Hugging Face's logo

al1808th
/
stanza-digphil

Model card Files
xet
Community
Albin Thörn Cleland commited on
Commit
0b4690a
·
1 Parent(s): b01c18c

new models

Browse files
Files changed (33) hide show
  1. logs/latest.txt +1 -1
  2. logs/log_conll17.pt_sv_diachron_de_lit_20251203_154509.txt +740 -0
  3. logs/log_conll17.pt_sv_diachron_is_20251203_164326.txt +911 -0
  4. logs/log_conll17.pt_sv_diachron_is_20251203_221228.txt +147 -0
  5. logs/log_conll17.pt_sv_diachron_nn_20251203_160846.txt +1146 -0
  6. logs/log_diachronic.pt_sv_diachron_de_lit_20251203_150832.txt +188 -0
  7. logs/log_diachronic.pt_sv_diachron_de_lit_20251203_151554.txt +650 -0
  8. logs/log_diachronic.pt_sv_diachron_is_20251203_214751.txt +757 -0
  9. logs/log_diachronic.pt_sv_diachron_nn_20251203_003001.txt +577 -0
  10. saved_models/depparse/conll17_de_lit/sv_diachronic_charlm_parser.pt +3 -0
  11. saved_models/depparse/conll17_de_lit/sv_diachronic_charlm_parser_checkpoint.pt +3 -0
  12. saved_models/depparse/conll17_icepahc/sv_diachronic_charlm_parser.pt +3 -0
  13. saved_models/depparse/conll17_icepahc/sv_diachronic_charlm_parser_checkpoint.pt +3 -0
  14. saved_models/depparse/conll17_isPUD-pahc-gc/sv_diachronic_charlm_parser.pt +3 -0
  15. saved_models/depparse/conll17_isPUD-pahc-gc/sv_diachronic_charlm_parser_checkpoint.pt +3 -0
  16. saved_models/depparse/conll17_isPUD/sv_diachronic_charlm_parser.pt +3 -0
  17. saved_models/depparse/conll17_isPUD/sv_diachronic_charlm_parser_checkpoint.pt +3 -0
  18. saved_models/depparse/conll17_nn/sv_diachronic_charlm_parser.pt +3 -0
  19. saved_models/depparse/conll17_nn/sv_diachronic_charlm_parser_checkpoint.pt +3 -0
  20. saved_models/depparse/kubhist2-isPUD/sv_diachronic_charlm_parser.pt +3 -0
  21. saved_models/depparse/kubhist2-isPUD/sv_diachronic_charlm_parser_checkpoint.pt +3 -0
  22. saved_models/depparse/kubhist2_de_lit/sv_diachronic_charlm_parser.pt +3 -0
  23. saved_models/depparse/kubhist2_de_lit/sv_diachronic_charlm_parser_checkpoint.pt +3 -0
  24. saved_models/depparse/kubhist2_is-modern/sv_diachronic_charlm_parser.pt +3 -0
  25. saved_models/depparse/kubhist2_is-modern/sv_diachronic_charlm_parser_checkpoint.pt +3 -0
  26. saved_models/depparse/kubhist2_nn/sv_diachronic_charlm_parser.pt +3 -0
  27. saved_models/depparse/kubhist2_nn/sv_diachronic_charlm_parser_checkpoint.pt +3 -0
  28. ud-treebanks-is/is_gc-ud-dev.conllu +3 -0
  29. ud-treebanks-is/is_gc-ud-test.conllu +3 -0
  30. ud-treebanks-is/is_gc-ud-train.conllu +3 -0
  31. ud-treebanks-is/is_modern-ud-dev.conllu +3 -0
  32. ud-treebanks-is/is_modern-ud-test.conllu +3 -0
  33. ud-treebanks-is/is_modern-ud-train.conllu +3 -0
logs/latest.txt CHANGED
@@ -1 +1 @@
1
- log_diachronic.pt_sv_diachron_nn_20251203_001047.txt
 
1
+ log_diachronic.pt_sv_diachron_is_20251203_214751.txt
logs/log_conll17.pt_sv_diachron_de_lit_20251203_154509.txt ADDED
@@ -0,0 +1,740 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ === LOGFILE: logs/log_conll17.pt_sv_diachron_de_lit_20251203_154509.txt ===
2
+ Language codes: sv diachron de_lit
3
+ Using pretrained model: conll17.pt
4
+
5
+ Running: python prepare-train-val-test.py sv diachron de_lit
6
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
7
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
8
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
9
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
10
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
11
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
12
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
13
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
14
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
15
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
16
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
17
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-de_lit/de_lit-ud-test.conllu
18
+ Including DigPhil MACHINE in TRAIN (minus gold)…
19
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
20
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
21
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
22
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
23
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
24
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
25
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
26
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
27
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
28
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
29
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
30
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
31
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
32
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
33
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
34
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
35
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
36
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
37
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
38
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
39
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
40
+ Cleaning TRAIN...
41
+ [REMOVED] sent_id=6 ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
42
+ [REMOVED] sent_id=7_8 ERRORS=['Multiple roots found: [5, 10]']
43
+ [REMOVED] sent_id=30_31 ERRORS=['Multiple roots found: [3, 18]']
44
+ [REMOVED] sent_id=35 ERRORS=['Line 36: Invalid token ID or head']
45
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [1, 5]']
46
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [7, 20]']
47
+ [REMOVED] sent_id=8_9 ERRORS=['Multiple roots found: [24, 57]']
48
+ [REMOVED] sent_id=12_13 ERRORS=['Multiple roots found: [11, 16]']
49
+ [REMOVED] sent_id=124_split2 ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
50
+ [REMOVED] sent_id=396 ERRORS=['Token 2: Missing form']
51
+ [REMOVED] sent_id=416 ERRORS=['Token 2: Missing form']
52
+ [REMOVED] sent_id=589 ERRORS=['Token 2: Missing form']
53
+ [REMOVED] sent_id=909 ERRORS=['Token 2: Missing form']
54
+ [REMOVED] sent_id=912 ERRORS=['Token 2: Missing form']
55
+ [REMOVED] sent_id=3_split1 ERRORS=['Multiple roots found: [4, 15, 17]']
56
+ [REMOVED] sent_id=3_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
57
+ [REMOVED] sent_id=3_4 ERRORS=['Multiple roots found: [1, 5]']
58
+ [REMOVED] sent_id=5_6 ERRORS=['Multiple roots found: [3, 24]']
59
+ [REMOVED] sent_id=11_12_13 ERRORS=['Multiple roots found: [5, 17, 25]']
60
+ [REMOVED] sent_id=119 ERRORS=['Token 2: Missing form']
61
+ [REMOVED] sent_id=179 ERRORS=['Token 2: Missing form']
62
+ [REMOVED] sent_id=188 ERRORS=['Token 2: Missing form']
63
+ [REMOVED] sent_id=223 ERRORS=['Token 2: Missing form']
64
+ [REMOVED] sent_id=268 ERRORS=['Token 2: Missing form']
65
+ [REMOVED] sent_id=325 ERRORS=['Token 2: Missing form']
66
+ [REMOVED] sent_id=388 ERRORS=['Token 2: Missing form']
67
+ [REMOVED] sent_id=399 ERRORS=['Token 2: Missing form']
68
+ [REMOVED] sent_id=475 ERRORS=['Token 2: Missing form']
69
+ [REMOVED] sent_id=505 ERRORS=['Token 2: Missing form']
70
+ [REMOVED] sent_id=520 ERRORS=['Token 2: Missing form']
71
+ [REMOVED] sent_id=562 ERRORS=['Token 2: Missing form']
72
+ [REMOVED] sent_id=669 ERRORS=['Token 2: Missing form']
73
+ [REMOVED] sent_id=711 ERRORS=['Token 2: Missing form']
74
+ [REMOVED] sent_id=731 ERRORS=['Token 2: Missing form']
75
+ [REMOVED] sent_id=867 ERRORS=['Token 2: Missing form']
76
+ [REMOVED] sent_id=884 ERRORS=['Token 2: Missing form']
77
+ [REMOVED] sent_id=923 ERRORS=['Token 2: Missing form']
78
+ [REMOVED] sent_id=939 ERRORS=['Token 2: Missing form']
79
+ [REMOVED] sent_id=1086 ERRORS=['Token 2: Missing form']
80
+ [REMOVED] sent_id=1179 ERRORS=['Token 2: Missing form']
81
+ [REMOVED] sent_id=1251 ERRORS=['Token 2: Missing form']
82
+ [REMOVED] sent_id=1345 ERRORS=['Token 2: Missing form']
83
+ [REMOVED] sent_id=1459 ERRORS=['Token 2: Missing form']
84
+ [REMOVED] sent_id=1656 ERRORS=['Token 2: Missing form']
85
+ [REMOVED] sent_id=1669 ERRORS=['Token 2: Missing form']
86
+ [REMOVED] sent_id=87_88 ERRORS=['Multiple roots found: [3, 6]']
87
+ [REMOVED] sent_id=65_split2_66_split2 ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
88
+ [REMOVED] sent_id=25 ERRORS=['Token 2: Missing form']
89
+ [REMOVED] sent_id=136 ERRORS=['Token 2: Missing form']
90
+ [REMOVED] sent_id=208 ERRORS=['Token 2: Missing form']
91
+ [REMOVED] sent_id=230 ERRORS=['Token 2: Missing form']
92
+ [REMOVED] sent_id=245 ERRORS=['Token 2: Missing form']
93
+ [REMOVED] sent_id=276 ERRORS=['Token 2: Missing form']
94
+ [REMOVED] sent_id=320 ERRORS=['Token 2: Missing form']
95
+ [REMOVED] sent_id=366 ERRORS=['Token 2: Missing form']
96
+ [REMOVED] sent_id=519 ERRORS=['Token 2: Missing form']
97
+ [REMOVED] sent_id=569 ERRORS=['Token 2: Missing form']
98
+ [REMOVED] sent_id=50_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
99
+ [REMOVED] sent_id=53_54 ERRORS=['Multiple roots found: [27, 91]']
100
+ [REMOVED] sent_id=55_56_57 ERRORS=['Multiple roots found: [2, 4, 13]']
101
+ [REMOVED] sent_id=17_split1 ERRORS=['Multiple roots found: [2, 14, 17]']
102
+ [REMOVED] sent_id=17_split2 ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
103
+ [REMOVED] sent_id=19_split1 ERRORS=['Multiple roots found: [3, 31]']
104
+ Cleaning DEV...
