 3.2699
431
+ [2026-03-25 23:00:29,133][root][INFO] - [Round 0] Epoch 32 | Batch 800 | CTC Loss: 3.7242
432
+ [2026-03-25 23:05:26,706][root][INFO] - --- [Round 0] Epoch 32 Complete | Avg CTC Loss: 3.5859 ---
433
+ [2026-03-25 23:05:51,890][root][INFO] - [Round 0] Epoch 33 | Batch 0 | CTC Loss: 3.7860
434
+ [2026-03-25 23:11:36,094][root][INFO] - [Round 0] Epoch 33 | Batch 100 | CTC Loss: 3.5500
435
+ [2026-03-25 23:17:16,510][root][INFO] - [Round 0] Epoch 33 | Batch 200 | CTC Loss: 3.6155
436
+ [2026-03-25 23:22:22,443][root][INFO] - [Round 0] Epoch 33 | Batch 300 | CTC Loss: 3.9384
437
+ [2026-03-25 23:27:28,553][root][INFO] - [Round 0] Epoch 33 | Batch 400 | CTC Loss: 4.1554
438
+ [2026-03-25 23:32:39,228][root][INFO] - [Round 0] Epoch 33 | Batch 500 | CTC Loss: 3.7352
439
+ [2026-03-25 23:37:52,392][root][INFO] - [Round 0] Epoch 33 | Batch 600 | CTC Loss: 4.5982
440
+ [2026-03-25 23:43:27,863][root][INFO] - [Round 0] Epoch 33 | Batch 700 | CTC Loss: 3.4278
441
+ [2026-03-25 23:49:24,169][root][INFO] - [Round 0] Epoch 33 | Batch 800 | CTC Loss: 3.1218
442
+ [2026-03-25 23:54:16,245][root][INFO] - --- [Round 0] Epoch 33 Complete | Avg CTC Loss: 3.5742 ---
443
+ [2026-03-25 23:54:36,012][root][INFO] - [Round 0] Epoch 34 | Batch 0 | CTC Loss: 3.3804
444
+ [2026-03-26 00:00:14,624][root][INFO] - [Round 0] Epoch 34 | Batch 100 | CTC Loss: 3.6649
445
+ [2026-03-26 00:05:22,157][root][INFO] - [Round 0] Epoch 34 | Batch 200 | CTC Loss: 3.8063
446
+ [2026-03-26 00:10:18,432][root][INFO] - [Round 0] Epoch 34 | Batch 300 | CTC Loss: 3.4326
447
+ [2026-03-26 00:15:30,700][root][INFO] - [Round 0] Epoch 34 | Batch 400 | CTC Loss: 3.5160
448
+ [2026-03-26 00:20:40,113][root][INFO] - [Round 0] Epoch 34 | Batch 500 | CTC Loss: 3.5773
449
+ [2026-03-26 00:25:28,280][root][INFO] - [Round 0] Epoch 34 | Batch 600 | CTC Loss: 3.6058
450
+ [2026-03-26 00:30:15,079][root][INFO] - [Round 0] Epoch 34 | Batch 700 | CTC Loss: 3.2571
451
+ [2026-03-26 00:36:07,509][root][INFO] - [Round 0] Epoch 34 | Batch 800 | CTC Loss: 3.5468
452
+ [2026-03-26 00:41:23,351][root][INFO] - --- [Round 0] Epoch 34 Complete | Avg CTC Loss: 3.5821 ---
453
+ [2026-03-26 00:41:27,012][root][INFO] - [Round 0] Epoch 35 | Batch 0 | CTC Loss: 3.3722
454
+ [2026-03-26 00:47:21,585][root][INFO] - [Round 0] Epoch 35 | Batch 100 | CTC Loss: 3.4181
455
+ [2026-03-26 00:52:45,031][root][INFO] - [Round 0] Epoch 35 | Batch 200 | CTC Loss: 3.7325
456
+ [2026-03-26 00:58:03,487][root][INFO] - [Round 0] Epoch 35 | Batch 300 | CTC Loss: 3.7155
457
+ [2026-03-26 01:03:10,684][root][INFO] - [Round 0] Epoch 35 | Batch 400 | CTC Loss: 3.5660
458
+ [2026-03-26 01:08:28,878][root][INFO] - [Round 0] Epoch 35 | Batch 500 | CTC Loss: 3.4437
459
+ [2026-03-26 01:14:11,612][root][INFO] - [Round 0] Epoch 35 | Batch 600 | CTC Loss: 3.5124
460