105
+ [REMOVED] sent_id=33 ERRORS=['Token 15: Missing deprel']
106
+ Cleaning TEST...
107
+ Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (68172 valid sentences)
108
+ Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
109
+ Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
110
+ Done.
111
+ Sourcing scripts/config_alvis.sh
112
+ Running stanza dataset preparation…
113
+ 2025-12-03 15:45:18 INFO: Datasets program called with:
114
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
115
+ 2025-12-03 15:45:18 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
116
+
117
+ 2025-12-03 15:45:18 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
118
+ 2025-12-03 15:45:18 DEBUG: Processing parameter "processors"...
119
+ 2025-12-03 15:45:18 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
120
+ 2025-12-03 15:45:18 INFO: Downloading these customized packages for language: sv (Swedish)...
121
+ =======================
122
+ | Processor | Package |
123
+ -----------------------
124
+ =======================
125
+
126
+ 2025-12-03 15:45:18 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
127
+ 2025-12-03 15:45:18 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
128
+ 2025-12-03 15:45:18 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
129
+ 2025-12-03 15:45:18 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
130
+ Augmented 188 quotes: Counter({'«»': 23, '»«': 22, '″″': 22, '""': 20, '““': 20, '„”': 19, '「」': 18, '《》': 16, '„“': 15, '””': 13})
131
+ 2025-12-03 15:45:21 INFO: Running tagger to retag /local/tmp.5440223/tmplttwhhzq/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
132
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmplttwhhzq/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
133
+ 2025-12-03 15:45:22 INFO: Running tagger in predict mode
134
+ 2025-12-03 15:45:22 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
135
+ 2025-12-03 15:45:23 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
136
+ 2025-12-03 15:45:23 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
137
+ 2025-12-03 15:45:23 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
138
+ 2025-12-03 15:45:23 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
139
+ 2025-12-03 15:45:24 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
140
+ 2025-12-03 15:45:26 INFO: Loading data with batch size 250...
141
+ 2025-12-03 15:46:06 INFO: Start evaluation...
142
+ 2025-12-03 15:49:44 INFO: UPOS XPOS UFeats AllTags
143
+ 2025-12-03 15:49:44 INFO: 96.83 85.29 95.46 84.47
144
+ 2025-12-03 15:49:44 INFO: POS Tagger score: sv_diachronic 84.47
145
+ 2025-12-03 15:49:45 INFO: Running tagger to retag /local/tmp.5440223/tmplttwhhzq/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
146
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmplttwhhzq/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
147
+ 2025-12-03 15:49:45 INFO: Running tagger in predict mode
148
+ 2025-12-03 15:49:45 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
149
+ 2025-12-03 15:49:47 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
150
+ 2025-12-03 15:49:47 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
151
+ 2025-12-03 15:49:47 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
152
+ 2025-12-03 15:49:47 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
153
+ 2025-12-03 15:49:47 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
154
+ 2025-12-03 15:49:47 INFO: Loading data with batch size 250...
155
+ 2025-12-03 15:49:47 INFO: Start evaluation...
156
+ 2025-12-03 15:49:47 INFO: UPOS XPOS UFeats AllTags
157
+ 2025-12-03 15:49:47 INFO: 93.32 90.84 93.32 85.64
158
+ 2025-12-03 15:49:47 INFO: POS Tagger score: sv_diachronic 85.64
159
+ 2025-12-03 15:49:47 INFO: Running tagger to retag /local/tmp.5440223/tmplttwhhzq/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
160
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmplttwhhzq/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
161
+ 2025-12-03 15:49:47 INFO: Running tagger in predict mode
162
+ 2025-12-03 15:49:47 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
163
+ 2025-12-03 15:49:49 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
164
+ 2025-12-03 15:49:49 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
165
+ 2025-12-03 15:49:49 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
166
+ 2025-12-03 15:49:49 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
167
+ 2025-12-03 15:49:49 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
168
+ 2025-12-03 15:49:49 INFO: Loading data with batch size 250...
169
+ 2025-12-03 15:49:49 INFO: Start evaluation...
170
+ 2025-12-03 15:49:50 INFO: UPOS XPOS UFeats AllTags
171
+ 2025-12-03 15:49:50 INFO: 93.14 96.78 95.32 90.28
172
+ 2025-12-03 15:49:50 INFO: POS Tagger score: sv_diachronic 90.28
173
+ Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
174
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5440223/tmplttwhhzq/sv_diachronic.train.gold.conllu
175
+ Swapped 'w1, w2' for 'w1 ,w2' 141 times
176
+ Added 573 new sentences with asdf, zzzz -> asdf,zzzz
177
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5440223/tmplttwhhzq/sv_diachronic.dev.gold.conllu
178
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5440223/tmplttwhhzq/sv_diachronic.test.gold.conllu
179
+ Running stanza dependency parser training…
180
+ 2025-12-03 15:50:04 INFO: Training program called with:
181
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt --batch_size 32 --dropout 0.33
182
+ 2025-12-03 15:50:04 DEBUG: UD_Swedish-diachronic: sv_diachronic
183
+ 2025-12-03 15:50:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
184
+ 2025-12-03 15:50:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
185
+ 2025-12-03 15:50:04 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
186
+ 2025-12-03 15:50:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
187
+ 2025-12-03 15:50:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
188
+ 2025-12-03 15:50:04 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
189
+ 2025-12-03 15:50:04 INFO: Running parser in train mode
190
+ 2025-12-03 15:50:04 INFO: Using pretrained contextualized char embedding
191
+ 2025-12-03 15:50:04 INFO: Loading data with batch size 32...
192
+ 2025-12-03 15:50:11 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 68745
193
+ 2025-12-03 15:50:11 INFO: Original data size: 68745
194
+ 2025-12-03 15:50:12 INFO: Augmented data size: 69046
195
+ 2025-12-03 15:50:31 WARNING: sv_diachronic is not a known dataset. Examining the data to choose which xpos vocab to use
196
+ 2025-12-03 15:50:31 INFO: Original length = 69046
197
+ 2025-12-03 15:50:31 INFO: Filtered length = 69046
198
+ 2025-12-03 15:50:47 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
199
+ 2025-12-03 15:50:55 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
200
+ 2025-12-03 15:51:09 DEBUG: 40457 batches created.
201
+ 2025-12-03 15:51:09 DEBUG: 9 batches created.
202
+ 2025-12-03 15:51:09 INFO: Training parser...
203
+ 2025-12-03 15:51:09 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
204
+ 2025-12-03 15:51:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
205
+ 2025-12-03 15:51:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
206
+ 2025-12-03 15:51:10 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
207
+ 2025-12-03 15:51:17 INFO: Finished STEP 20/50000, loss = 6.791992 (0.200 sec/batch), lr: 0.003000
208
+ 2025-12-03 15:51:21 INFO: Finished STEP 40/50000, loss = 5.211706 (0.167 sec/batch), lr: 0.003000
209
+ 2025-12-03 15:51:24 INFO: Finished STEP 60/50000, loss = 5.668643 (0.147 sec/batch), lr: 0.003000
210
+ 2025-12-03 15:51:27 INFO: Finished STEP 80/50000, loss = 4.197357 (0.151 sec/batch), lr: 0.003000
211
+ 2025-12-03 15:51:30 INFO: Finished STEP 100/50000, loss = 5.356755 (0.129 sec/batch), lr: 0.003000
212
+ 2025-12-03 15:51:30 INFO: Evaluating on dev set...
213
+ 2025-12-03 15:51:31 INFO: LAS MLAS BLEX
214
+ 2025-12-03 15:51:31 INFO: 39.85 26.44 33.26
215
+ 2025-12-03 15:51:31 INFO: step 100: train_loss = 8.587716, dev_score = 0.3985
216
+ 2025-12-03 15:51:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
217
+ 2025-12-03 15:51:31 INFO: new best model saved.
218
+ 2025-12-03 15:51:32 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
219
+ 2025-12-03 15:51:32 INFO: new model checkpoint saved.
220
+ 2025-12-03 15:51:34 INFO: Finished STEP 120/50000, loss = 4.996027 (0.120 sec/batch), lr: 0.003000
221
+ 2025-12-03 15:51:37 INFO: Finished STEP 140/50000, loss = 5.125143 (0.131 sec/batch), lr: 0.003000
222
+ 2025-12-03 15:51:40 INFO: Finished STEP 160/50000, loss = 4.139769 (0.138 sec/batch), lr: 0.003000
223
+ 2025-12-03 15:51:42 INFO: Finished STEP 180/50000, loss = 4.071211 (0.131 sec/batch), lr: 0.003000
224
+ 2025-12-03 15:51:45 INFO: Finished STEP 200/50000, loss = 4.302421 (0.121 sec/batch), lr: 0.003000
225
+ 2025-12-03 15:51:45 INFO: Evaluating on dev set...
226
+ 2025-12-03 15:51:45 INFO: LAS MLAS BLEX
227
+ 2025-12-03 15:51:45 INFO: 50.00 35.44 40.93
228
+ 2025-12-03 15:51:45 INFO: step 200: train_loss = 4.776229, dev_score = 0.5000
229
+ 2025-12-03 15:51:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
230
+ 2025-12-03 15:51:45 INFO: new best model saved.
231
+ 2025-12-03 15:51:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
232
+ 2025-12-03 15:51:46 INFO: new model checkpoint saved.