+ [2026-03-26 01:19:07,233][root][INFO] - [Round 0] Epoch 35 | Batch 700 | CTC Loss: 3.5705
461
+ [2026-03-26 01:23:57,958][root][INFO] - [Round 0] Epoch 35 | Batch 800 | CTC Loss: 3.3420
462
+ [2026-03-26 01:28:38,324][root][INFO] - --- [Round 0] Epoch 35 Complete | Avg CTC Loss: 3.5758 ---
463
+ [2026-03-26 01:29:03,249][root][INFO] - [Round 0] Epoch 36 | Batch 0 | CTC Loss: 3.2298
464
+ [2026-03-26 01:34:17,570][root][INFO] - [Round 0] Epoch 36 | Batch 100 | CTC Loss: 3.5712
465
+ [2026-03-26 01:39:17,193][root][INFO] - [Round 0] Epoch 36 | Batch 200 | CTC Loss: 3.6631
466
+ [2026-03-26 01:44:06,252][root][INFO] - [Round 0] Epoch 36 | Batch 300 | CTC Loss: 4.0051
467
+ [2026-03-26 01:48:53,563][root][INFO] - [Round 0] Epoch 36 | Batch 400 | CTC Loss: 3.5669
468
+ [2026-03-26 01:53:41,026][root][INFO] - [Round 0] Epoch 36 | Batch 500 | CTC Loss: 3.4206
469
+ [2026-03-26 01:58:24,997][root][INFO] - [Round 0] Epoch 36 | Batch 600 | CTC Loss: 3.0495
470
+ [2026-03-26 02:03:47,510][root][INFO] - [Round 0] Epoch 36 | Batch 700 | CTC Loss: 3.1517
471
+ [2026-03-26 02:08:59,290][root][INFO] - [Round 0] Epoch 36 | Batch 800 | CTC Loss: 3.8325
472
+ [2026-03-26 02:13:50,179][root][INFO] - --- [Round 0] Epoch 36 Complete | Avg CTC Loss: 3.5677 ---
473
+ [2026-03-26 02:13:50,198][root][INFO] - New best model saved (loss=3.5677)
474
+ [2026-03-26 02:14:08,676][root][INFO] - [Round 0] Epoch 37 | Batch 0 | CTC Loss: 3.4056
475
+ [2026-03-26 02:19:10,480][root][INFO] - [Round 0] Epoch 37 | Batch 100 | CTC Loss: 3.5193
476
+ [2026-03-26 02:24:07,060][root][INFO] - [Round 0] Epoch 37 | Batch 200 | CTC Loss: 3.7741
477
+ [2026-03-26 02:28:56,227][root][INFO] - [Round 0] Epoch 37 | Batch 300 | CTC Loss: 3.2177
478
+ [2026-03-26 02:33:38,095][root][INFO] - [Round 0] Epoch 37 | Batch 400 | CTC Loss: 3.8843
479
+ [2026-03-26 02:38:14,812][root][INFO] - [Round 0] Epoch 37 | Batch 500 | CTC Loss: 3.8606
480
+ [2026-03-26 02:42:54,628][root][INFO] - [Round 0] Epoch 37 | Batch 600 | CTC Loss: 3.7446
481
+ [2026-03-26 02:48:43,814][root][INFO] - [Round 0] Epoch 37 | Batch 700 | CTC Loss: 3.4902
482
+ [2026-03-26 02:54:12,807][root][INFO] - [Round 0] Epoch 37 | Batch 800 | CTC Loss: 3.2551
483
+ [2026-03-26 02:58:59,471][root][INFO] - --- [Round 0] Epoch 37 Complete | Avg CTC Loss: 3.5597 ---
484
+ [2026-03-26 02:58:59,486][root][INFO] - New best model saved (loss=3.5597)
485
+ [2026-03-26 02:59:09,537][root][INFO] - [Round 0] Epoch 38 | Batch 0 | CTC Loss: 3.7240
486
+ [2026-03-26 03:04:54,393][root][INFO] - [Round 0] Epoch 38 | Batch 100 | CTC Loss: 3.7069
487
+ [2026-03-26 03:10:37,599][root][INFO] - [Round 0] Epoch 38 | Batch 200 | CTC Loss: 3.5306
488
+ [2026-03-26 03:16:04,451][root][INFO] - [Round 0] Epoch 38 | Batch 300 | CTC Loss: 3.5254
489
+ [2026-03-26 03:21:27,269][root][INFO] - [Round 0] Epoch 38 | Batch 400 | CTC Loss: 3.2078
490
+ [2026-03-26 03:26:45,656][root][INFO] - [Round 0] Epoch 38 | Batch 500 | CTC Loss: 3.6870
491
+ [2026-03-26 03:32:04,305][root][INFO] - [Round 0] Epoch 38 | Batch 600 | CTC Loss: 3.8600
492
+ [2026-03-26 03:37:45,051][root][INFO] - [Round 0] Epoch 38 | Batch 700 | CTC Loss: 4.5452
493
+ [2026-03-26 03:43:08,303][root][INFO] - [Round 0] Epoch 38 | Batch 800 | CTC Loss: 3.6118
494
+ [2026-03-26 03:48:15,500][root][INFO] - --- [Round 0] Epoch 38 Complete | Avg CTC Loss: 3.5586 ---
495
+ [2026-03-26 03:48:15,516][root][INFO] - New best model saved (loss=3.5586)
496
+ [2026-03-26 03:48:28,617][root][INFO] - [Round 0] Epoch 39 | Batch 0 | CTC Loss: 3.5744
497
+ [2026-03-26 03:54:38,971][root][INFO] - [Round 0] Epoch 39 | Batch 100 | CTC Loss: 3.5238
498
+ [2026-03-26 04:00:11,461][root][INFO] - [Round 0] Epoch 39 | Batch 200 | CTC Loss: 2.8518
499
+ [2026-03-26 04:04:58,015][root][INFO] - [Round 0] Epoch 39 | Batch 300 | CTC Loss: 3.6318
500
+ [2026-03-26 04:10:01,040][root][INFO] - [Round 0] Epoch 39 | Batch 400 | CTC Loss: 3.5391
501
+ [2026-03-26 04:16:26,951][root][INFO] - [Round 0] Epoch 39 | Batch 500 | CTC Loss: 3.3174
502
+ [2026-03-26 04:22:13,284][root][INFO] - [Round 0] Epoch 39 | Batch 600 | CTC Loss: 3.9557
503
+ [2026-03-26 04:28:25,153][root][INFO] - [Round 0] Epoch 39 | Batch 700 | CTC Loss: 2.8948
504
+ [2026-03-26 04:35:06,313][root][INFO] - [Round 0] Epoch 39 | Batch 800 | CTC Loss: 3.2684
505
+ [2026-03-26 04:39:20,493][root][INFO] - --- [Round 0] Epoch 39 Complete | Avg CTC Loss: 3.5731 ---
506
+ [2026-03-26 04:39:39,929][root][INFO] - [Round 0] Epoch 40 | Batch 0 | CTC Loss: 3.9458
507
+ [2026-03-26 04:46:32,485][root][INFO] - [Round 0] Epoch 40 | Batch 100 | CTC Loss: 3.4971
508
+ [2026-03-26 04:52:53,388][root][INFO] - [Round 0] Epoch 40 | Batch 200 | CTC Loss: 3.5464
509
+ [2026-03-26 04:58:10,504][root][INFO] - [Round 0] Epoch 40 | Batch 300 | CTC Loss: 3.9299
510
+ [2026-03-26 05:03:52,530][root][INFO] - [Round 0] Epoch 40 | Batch 400 | CTC Loss: 4.3317
511
+ [2026-03-26 05:09:18,901][root][INFO] - [Round 0] Epoch 40 | Batch 500 | CTC Loss: 3.2882
512
+ [2026-03-26 05:14:05,511][root][INFO] - [Round 0] Epoch 40 | Batch 600 | CTC Loss: 3.4346
513
+ [2026-03-26 05:19:32,921][root][INFO] - [Round 0] Epoch 40 | Batch 700 | CTC Loss: 3.9763
514
+ [2026-03-26 05:24:52,423][root][INFO] - [Round 0] Epoch 40 | Batch 800 | CTC Loss: 3.6846
515
+ [2026-03-26 05:29:37,571][root][INFO] - --- [Round 0] Epoch 40 Complete | Avg CTC Loss: 3.5703 ---
516
+ [2026-03-26 05:29:56,977][root][INFO] - [Round 0] Epoch 41 | Batch 0 | CTC Loss: 3.2045
517
+ [2026-03-26 05:35:32,095][root][INFO] - [Round 0] Epoch 41 | Batch 100 | CTC Loss: 3.7598
518
+ [2026-03-26 05:39:56,441][root][INFO] - [Round 0] Epoch 41 | Batch 200 | CTC Loss: 3.4276