233
+ 2025-12-03 15:51:48 INFO: Finished STEP 220/50000, loss = 4.349181 (0.124 sec/batch), lr: 0.003000
234
+ 2025-12-03 15:51:51 INFO: Finished STEP 240/50000, loss = 5.862440 (0.125 sec/batch), lr: 0.003000
235
+ 2025-12-03 15:51:53 INFO: Finished STEP 260/50000, loss = 4.765276 (0.118 sec/batch), lr: 0.003000
236
+ 2025-12-03 15:51:56 INFO: Finished STEP 280/50000, loss = 4.617913 (0.116 sec/batch), lr: 0.003000
237
+ 2025-12-03 15:51:58 INFO: Finished STEP 300/50000, loss = 5.353960 (0.113 sec/batch), lr: 0.003000
238
+ 2025-12-03 15:51:58 INFO: Evaluating on dev set...
239
+ 2025-12-03 15:51:58 INFO: LAS MLAS BLEX
240
+ 2025-12-03 15:51:58 INFO: 51.49 38.56 43.22
241
+ 2025-12-03 15:51:58 INFO: step 300: train_loss = 4.421819, dev_score = 0.5149
242
+ 2025-12-03 15:51:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
243
+ 2025-12-03 15:51:59 INFO: new best model saved.
244
+ 2025-12-03 15:51:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
245
+ 2025-12-03 15:51:59 INFO: new model checkpoint saved.
246
+ 2025-12-03 15:52:02 INFO: Finished STEP 320/50000, loss = 5.069269 (0.121 sec/batch), lr: 0.003000
247
+ 2025-12-03 15:52:04 INFO: Finished STEP 340/50000, loss = 3.614627 (0.112 sec/batch), lr: 0.003000
248
+ 2025-12-03 15:52:06 INFO: Finished STEP 360/50000, loss = 3.878788 (0.115 sec/batch), lr: 0.003000
249
+ 2025-12-03 15:52:08 INFO: Finished STEP 380/50000, loss = 4.367438 (0.100 sec/batch), lr: 0.003000
250
+ 2025-12-03 15:52:10 INFO: Finished STEP 400/50000, loss = 4.081972 (0.107 sec/batch), lr: 0.003000
251
+ 2025-12-03 15:52:10 INFO: Evaluating on dev set...
252
+ 2025-12-03 15:52:11 INFO: LAS MLAS BLEX
253
+ 2025-12-03 15:52:11 INFO: 55.69 43.55 46.51
254
+ 2025-12-03 15:52:11 INFO: step 400: train_loss = 4.280971, dev_score = 0.5569
255
+ 2025-12-03 15:52:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
256
+ 2025-12-03 15:52:11 INFO: new best model saved.
257
+ 2025-12-03 15:52:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
258
+ 2025-12-03 15:52:12 INFO: new model checkpoint saved.
259
+ 2025-12-03 15:52:14 INFO: Finished STEP 420/50000, loss = 3.892664 (0.110 sec/batch), lr: 0.003000
260
+ 2025-12-03 15:52:16 INFO: Finished STEP 440/50000, loss = 3.998952 (0.100 sec/batch), lr: 0.003000
261
+ 2025-12-03 15:52:18 INFO: Finished STEP 460/50000, loss = 5.833274 (0.112 sec/batch), lr: 0.003000
262
+ 2025-12-03 15:52:20 INFO: Finished STEP 480/50000, loss = 3.561453 (0.106 sec/batch), lr: 0.003000
263
+ 2025-12-03 15:52:22 INFO: Finished STEP 500/50000, loss = 3.267099 (0.106 sec/batch), lr: 0.003000
264
+ 2025-12-03 15:52:22 INFO: Evaluating on dev set...
265
+ 2025-12-03 15:52:23 INFO: LAS MLAS BLEX
266
+ 2025-12-03 15:52:23 INFO: 55.45 44.30 48.95
267
+ 2025-12-03 15:52:23 INFO: step 500: train_loss = 4.074671, dev_score = 0.5545
268
+ 2025-12-03 15:52:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
269
+ 2025-12-03 15:52:24 INFO: new model checkpoint saved.
270
+ 2025-12-03 15:52:26 INFO: Finished STEP 520/50000, loss = 3.987905 (0.101 sec/batch), lr: 0.003000
271
+ 2025-12-03 15:52:28 INFO: Finished STEP 540/50000, loss = 4.040174 (0.105 sec/batch), lr: 0.003000
272
+ 2025-12-03 15:52:30 INFO: Finished STEP 560/50000, loss = 5.425183 (0.096 sec/batch), lr: 0.003000
273
+ 2025-12-03 15:52:32 INFO: Finished STEP 580/50000, loss = 3.171549 (0.098 sec/batch), lr: 0.003000
274
+ 2025-12-03 15:52:34 INFO: Finished STEP 600/50000, loss = 5.054599 (0.087 sec/batch), lr: 0.003000
275
+ 2025-12-03 15:52:34 INFO: Evaluating on dev set...
276
+ 2025-12-03 15:52:34 INFO: LAS MLAS BLEX
277
+ 2025-12-03 15:52:34 INFO: 59.41 50.85 53.39
278
+ 2025-12-03 15:52:34 INFO: step 600: train_loss = 3.989031, dev_score = 0.5941
279
+ 2025-12-03 15:52:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
280
+ 2025-12-03 15:52:34 INFO: new best model saved.
281
+ 2025-12-03 15:52:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
282
+ 2025-12-03 15:52:35 INFO: new model checkpoint saved.
283
+ 2025-12-03 15:52:37 INFO: Finished STEP 620/50000, loss = 5.185298 (0.107 sec/batch), lr: 0.003000
284
+ 2025-12-03 15:52:39 INFO: Finished STEP 640/50000, loss = 3.282787 (0.098 sec/batch), lr: 0.003000
285
+ 2025-12-03 15:52:41 INFO: Finished STEP 660/50000, loss = 7.106239 (0.095 sec/batch), lr: 0.003000
286
+ 2025-12-03 15:52:43 INFO: Finished STEP 680/50000, loss = 3.663255 (0.099 sec/batch), lr: 0.003000
287
+ 2025-12-03 15:52:45 INFO: Finished STEP 700/50000, loss = 4.011485 (0.098 sec/batch), lr: 0.003000
288
+ 2025-12-03 15:52:45 INFO: Evaluating on dev set...
289
+ 2025-12-03 15:52:46 INFO: LAS MLAS BLEX
290
+ 2025-12-03 15:52:46 INFO: 59.16 46.74 52.21
291
+ 2025-12-03 15:52:46 INFO: step 700: train_loss = 3.973323, dev_score = 0.5916
292
+ 2025-12-03 15:52:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
293
+ 2025-12-03 15:52:46 INFO: new model checkpoint saved.
294
+ 2025-12-03 15:52:48 INFO: Finished STEP 720/50000, loss = 4.034673 (0.094 sec/batch), lr: 0.003000
295
+ 2025-12-03 15:52:50 INFO: Finished STEP 740/50000, loss = 4.719307 (0.099 sec/batch), lr: 0.003000
296
+ 2025-12-03 15:52:52 INFO: Finished STEP 760/50000, loss = 3.175733 (0.089 sec/batch), lr: 0.003000
297
+ 2025-12-03 15:52:54 INFO: Finished STEP 780/50000, loss = 3.158538 (0.098 sec/batch), lr: 0.003000
298
+ 2025-12-03 15:52:56 INFO: Finished STEP 800/50000, loss = 2.881503 (0.099 sec/batch), lr: 0.003000
299
+ 2025-12-03 15:52:56 INFO: Evaluating on dev set...
300
+ 2025-12-03 15:52:56 INFO: LAS MLAS BLEX
301
+ 2025-12-03 15:52:56 INFO: 57.92 46.19 49.15
302
+ 2025-12-03 15:52:56 INFO: step 800: train_loss = 3.763542, dev_score = 0.5792
303
+ 2025-12-03 15:52:57 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
304
+ 2025-12-03 15:52:57 INFO: new model checkpoint saved.
305
+ 2025-12-03 15:52:59 INFO: Finished STEP 820/50000, loss = 4.045841 (0.087 sec/batch), lr: 0.003000
306
+ 2025-12-03 15:53:01 INFO: Finished STEP 840/50000, loss = 3.735970 (0.088 sec/batch), lr: 0.003000
307
+ 2025-12-03 15:53:02 INFO: Finished STEP 860/50000, loss = 4.020472 (0.080 sec/batch), lr: 0.003000
308
+ 2025-12-03 15:53:04 INFO: Finished STEP 880/50000, loss = 5.290604 (0.091 sec/batch), lr: 0.003000
309
+ 2025-12-03 15:53:06 INFO: Finished STEP 900/50000, loss = 3.524802 (0.089 sec/batch), lr: 0.003000
310
+ 2025-12-03 15:53:06 INFO: Evaluating on dev set...
311
+ 2025-12-03 15:53:07 INFO: LAS MLAS BLEX
312
+ 2025-12-03 15:53:07 INFO: 56.93 45.67 49.47
313
+ 2025-12-03 15:53:07 INFO: step 900: train_loss = 3.948068, dev_score = 0.5693
314
+ 2025-12-03 15:53:07 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
315
+ 2025-12-03 15:53:07 INFO: new model checkpoint saved.
316
+ 2025-12-03 15:53:09 INFO: Finished STEP 920/50000, loss = 3.348033 (0.088 sec/batch), lr: 0.003000
317
+ 2025-12-03 15:53:11 INFO: Finished STEP 940/50000, loss = 3.860394 (0.091 sec/batch), lr: 0.003000
318
+ 2025-12-03 15:53:13 INFO: Finished STEP 960/50000, loss = 3.459932 (0.082 sec/batch), lr: 0.003000
319
+ 2025-12-03 15:53:15 INFO: Finished STEP 980/50000, loss = 4.335596 (0.085 sec/batch), lr: 0.003000
320
+ 2025-12-03 15:53:17 INFO: Finished STEP 1000/50000, loss = 3.108686 (0.089 sec/batch), lr: 0.003000
321
+ 2025-12-03 15:53:17 INFO: Evaluating on dev set...
322
+ 2025-12-03 15:53:17 INFO: LAS MLAS BLEX
323
+ 2025-12-03 15:53:17 INFO: 60.40 51.27 53.81
324
+ 2025-12-03 15:53:17 INFO: step 1000: train_loss = 3.976391, dev_score = 0.6040
325
+ 2025-12-03 15:53:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
326
+ 2025-12-03 15:53:17 INFO: new best model saved.
327
+ 2025-12-03 15:53:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
328
+ 2025-12-03 15:53:18 INFO: new model checkpoint saved.
329
+ 2025-12-03 15:53:20 INFO: Finished STEP 1020/50000, loss = 6.081649 (0.089 sec/batch), lr: 0.003000
330
+ 2025-12-03 15:53:21 INFO: Finished STEP 1040/50000, loss = 3.454382 (0.089 sec/batch), lr: 0.003000
331
+ 2025-12-03 15:53:23 INFO: Finished STEP 1060/50000, loss = 3.093398 (0.085 sec/batch), lr: 0.003000
332
+ 2025-12-03 15:53:25 INFO: Finished STEP 1080/50000, loss = 3.780082 (0.088 sec/batch), lr: 0.003000
333
+ 2025-12-03 15:53:27 INFO: Finished STEP 1100/50000, loss = 3.758047 (0.075 sec/batch), lr: 0.003000
334
+ 2025-12-03 15:53:27 INFO: Evaluating on dev set...
335
+ 2025-12-03 15:53:27 INFO: LAS MLAS BLEX
336
+ 2025-12-03 15:53:27 INFO: 59.41 49.05 52.85
337
+ 2025-12-03 15:53:27 INFO: step 1100: train_loss = 3.779099, dev_score = 0.5941
338
+ 2025-12-03 15:53:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
339
+ 2025-12-03 15:53:28 INFO: new model checkpoint saved.
340
+ 2025-12-03 15:53:30 INFO: Finished STEP 1120/50000, loss = 3.774802 (0.088 sec/batch), lr: 0.003000
341
+ 2025-12-03 15:53:31 INFO: Finished STEP 1140/50000, loss = 2.977976 (0.089 sec/batch), lr: 0.003000
342
+ 2025-12-03 15:53:33 INFO: Finished STEP 1160/50000, loss = 3.838738 (0.078 sec/batch), lr: 0.003000
343
+ 2025-12-03 15:53:35 INFO: Finished STEP 1180/50000, loss = 4.660880 (0.080 sec/batch), lr: 0.003000
344
+ 2025-12-03 15:53:36 INFO: Finished STEP 1200/50000, loss = 2.489745 (0.088 sec/batch), lr: 0.003000
345
+ 2025-12-03 15:53:36 INFO: Evaluating on dev set...
346
+ 2025-12-03 15:53:37 INFO: LAS MLAS BLEX
347
+ 2025-12-03 15:53:37 INFO: 58.66 48.31 51.69
348
+ 2025-12-03 15:53:37 INFO: step 1200: train_loss = 3.720397, dev_score = 0.5866
349
+ 2025-12-03 15:53:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
350
+ 2025-12-03 15:53:37 INFO: new model checkpoint saved.
351
+ 2025-12-03 15:53:39 INFO: Finished STEP 1220/50000, loss = 3.247226 (0.090 sec/batch), lr: 0.003000
352
+ 2025-12-03 15:53:41 INFO: Finished STEP 1240/50000, loss = 4.093720 (0.079 sec/batch), lr: 0.003000
353
+ 2025-12-03 15:53:43 INFO: Finished STEP 1260/50000, loss = 2.970687 (0.080 sec/batch), lr: 0.003000
354
+ 2025-12-03 15:53:44 INFO: Finished STEP 1280/50000, loss = 3.959898 (0.087 sec/batch), lr: 0.003000
355
+ 2025-12-03 15:53:46 INFO: Finished STEP 1300/50000, loss = 3.890834 (0.081 sec/batch), lr: 0.003000
356
+ 2025-12-03 15:53:46 INFO: Evaluating on dev set...
357
+ 2025-12-03 15:53:47 INFO: LAS MLAS BLEX
358
+ 2025-12-03 15:53:47 INFO: 57.67 50.00 52.97
359
+ 2025-12-03 15:53:47 INFO: step 1300: train_loss = 3.945667, dev_score = 0.5767
360
+ 2025-12-03 15:53:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
361
+ 2025-12-03 15:53:47 INFO: new model checkpoint saved.
362
+ 2025-12-03 15:53:49 INFO: Finished STEP 1320/50000, loss = 4.299366 (0.076 sec/batch), lr: 0.003000
363
+ 2025-12-03 15:53:51 INFO: Finished STEP 1340/50000, loss = 3.080094 (0.082 sec/batch), lr: 0.003000
364
+ 2025-12-03 15:53:52 INFO: Finished STEP 1360/50000, loss = 5.326971 (0.087 sec/batch), lr: 0.003000
365
+ 2025-12-03 15:53:54 INFO: Finished STEP 1380/50000, loss = 4.148820 (0.083 sec/batch), lr: 0.003000
366
+ 2025-12-03 15:53:56 INFO: Finished STEP 1400/50000, loss = 3.131835 (0.085 sec/batch), lr: 0.003000
367
+ 2025-12-03 15:53:56 INFO: Evaluating on dev set...
368
+ 2025-12-03 15:53:56 INFO: LAS MLAS BLEX
369
+ 2025-12-03 15:53:56 INFO: 59.16 49.47 52.85
370
+ 2025-12-03 15:53:56 INFO: step 1400: train_loss = 3.828261, dev_score = 0.5916
371
+ 2025-12-03 15:53:57 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
372
+ 2025-12-03 15:53:57 INFO: new model checkpoint saved.
373
+ 2025-12-03 15:53:58 INFO: Finished STEP 1420/50000, loss = 3.877927 (0.079 sec/batch), lr: 0.003000
374
+ 2025-12-03 15:54:00 INFO: Finished STEP 1440/50000, loss = 2.828221 (0.077 sec/batch), lr: 0.003000
375
+ 2025-12-03 15:54:02 INFO: Finished STEP 1460/50000, loss = 2.771763 (0.088 sec/batch), lr: 0.003000
376
+ 2025-12-03 15:54:03 INFO: Finished STEP 1480/50000, loss = 3.887072 (0.082 sec/batch), lr: 0.003000
377
+ 2025-12-03 15:54:05 INFO: Finished STEP 1500/50000, loss = 3.089223 (0.080 sec/batch), lr: 0.003000
378
+ 2025-12-03 15:54:05 INFO: Evaluating on dev set...
379
+ 2025-12-03 15:54:05 INFO: LAS MLAS BLEX
380
+ 2025-12-03 15:54:05 INFO: 60.89 49.47 54.12
381
+ 2025-12-03 15:54:05 INFO: step 1500: train_loss = 3.662534, dev_score = 0.6089
382
+ 2025-12-03 15:54:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
383
+ 2025-12-03 15:54:06 INFO: new best model saved.
384
+ 2025-12-03 15:54:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
385
+ 2025-12-03 15:54:06 INFO: new model checkpoint saved.
386
+ 2025-12-03 15:54:08 INFO: Finished STEP 1520/50000, loss = 3.424534 (0.079 sec/batch), lr: 0.003000
387
+ 2025-12-03 15:54:10 INFO: Finished STEP 1540/50000, loss = 4.257630 (0.081 sec/batch), lr: 0.003000
388
+ 2025-12-03 15:54:11 INFO: Finished STEP 1560/50000, loss = 4.004986 (0.078 sec/batch), lr: 0.003000
389
+ 2025-12-03 15:54:13 INFO: Finished STEP 1580/50000, loss = 4.259016 (0.076 sec/batch), lr: 0.003000
390
+ 2025-12-03 15:54:14 INFO: Finished STEP 1600/50000, loss = 4.886831 (0.071 sec/batch), lr: 0.003000
391
+ 2025-12-03 15:54:14 INFO: Evaluating on dev set...
392
+ 2025-12-03 15:54:15 INFO: LAS MLAS BLEX
393
+ 2025-12-03 15:54:15 INFO: 58.66 49.47 52.01
394
+ 2025-12-03 15:54:15 INFO: step 1600: train_loss = 3.940190, dev_score = 0.5866
395
+ 2025-12-03 15:54:15 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
396
+ 2025-12-03 15:54:15 INFO: new model checkpoint saved.
397
+ 2025-12-03 15:54:17 INFO: Finished STEP 1620/50000, loss = 3.486598 (0.072 sec/batch), lr: 0.003000
398
+ 2025-12-03 15:54:19 INFO: Finished STEP 1640/50000, loss = 3.426835 (0.076 sec/batch), lr: 0.003000
399
+ 2025-12-03 15:54:20 INFO: Finished STEP 1660/50000, loss = 3.082500 (0.077 sec/batch), lr: 0.003000
400
+ 2025-12-03 15:54:22 INFO: Finished STEP 1680/50000, loss = 2.587858 (0.080 sec/batch), lr: 0.003000
401
+ 2025-12-03 15:54:23 INFO: Finished STEP 1700/50000, loss = 4.049259 (0.081 sec/batch), lr: 0.003000
402
+ 2025-12-03 15:54:23 INFO: Evaluating on dev set...
403
+ 2025-12-03 15:54:24 INFO: LAS MLAS BLEX
404
+ 2025-12-03 15:54:24 INFO: 59.41 50.63 54.43
405
+ 2025-12-03 15:54:24 INFO: step 1700: train_loss = 3.731144, dev_score = 0.5941
406
+ 2025-12-03 15:54:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
407
+ 2025-12-03 15:54:24 INFO: new model checkpoint saved.