519
+ [2026-03-26 05:45:08,358][root][INFO] - [Round 0] Epoch 41 | Batch 300 | CTC Loss: 3.4696
520
+ [2026-03-26 05:50:26,916][root][INFO] - [Round 0] Epoch 41 | Batch 400 | CTC Loss: 3.4942
521
+ [2026-03-26 05:55:41,958][root][INFO] - [Round 0] Epoch 41 | Batch 500 | CTC Loss: 3.2299
522
+ [2026-03-26 06:00:47,912][root][INFO] - [Round 0] Epoch 41 | Batch 600 | CTC Loss: 3.3863
523
+ [2026-03-26 06:05:28,565][root][INFO] - [Round 0] Epoch 41 | Batch 700 | CTC Loss: 3.1491
524
+ [2026-03-26 06:10:08,044][root][INFO] - [Round 0] Epoch 41 | Batch 800 | CTC Loss: 3.9840
525
+ [2026-03-26 06:15:13,917][root][INFO] - --- [Round 0] Epoch 41 Complete | Avg CTC Loss: 3.5674 ---
526
+ [2026-03-26 06:15:41,229][root][INFO] - [Round 0] Epoch 42 | Batch 0 | CTC Loss: 3.5992
527
+ [2026-03-26 06:21:15,181][root][INFO] - [Round 0] Epoch 42 | Batch 100 | CTC Loss: 3.3661
528
+ [2026-03-26 06:26:33,226][root][INFO] - [Round 0] Epoch 42 | Batch 200 | CTC Loss: 3.5716
529
+ [2026-03-26 06:31:54,532][root][INFO] - [Round 0] Epoch 42 | Batch 300 | CTC Loss: 3.1570
530
+ [2026-03-26 06:37:22,354][root][INFO] - [Round 0] Epoch 42 | Batch 400 | CTC Loss: 4.1496
531
+ [2026-03-26 06:43:02,406][root][INFO] - [Round 0] Epoch 42 | Batch 500 | CTC Loss: 3.6420
532
+ [2026-03-26 06:48:47,514][root][INFO] - [Round 0] Epoch 42 | Batch 600 | CTC Loss: 3.2239
533
+ [2026-03-26 06:54:34,148][root][INFO] - [Round 0] Epoch 42 | Batch 700 | CTC Loss: 3.5875
534
+ [2026-03-26 06:59:17,411][root][INFO] - [Round 0] Epoch 42 | Batch 800 | CTC Loss: 3.1832
535
+ [2026-03-26 07:04:44,791][root][INFO] - --- [Round 0] Epoch 42 Complete | Avg CTC Loss: 3.5460 ---
536
+ [2026-03-26 07:04:44,807][root][INFO] - New best model saved (loss=3.5460)
537
+ [2026-03-26 07:05:00,007][root][INFO] - [Round 0] Epoch 43 | Batch 0 | CTC Loss: 4.2800
538
+ [2026-03-26 07:11:39,527][root][INFO] - [Round 0] Epoch 43 | Batch 100 | CTC Loss: 3.6455
539
+ [2026-03-26 07:17:23,278][root][INFO] - [Round 0] Epoch 43 | Batch 200 | CTC Loss: 3.4507
540
+ [2026-03-26 07:21:57,791][root][INFO] - [Round 0] Epoch 43 | Batch 300 | CTC Loss: 3.0454
541
+ [2026-03-26 07:27:27,832][root][INFO] - [Round 0] Epoch 43 | Batch 400 | CTC Loss: 3.1642
542
+ [2026-03-26 07:32:47,745][root][INFO] - [Round 0] Epoch 43 | Batch 500 | CTC Loss: 3.8408
543
+ [2026-03-26 07:37:24,992][root][INFO] - [Round 0] Epoch 43 | Batch 600 | CTC Loss: 3.3911
544
+ [2026-03-26 07:42:04,089][root][INFO] - [Round 0] Epoch 43 | Batch 700 | CTC Loss: 2.9972
545
+ [2026-03-26 07:46:52,825][root][INFO] - [Round 0] Epoch 43 | Batch 800 | CTC Loss: 3.8514
546
+ [2026-03-26 07:51:43,875][root][INFO] - --- [Round 0] Epoch 43 Complete | Avg CTC Loss: 3.5516 ---
547
+ [2026-03-26 07:52:07,458][root][INFO] - [Round 0] Epoch 44 | Batch 0 | CTC Loss: 3.4208