408
+ 2025-12-03 15:54:26 INFO: Finished STEP 1720/50000, loss = 2.259671 (0.076 sec/batch), lr: 0.003000
409
+ 2025-12-03 15:54:28 INFO: Finished STEP 1740/50000, loss = 5.103413 (0.079 sec/batch), lr: 0.003000
410
+ 2025-12-03 15:54:29 INFO: Finished STEP 1760/50000, loss = 3.651329 (0.072 sec/batch), lr: 0.003000
411
+ 2025-12-03 15:54:31 INFO: Finished STEP 1780/50000, loss = 3.879523 (0.078 sec/batch), lr: 0.003000
412
+ 2025-12-03 15:54:32 INFO: Finished STEP 1800/50000, loss = 3.246203 (0.078 sec/batch), lr: 0.003000
413
+ 2025-12-03 15:54:32 INFO: Evaluating on dev set...
414
+ 2025-12-03 15:54:33 INFO: LAS MLAS BLEX
415
+ 2025-12-03 15:54:33 INFO: 60.89 53.39 55.93
416
+ 2025-12-03 15:54:33 INFO: step 1800: train_loss = 3.826308, dev_score = 0.6089
417
+ 2025-12-03 15:54:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
418
+ 2025-12-03 15:54:33 INFO: new best model saved.
419
+ 2025-12-03 15:54:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
420
+ 2025-12-03 15:54:34 INFO: new model checkpoint saved.
421
+ 2025-12-03 15:54:35 INFO: Finished STEP 1820/50000, loss = 4.498244 (0.072 sec/batch), lr: 0.003000
422
+ 2025-12-03 15:54:37 INFO: Finished STEP 1840/50000, loss = 3.589321 (0.078 sec/batch), lr: 0.003000
423
+ 2025-12-03 15:54:39 INFO: Finished STEP 1860/50000, loss = 2.884829 (0.077 sec/batch), lr: 0.003000
424
+ 2025-12-03 15:54:40 INFO: Finished STEP 1880/50000, loss = 3.823880 (0.082 sec/batch), lr: 0.003000
425
+ 2025-12-03 15:54:42 INFO: Finished STEP 1900/50000, loss = 2.696285 (0.082 sec/batch), lr: 0.003000
426
+ 2025-12-03 15:54:42 INFO: Evaluating on dev set...
427
+ 2025-12-03 15:54:42 INFO: LAS MLAS BLEX
428
+ 2025-12-03 15:54:42 INFO: 60.64 51.16 55.39
429
+ 2025-12-03 15:54:42 INFO: step 1900: train_loss = 4.029873, dev_score = 0.6064
430
+ 2025-12-03 15:54:43 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
431
+ 2025-12-03 15:54:43 INFO: new model checkpoint saved.
432
+ 2025-12-03 15:54:44 INFO: Finished STEP 1920/50000, loss = 5.128090 (0.068 sec/batch), lr: 0.003000
433
+ 2025-12-03 15:54:46 INFO: Finished STEP 1940/50000, loss = 3.145783 (0.078 sec/batch), lr: 0.003000
434
+ 2025-12-03 15:54:47 INFO: Finished STEP 1960/50000, loss = 3.195472 (0.075 sec/batch), lr: 0.003000
435
+ 2025-12-03 15:54:49 INFO: Finished STEP 1980/50000, loss = 3.154953 (0.072 sec/batch), lr: 0.003000
436
+ 2025-12-03 15:54:51 INFO: Finished STEP 2000/50000, loss = 3.763106 (0.080 sec/batch), lr: 0.003000
437
+ 2025-12-03 15:54:51 INFO: Evaluating on dev set...
438
+ 2025-12-03 15:54:51 INFO: LAS MLAS BLEX
439
+ 2025-12-03 15:54:51 INFO: 61.39 50.85 54.66
440
+ 2025-12-03 15:54:51 INFO: step 2000: train_loss = 3.764299, dev_score = 0.6139
441
+ 2025-12-03 15:54:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
442
+ 2025-12-03 15:54:51 INFO: new best model saved.
443
+ 2025-12-03 15:54:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
444
+ 2025-12-03 15:54:52 INFO: new model checkpoint saved.
445
+ 2025-12-03 15:54:53 INFO: Finished STEP 2020/50000, loss = 4.129125 (0.077 sec/batch), lr: 0.003000
446
+ 2025-12-03 15:54:55 INFO: Finished STEP 2040/50000, loss = 4.000008 (0.076 sec/batch), lr: 0.003000
447
+ 2025-12-03 15:54:56 INFO: Finished STEP 2060/50000, loss = 3.594474 (0.076 sec/batch), lr: 0.003000
448
+ 2025-12-03 15:54:58 INFO: Finished STEP 2080/50000, loss = 5.156559 (0.070 sec/batch), lr: 0.003000
449
+ 2025-12-03 15:54:59 INFO: Finished STEP 2100/50000, loss = 2.780583 (0.076 sec/batch), lr: 0.003000
450
+ 2025-12-03 15:54:59 INFO: Evaluating on dev set...
451
+ 2025-12-03 15:55:00 INFO: LAS MLAS BLEX
452
+ 2025-12-03 15:55:00 INFO: 58.91 50.32 53.28
453
+ 2025-12-03 15:55:00 INFO: step 2100: train_loss = 3.970306, dev_score = 0.5891
454
+ 2025-12-03 15:55:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
455
+ 2025-12-03 15:55:01 INFO: new model checkpoint saved.
456
+ 2025-12-03 15:55:02 INFO: Finished STEP 2120/50000, loss = 3.461082 (0.078 sec/batch), lr: 0.003000
457
+ 2025-12-03 15:55:04 INFO: Finished STEP 2140/50000, loss = 4.371090 (0.068 sec/batch), lr: 0.003000
458
+ 2025-12-03 15:55:05 INFO: Finished STEP 2160/50000, loss = 3.734883 (0.074 sec/batch), lr: 0.003000
459
+ 2025-12-03 15:55:07 INFO: Finished STEP 2180/50000, loss = 2.186298 (0.077 sec/batch), lr: 0.003000
460
+ 2025-12-03 15:55:08 INFO: Finished STEP 2200/50000, loss = 3.871332 (0.078 sec/batch), lr: 0.003000
461
+ 2025-12-03 15:55:08 INFO: Evaluating on dev set...
462
+ 2025-12-03 15:55:09 INFO: LAS MLAS BLEX
463
+ 2025-12-03 15:55:09 INFO: 58.66 48.73 52.97
464
+ 2025-12-03 15:55:09 INFO: step 2200: train_loss = 3.926589, dev_score = 0.5866
465
+ 2025-12-03 15:55:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
466
+ 2025-12-03 15:55:09 INFO: new model checkpoint saved.
467
+ 2025-12-03 15:55:11 INFO: Finished STEP 2220/50000, loss = 2.865967 (0.073 sec/batch), lr: 0.003000
468
+ 2025-12-03 15:55:12 INFO: Finished STEP 2240/50000, loss = 2.578629 (0.069 sec/batch), lr: 0.003000
469
+ 2025-12-03 15:55:14 INFO: Finished STEP 2260/50000, loss = 3.295215 (0.068 sec/batch), lr: 0.003000
470
+ 2025-12-03 15:55:15 INFO: Finished STEP 2280/50000, loss = 3.903890 (0.074 sec/batch), lr: 0.003000
471
+ 2025-12-03 15:55:17 INFO: Finished STEP 2300/50000, loss = 3.614969 (0.071 sec/batch), lr: 0.003000
472
+ 2025-12-03 15:55:17 INFO: Evaluating on dev set...
473
+ 2025-12-03 15:55:17 INFO: LAS MLAS BLEX
474
+ 2025-12-03 15:55:17 INFO: 60.40 52.43 56.66
475
+ 2025-12-03 15:55:17 INFO: step 2300: train_loss = 3.953848, dev_score = 0.6040
476
+ 2025-12-03 15:55:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
477
+ 2025-12-03 15:55:18 INFO: new model checkpoint saved.
478
+ 2025-12-03 15:55:19 INFO: Finished STEP 2320/50000, loss = 3.836884 (0.076 sec/batch), lr: 0.003000
479
+ 2025-12-03 15:55:21 INFO: Finished STEP 2340/50000, loss = 2.887274 (0.079 sec/batch), lr: 0.003000
480
+ 2025-12-03 15:55:22 INFO: Finished STEP 2360/50000, loss = 4.795152 (0.079 sec/batch), lr: 0.003000
481
+ 2025-12-03 15:55:24 INFO: Finished STEP 2380/50000, loss = 4.483939 (0.074 sec/batch), lr: 0.003000
482
+ 2025-12-03 15:55:25 INFO: Finished STEP 2400/50000, loss = 4.851534 (0.075 sec/batch), lr: 0.003000
483
+ 2025-12-03 15:55:25 INFO: Evaluating on dev set...
484
+ 2025-12-03 15:55:26 INFO: LAS MLAS BLEX
485
+ 2025-12-03 15:55:26 INFO: 58.91 50.21 54.85
486
+ 2025-12-03 15:55:26 INFO: step 2400: train_loss = 3.932371, dev_score = 0.5891
487
+ 2025-12-03 15:55:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
488
+ 2025-12-03 15:55:27 INFO: new model checkpoint saved.
489
+ 2025-12-03 15:55:28 INFO: Finished STEP 2420/50000, loss = 4.692697 (0.067 sec/batch), lr: 0.003000
490
+ 2025-12-03 15:55:30 INFO: Finished STEP 2440/50000, loss = 2.805047 (0.068 sec/batch), lr: 0.003000
491
+ 2025-12-03 15:55:31 INFO: Finished STEP 2460/50000, loss = 3.715670 (0.077 sec/batch), lr: 0.003000
492
+ 2025-12-03 15:55:32 INFO: Finished STEP 2480/50000, loss = 5.003852 (0.072 sec/batch), lr: 0.003000
493
+ 2025-12-03 15:55:34 INFO: Finished STEP 2500/50000, loss = 5.362862 (0.071 sec/batch), lr: 0.003000
494
+ 2025-12-03 15:55:34 INFO: Evaluating on dev set...
495
+ 2025-12-03 15:55:34 INFO: LAS MLAS BLEX
496
+ 2025-12-03 15:55:34 INFO: 62.87 54.12 57.51
497
+ 2025-12-03 15:55:34 INFO: step 2500: train_loss = 3.716296, dev_score = 0.6287
498
+ 2025-12-03 15:55:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
499
+ 2025-12-03 15:55:35 INFO: new best model saved.
500
+ 2025-12-03 15:55:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
501
+ 2025-12-03 15:55:35 INFO: new model checkpoint saved.
502
+ 2025-12-03 15:55:37 INFO: Finished STEP 2520/50000, loss = 4.285573 (0.073 sec/batch), lr: 0.003000
503
+ 2025-12-03 15:55:38 INFO: Finished STEP 2540/50000, loss = 3.220155 (0.071 sec/batch), lr: 0.003000
504
+ 2025-12-03 15:55:40 INFO: Finished STEP 2560/50000, loss = 4.155645 (0.077 sec/batch), lr: 0.003000
505
+ 2025-12-03 15:55:41 INFO: Finished STEP 2580/50000, loss = 3.434030 (0.073 sec/batch), lr: 0.003000
506
+ 2025-12-03 15:55:43 INFO: Finished STEP 2600/50000, loss = 2.680579 (0.076 sec/batch), lr: 0.003000
507
+ 2025-12-03 15:55:43 INFO: Evaluating on dev set...
508
+ 2025-12-03 15:55:43 INFO: LAS MLAS BLEX
509
+ 2025-12-03 15:55:43 INFO: 62.87 54.66 58.05
510
+ 2025-12-03 15:55:43 INFO: step 2600: train_loss = 4.018973, dev_score = 0.6287
511
+ 2025-12-03 15:55:43 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
512
+ 2025-12-03 15:55:43 INFO: new best model saved.
513
+ 2025-12-03 15:55:44 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
514
+ 2025-12-03 15:55:44 INFO: new model checkpoint saved.
515
+ 2025-12-03 15:55:45 INFO: Finished STEP 2620/50000, loss = 2.524696 (0.072 sec/batch), lr: 0.003000
516
+ 2025-12-03 15:55:47 INFO: Finished STEP 2640/50000, loss = 3.829545 (0.065 sec/batch), lr: 0.003000
517
+ 2025-12-03 15:55:48 INFO: Finished STEP 2660/50000, loss = 3.246467 (0.069 sec/batch), lr: 0.003000
518
+ 2025-12-03 15:55:50 INFO: Finished STEP 2680/50000, loss = 2.396587 (0.068 sec/batch), lr: 0.003000
519
+ 2025-12-03 15:55:51 INFO: Finished STEP 2700/50000, loss = 3.570580 (0.071 sec/batch), lr: 0.003000
520
+ 2025-12-03 15:55:51 INFO: Evaluating on dev set...
521
+ 2025-12-03 15:55:52 INFO: LAS MLAS BLEX
522
+ 2025-12-03 15:55:52 INFO: 58.91 49.58 54.24
523
+ 2025-12-03 15:55:52 INFO: step 2700: train_loss = 3.918322, dev_score = 0.5891
524
+ 2025-12-03 15:55:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
525
+ 2025-12-03 15:55:52 INFO: new model checkpoint saved.
526
+ 2025-12-03 15:55:54 INFO: Finished STEP 2720/50000, loss = 2.712030 (0.072 sec/batch), lr: 0.003000
527
+ 2025-12-03 15:55:55 INFO: Finished STEP 2740/50000, loss = 5.811014 (0.077 sec/batch), lr: 0.003000
528
+ 2025-12-03 15:55:57 INFO: Finished STEP 2760/50000, loss = 3.773255 (0.076 sec/batch), lr: 0.003000
529
+ 2025-12-03 15:55:58 INFO: Finished STEP 2780/50000, loss = 2.087427 (0.075 sec/batch), lr: 0.003000
530
+ 2025-12-03 15:56:00 INFO: Finished STEP 2800/50000, loss = 3.811720 (0.063 sec/batch), lr: 0.003000
531
+ 2025-12-03 15:56:00 INFO: Evaluating on dev set...
532
+ 2025-12-03 15:56:00 INFO: LAS MLAS BLEX
533
+ 2025-12-03 15:56:00 INFO: 61.63 51.59 55.39
534
+ 2025-12-03 15:56:00 INFO: step 2800: train_loss = 3.902705, dev_score = 0.6163
535
+ 2025-12-03 15:56:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
536
+ 2025-12-03 15:56:01 INFO: new model checkpoint saved.
537
+ 2025-12-03 15:56:02 INFO: Finished STEP 2820/50000, loss = 3.443505 (0.070 sec/batch), lr: 0.003000
538
+ 2025-12-03 15:56:04 INFO: Finished STEP 2840/50000, loss = 3.684120 (0.064 sec/batch), lr: 0.003000
539
+ 2025-12-03 15:56:05 INFO: Finished STEP 2860/50000, loss = 5.325919 (0.071 sec/batch), lr: 0.003000
540
+ 2025-12-03 15:56:06 INFO: Finished STEP 2880/50000, loss = 3.516696 (0.068 sec/batch), lr: 0.003000
541
+ 2025-12-03 15:56:08 INFO: Finished STEP 2900/50000, loss = 3.573216 (0.072 sec/batch), lr: 0.003000
542
+ 2025-12-03 15:56:08 INFO: Evaluating on dev set...
543
+ 2025-12-03 15:56:09 INFO: LAS MLAS BLEX
544
+ 2025-12-03 15:56:09 INFO: 58.17 48.20 53.28
545
+ 2025-12-03 15:56:09 INFO: step 2900: train_loss = 4.154457, dev_score = 0.5817
546
+ 2025-12-03 15:56:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
547
+ 2025-12-03 15:56:09 INFO: new model checkpoint saved.
548
+ 2025-12-03 15:56:11 INFO: Finished STEP 2920/50000, loss = 2.522206 (0.077 sec/batch), lr: 0.003000
549
+ 2025-12-03 15:56:12 INFO: Finished STEP 2940/50000, loss = 2.687997 (0.082 sec/batch), lr: 0.003000
550
+ 2025-12-03 15:56:13 INFO: Finished STEP 2960/50000, loss = 3.291323 (0.078 sec/batch), lr: 0.003000
551
+ 2025-12-03 15:56:15 INFO: Finished STEP 2980/50000, loss = 4.553081 (0.069 sec/batch), lr: 0.003000
552
+ 2025-12-03 15:56:16 INFO: Finished STEP 3000/50000, loss = 3.136598 (0.071 sec/batch), lr: 0.003000
553
+ 2025-12-03 15:56:16 INFO: Evaluating on dev set...
554
+ 2025-12-03 15:56:17 INFO: LAS MLAS BLEX
555
+ 2025-12-03 15:56:17 INFO: 61.14 52.43 55.39
556
+ 2025-12-03 15:56:17 INFO: step 3000: train_loss = 3.949049, dev_score = 0.6114
557
+ 2025-12-03 15:56:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
558
+ 2025-12-03 15:56:17 INFO: new model checkpoint saved.
559
+ 2025-12-03 15:56:19 INFO: Finished STEP 3020/50000, loss = 3.020242 (0.077 sec/batch), lr: 0.003000
560
+ 2025-12-03 15:56:20 INFO: Finished STEP 3040/50000, loss = 4.876935 (0.069 sec/batch), lr: 0.003000
561
+ 2025-12-03 15:56:22 INFO: Finished STEP 3060/50000, loss = 3.937078 (0.080 sec/batch), lr: 0.003000
562
+ 2025-12-03 15:56:23 INFO: Finished STEP 3080/50000, loss = 4.811732 (0.064 sec/batch), lr: 0.003000
563
+ 2025-12-03 15:56:25 INFO: Finished STEP 3100/50000, loss = 3.313049 (0.073 sec/batch), lr: 0.003000
564
+ 2025-12-03 15:56:25 INFO: Evaluating on dev set...
565
+ 2025-12-03 15:56:25 INFO: LAS MLAS BLEX
566
+ 2025-12-03 15:56:25 INFO: 62.13 54.12 58.35
567
+ 2025-12-03 15:56:25 INFO: step 3100: train_loss = 4.106587, dev_score = 0.6213
568
+ 2025-12-03 15:56:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
569
+ 2025-12-03 15:56:26 INFO: new model checkpoint saved.
570
+ 2025-12-03 15:56:27 INFO: Finished STEP 3120/50000, loss = 3.817453 (0.067 sec/batch), lr: 0.003000
571
+ 2025-12-03 15:56:29 INFO: Finished STEP 3140/50000, loss = 5.620703 (0.070 sec/batch), lr: 0.003000
572
+ 2025-12-03 15:56:30 INFO: Finished STEP 3160/50000, loss = 4.408039 (0.078 sec/batch), lr: 0.003000
573
+ 2025-12-03 15:56:32 INFO: Finished STEP 3180/50000, loss = 2.225734 (0.074 sec/batch), lr: 0.003000
574
+ 2025-12-03 15:56:33 INFO: Finished STEP 3200/50000, loss = 6.085883 (0.082 sec/batch), lr: 0.003000
575
+ 2025-12-03 15:56:33 INFO: Evaluating on dev set...
576
+ 2025-12-03 15:56:34 INFO: LAS MLAS BLEX
577
+ 2025-12-03 15:56:34 INFO: 58.91 51.59 53.70
578
+ 2025-12-03 15:56:34 INFO: step 3200: train_loss = 4.045106, dev_score = 0.5891
579
+ 2025-12-03 15:56:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
580
+ 2025-12-03 15:56:34 INFO: new model checkpoint saved.
581
+ 2025-12-03 15:56:36 INFO: Finished STEP 3220/50000, loss = 3.643058 (0.064 sec/batch), lr: 0.003000
582
+ 2025-12-03 15:56:37 INFO: Finished STEP 3240/50000, loss = 3.506227 (0.068 sec/batch), lr: 0.003000
583
+ 2025-12-03 15:56:38 INFO: Finished STEP 3260/50000, loss = 2.930928 (0.072 sec/batch), lr: 0.003000
584
+ 2025-12-03 15:56:40 INFO: Finished STEP 3280/50000, loss = 4.896526 (0.071 sec/batch), lr: 0.003000
585
+ 2025-12-03 15:56:41 INFO: Finished STEP 3300/50000, loss = 4.648397 (0.077 sec/batch), lr: 0.003000
586
+ 2025-12-03 15:56:41 INFO: Evaluating on dev set...
587
+ 2025-12-03 15:56:42 INFO: LAS MLAS BLEX
588
+ 2025-12-03 15:56:42 INFO: 59.16 52.01 52.43
589
+ 2025-12-03 15:56:42 INFO: step 3300: train_loss = 3.880126, dev_score = 0.5916
590
+ 2025-12-03 15:56:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
591
+ 2025-12-03 15:56:42 INFO: new model checkpoint saved.
592
+ 2025-12-03 15:56:44 INFO: Finished STEP 3320/50000, loss = 4.656130 (0.065 sec/batch), lr: 0.003000
593
+ 2025-12-03 15:56:45 INFO: Finished STEP 3340/50000, loss = 4.644033 (0.070 sec/batch), lr: 0.003000
594
+ 2025-12-03 15:56:47 INFO: Finished STEP 3360/50000, loss = 4.342103 (0.070 sec/batch), lr: 0.003000
595
+ 2025-12-03 15:56:48 INFO: Finished STEP 3380/50000, loss = 1.757741 (0.069 sec/batch), lr: 0.003000
596
+ 2025-12-03 15:56:49 INFO: Finished STEP 3400/50000, loss = 3.296278 (0.077 sec/batch), lr: 0.003000
597
+ 2025-12-03 15:56:49 INFO: Evaluating on dev set...
598
+ 2025-12-03 15:56:50 INFO: LAS MLAS BLEX
599
+ 2025-12-03 15:56:50 INFO: 57.92 48.63 52.43
600
+ 2025-12-03 15:56:50 INFO: step 3400: train_loss = 3.761222, dev_score = 0.5792
601
+ 2025-12-03 15:56:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
602
+ 2025-12-03 15:56:51 INFO: new model checkpoint saved.
603
+ 2025-12-03 15:56:52 INFO: Finished STEP 3420/50000, loss = 4.175681 (0.063 sec/batch), lr: 0.003000
604
+ 2025-12-03 15:56:53 INFO: Finished STEP 3440/50000, loss = 3.109319 (0.072 sec/batch), lr: 0.003000
605
+ 2025-12-03 15:56:55 INFO: Finished STEP 3460/50000, loss = 5.698625 (0.070 sec/batch), lr: 0.003000
606
+ 2025-12-03 15:56:56 INFO: Finished STEP 3480/50000, loss = 4.906111 (0.069 sec/batch), lr: 0.003000
607
+ 2025-12-03 15:56:58 INFO: Finished STEP 3500/50000, loss = 2.886077 (0.068 sec/batch), lr: 0.003000
608
+ 2025-12-03 15:56:58 INFO: Evaluating on dev set...
609
+ 2025-12-03 15:56:58 INFO: LAS MLAS BLEX
610
+ 2025-12-03 15:56:58 INFO: 60.64 51.48 55.70
611
+ 2025-12-03 15:56:58 INFO: step 3500: train_loss = 3.804499, dev_score = 0.6064
612
+ 2025-12-03 15:56:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
613
+ 2025-12-03 15:56:59 INFO: new model checkpoint saved.
614
+ 2025-12-03 15:57:00 INFO: Finished STEP 3520/50000, loss = 2.301594 (0.065 sec/batch), lr: 0.003000
615
+ 2025-12-03 15:57:01 INFO: Finished STEP 3540/50000, loss = 4.423633 (0.075 sec/batch), lr: 0.003000
616
+ 2025-12-03 15:57:03 INFO: Finished STEP 3560/50000, loss = 3.613044 (0.063 sec/batch), lr: 0.003000
617
+ 2025-12-03 15:57:04 INFO: Finished STEP 3580/50000, loss = 3.496284 (0.067 sec/batch), lr: 0.003000
618
+ 2025-12-03 15:57:06 INFO: Finished STEP 3600/50000, loss = 6.527389 (0.066 sec/batch), lr: 0.003000
619
+ 2025-12-03 15:57:06 INFO: Evaluating on dev set...
620
+ 2025-12-03 15:57:06 INFO: LAS MLAS BLEX
621
+ 2025-12-03 15:57:06 INFO: 60.64 51.27 55.08
622
+ 2025-12-03 15:57:06 INFO: step 3600: train_loss = 4.020496, dev_score = 0.6064
623
+ 2025-12-03 15:57:06 INFO: Training ended with 3600 steps.
624
+ 2025-12-03 15:57:06 INFO: Best dev F1 = 62.87, at iteration = 2500
625
+ 2025-12-03 15:57:07 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
626
+ 2025-12-03 15:57:07 INFO: Running parser in predict mode
627
+ 2025-12-03 15:57:07 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
628
+ 2025-12-03 15:57:09 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
629
+ 2025-12-03 15:57:09 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
630
+ 2025-12-03 15:57:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
631
+ 2025-12-03 15:57:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
632
+ 2025-12-03 15:57:10 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
633
+ 2025-12-03 15:57:10 INFO: Loading data with batch size 32...
634
+ 2025-12-03 15:57:10 DEBUG: 9 batches created.
635
+ 2025-12-03 15:57:10 INFO: F1 scores for each dependency:
636
+ Note that unlabeled attachment errors hurt the labeled attachment scores
637
+ acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
638
+ acl:relcl: p 0.2222 r 0.2857 f1 0.2500 (7 actual)
639
+ advcl: p 0.1250 r 0.2000 f1 0.1538 (5 actual)
640
+ advmod: p 0.5357 r 0.6000 f1 0.5660 (25 actual)
641
+ amod: p 0.9231 r 0.7742 f1 0.8421 (31 actual)
642
+ appos: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
643
+ aux: p 0.8182 r 0.8182 f1 0.8182 (11 actual)
644
+ case: p 0.9623 r 0.9107 f1 0.9358 (56 actual)
645
+ cc: p 0.6429 r 0.6923 f1 0.6667 (13 actual)
646
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
647
+ conj: p 0.2667 r 0.3333 f1 0.2963 (12 actual)
648
+ cop: p 0.5000 r 0.3333 f1 0.4000 (3 actual)
649
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
650
+ det: p 0.8696 r 0.9091 f1 0.8889 (22 actual)
651
+ expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
652
+ iobj: p 0.5000 r 0.5000 f1 0.5000 (2 actual)
653
+ mark: p 0.4167 r 0.4167 f1 0.4167 (12 actual)
654
+ nmod: p 0.4000 r 0.5333 f1 0.4571 (15 actual)
655
+ nmod:poss: p 1.0000 r 0.8947 f1 0.9444 (19 actual)
656
+ nsubj: p 0.5417 r 0.7647 f1 0.6341 (17 actual)
657
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
658
+ obj: p 0.6071 r 0.7727 f1 0.6800 (22 actual)
659
+ obl: p 0.6154 r 0.5854 f1 0.6000 (41 actual)
660
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
661
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
662
+ parataxis: p 0.1111 r 0.3333 f1 0.1667 (3 actual)
663
+ punct: p 0.4231 r 0.4231 f1 0.4231 (52 actual)
664
+ root: p 0.4444 r 0.4444 f1 0.4444 (9 actual)
665
+ xcomp: p 0.5000 r 0.1250 f1 0.2000 (8 actual)
666
+ 2025-12-03 15:57:10 INFO: LAS MLAS BLEX
667
+ 2025-12-03 15:57:10 INFO: 62.87 54.66 58.05
668
+ 2025-12-03 15:57:10 INFO: Parser score:
669
+ 2025-12-03 15:57:10 INFO: sv_diachronic 62.87
670
+ 2025-12-03 15:57:10 INFO: Finished running dev set on
671
+ UD_Swedish-diachronic
672
+ UAS LAS CLAS MLAS BLEX
673
+ 68.56 62.87 58.05 54.66 58.05
674
+ 2025-12-03 15:57:10 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
675
+ 2025-12-03 15:57:10 INFO: Running parser in predict mode
676
+ 2025-12-03 15:57:10 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
677
+ 2025-12-03 15:57:12 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
678
+ 2025-12-03 15:57:12 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
679
+ 2025-12-03 15:57:12 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
680
+ 2025-12-03 15:57:12 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
681
+ 2025-12-03 15:57:12 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
682
+ 2025-12-03 15:57:12 INFO: Loading data with batch size 32...
683
+ 2025-12-03 15:57:12 DEBUG: 93 batches created.
684
+ 2025-12-03 15:57:17 INFO: F1 scores for each dependency:
685
+ Note that unlabeled attachment errors hurt the labeled attachment scores
686
+ acl: p 0.2778 r 0.1562 f1 0.2000 (32 actual)
687
+ acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
688
+ acl:relcl: p 0.2075 r 0.1467 f1 0.1719 (75 actual)
689
+ advcl: p 0.0842 r 0.1333 f1 0.1032 (60 actual)
690
+ advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
691
+ advmod: p 0.5520 r 0.5746 f1 0.5631 (268 actual)
692
+ amod: p 0.8596 r 0.8522 f1 0.8559 (230 actual)
693
+ appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
694
+ aux: p 0.8734 r 0.8214 f1 0.8466 (84 actual)
695
+ aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
696
+ case: p 0.8481 r 0.8231 f1 0.8354 (373 actual)
697
+ cc: p 0.5833 r 0.5871 f1 0.5852 (155 actual)
698
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
699
+ compound:prt: p 0.7727 r 0.8095 f1 0.7907 (21 actual)
700
+ conj: p 0.2397 r 0.2215 f1 0.2303 (158 actual)
701
+ cop: p 0.7714 r 0.5870 f1 0.6667 (46 actual)
702
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
703
+ dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
704
+ det: p 0.8960 r 0.8702 f1 0.8829 (208 actual)
705
+ discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
706
+ dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
707
+ expl: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
708
+ expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
709
+ fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
710
+ flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
711
+ flat:name: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
712
+ goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
713
+ iobj: p 0.3333 r 0.0714 f1 0.1176 (14 actual)
714
+ mark: p 0.6757 r 0.6536 f1 0.6645 (153 actual)
715
+ nmod: p 0.3798 r 0.4804 f1 0.4242 (102 actual)
716
+ nmod:poss: p 0.8707 r 0.9014 f1 0.8858 (142 actual)
717
+ nsubj: p 0.5673 r 0.6321 f1 0.5980 (280 actual)
718
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
719
+ nummod: p 0.7500 r 0.9000 f1 0.8182 (10 actual)
720
+ obj: p 0.5628 r 0.7104 f1 0.6280 (183 actual)
721
+ obl: p 0.4857 r 0.5504 f1 0.5160 (278 actual)
722
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
723
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
724
+ parataxis: p 0.0333 r 0.1667 f1 0.0556 (18 actual)
725
+ punct: p 0.4262 r 0.4282 f1 0.4272 (425 actual)
726
+ reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
727
+ root: p 0.5354 r 0.5354 f1 0.5354 (99 actual)
728
+ vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
729
+ xcomp: p 0.5750 r 0.3067 f1 0.4000 (75 actual)
730
+ 2025-12-03 15:57:17 INFO: LAS MLAS BLEX
731
+ 2025-12-03 15:57:17 INFO: 58.54 50.12 53.03
732
+ 2025-12-03 15:57:17 INFO: Parser score:
733
+ 2025-12-03 15:57:17 INFO: sv_diachronic 58.54
734
+ 2025-12-03 15:57:17 INFO: Finished running test set on
735
+ UD_Swedish-diachronic
736
+ UAS LAS CLAS MLAS BLEX
737
+ 67.16 58.54 53.03 50.12 53.03
738
+ DONE.
739
+ Full log saved to: logs/log_conll17.pt_sv_diachron_de_lit_20251203_154509.txt
740
+ Symlink updated: logs/latest.txt → log_conll17.pt_sv_diachron_de_lit_20251203_154509.txt
logs/log_conll17.pt_sv_diachron_is_20251203_164326.txt ADDED
@@ -0,0 +1,911 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ === LOGFILE: logs/log_conll17.pt_sv_diachron_is_20251203_164326.txt ===
2
+ Language codes: sv diachron is
3
+ Using pretrained model: conll17.pt
4
+
5
+ Running: python prepare-train-val-test.py sv diachron is
6
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
7
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
8
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
9
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
10
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
11
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
12
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
13
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
14
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
15
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
16
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
17
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_pud-ud-test.conllu
18
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_icepahc-ud-train.conllu
19
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_icepahc-ud-test.conllu
20
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_icepahc-ud-dev.conllu
21
+ Including DigPhil MACHINE in TRAIN (minus gold)…
22
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
23
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
24
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
25
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
26
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
27
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
28
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
29
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
30
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
31
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
32
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
33
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
34
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
35
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
36
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
37
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
38
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
39
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
40
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
41
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
42
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
43
+ Cleaning TRAIN...
44
+ [REMOVED] sent_id=6 ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
45
+ [REMOVED] sent_id=7_8 ERRORS=['Multiple roots found: [5, 10]']
46
+ [REMOVED] sent_id=30_31 ERRORS=['Multiple roots found: [3, 18]']
47
+ [REMOVED] sent_id=35 ERRORS=['Line 36: Invalid token ID or head']
48
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [1, 5]']
49
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [7, 20]']
50
+ [REMOVED] sent_id=8_9 ERRORS=['Multiple roots found: [24, 57]']
51
+ [REMOVED] sent_id=12_13 ERRORS=['Multiple roots found: [11, 16]']
52
+ [REMOVED] sent_id=124_split2 ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
53
+ [REMOVED] sent_id=396 ERRORS=['Token 2: Missing form']
54
+ [REMOVED] sent_id=416 ERRORS=['Token 2: Missing form']
55
+ [REMOVED] sent_id=589 ERRORS=['Token 2: Missing form']
56
+ [REMOVED] sent_id=909 ERRORS=['Token 2: Missing form']
57
+ [REMOVED] sent_id=912 ERRORS=['Token 2: Missing form']
58
+ [REMOVED] sent_id=3_split1 ERRORS=['Multiple roots found: [4, 15, 17]']
59
+ [REMOVED] sent_id=3_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
60
+ [REMOVED] sent_id=3_4 ERRORS=['Multiple roots found: [1, 5]']
61
+ [REMOVED] sent_id=5_6 ERRORS=['Multiple roots found: [3, 24]']
62
+ [REMOVED] sent_id=11_12_13 ERRORS=['Multiple roots found: [5, 17, 25]']
63
+ [REMOVED] sent_id=119 ERRORS=['Token 2: Missing form']
64
+ [REMOVED] sent_id=179 ERRORS=['Token 2: Missing form']
65
+ [REMOVED] sent_id=188 ERRORS=['Token 2: Missing form']
66
+ [REMOVED] sent_id=223 ERRORS=['Token 2: Missing form']
67
+ [REMOVED] sent_id=268 ERRORS=['Token 2: Missing form']
68
+ [REMOVED] sent_id=325 ERRORS=['Token 2: Missing form']
69
+ [REMOVED] sent_id=388 ERRORS=['Token 2: Missing form']
70
+ [REMOVED] sent_id=399 ERRORS=['Token 2: Missing form']
71
+ [REMOVED] sent_id=475 ERRORS=['Token 2: Missing form']
72
+ [REMOVED] sent_id=505 ERRORS=['Token 2: Missing form']
73
+ [REMOVED] sent_id=520 ERRORS=['Token 2: Missing form']
74
+ [REMOVED] sent_id=562 ERRORS=['Token 2: Missing form']
75
+ [REMOVED] sent_id=669 ERRORS=['Token 2: Missing form']
76
+ [REMOVED] sent_id=711 ERRORS=['Token 2: Missing form']
77
+ [REMOVED] sent_id=731 ERRORS=['Token 2: Missing form']
78
+ [REMOVED] sent_id=867 ERRORS=['Token 2: Missing form']
79
+ [REMOVED] sent_id=884 ERRORS=['Token 2: Missing form']
80
+ [REMOVED] sent_id=923 ERRORS=['Token 2: Missing form']
81
+ [REMOVED] sent_id=939 ERRORS=['Token 2: Missing form']
82
+ [REMOVED] sent_id=1086 ERRORS=['Token 2: Missing form']
83
+ [REMOVED] sent_id=1179 ERRORS=['Token 2: Missing form']
84
+ [REMOVED] sent_id=1251 ERRORS=['Token 2: Missing form']
85
+ [REMOVED] sent_id=1345 ERRORS=['Token 2: Missing form']
86
+ [REMOVED] sent_id=1459 ERRORS=['Token 2: Missing form']
87
+ [REMOVED] sent_id=1656 ERRORS=['Token 2: Missing form']
88
+ [REMOVED] sent_id=1669 ERRORS=['Token 2: Missing form']
89
+ [REMOVED] sent_id=87_88 ERRORS=['Multiple roots found: [3, 6]']
90
+ [REMOVED] sent_id=65_split2_66_split2 ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
91
+ [REMOVED] sent_id=25 ERRORS=['Token 2: Missing form']
92
+ [REMOVED] sent_id=136 ERRORS=['Token 2: Missing form']
93
+ [REMOVED] sent_id=208 ERRORS=['Token 2: Missing form']
94
+ [REMOVED] sent_id=230 ERRORS=['Token 2: Missing form']
95
+ [REMOVED] sent_id=245 ERRORS=['Token 2: Missing form']
96
+ [REMOVED] sent_id=276 ERRORS=['Token 2: Missing form']
97
+ [REMOVED] sent_id=320 ERRORS=['Token 2: Missing form']
98
+ [REMOVED] sent_id=366 ERRORS=['Token 2: Missing form']
99
+ [REMOVED] sent_id=519 ERRORS=['Token 2: Missing form']
100
+ [REMOVED] sent_id=569 ERRORS=['Token 2: Missing form']
101
+ [REMOVED] sent_id=50_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
102
+ [REMOVED] sent_id=53_54 ERRORS=['Multiple roots found: [27, 91]']
103
+ [REMOVED] sent_id=55_56_57 ERRORS=['Multiple roots found: [2, 4, 13]']
104
+ [REMOVED] sent_id=17_split1 ERRORS=['Multiple roots found: [2, 14, 17]']
105
+ [REMOVED] sent_id=17_split2 ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
106
+ [REMOVED] sent_id=19_split1 ERRORS=['Multiple roots found: [3, 31]']
107
+ Cleaning DEV...
108
+ [REMOVED] sent_id=33 ERRORS=['Token 15: Missing deprel']
109
+ Cleaning TEST...
110
+ Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (111281 valid sentences)
111
+ Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
112
+ Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
113
+ Done.
114
+ Sourcing scripts/config_alvis.sh
115
+ Running stanza dataset preparation…
116
+ 2025-12-03 16:43:38 INFO: Datasets program called with:
117
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
118
+ 2025-12-03 16:43:38 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
119
+
120
+ 2025-12-03 16:43:38 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
121
+ 2025-12-03 16:43:38 DEBUG: Processing parameter "processors"...
122
+ 2025-12-03 16:43:38 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
123
+ 2025-12-03 16:43:38 INFO: Downloading these customized packages for language: sv (Swedish)...
124
+ =======================
125
+ | Processor | Package |
126
+ -----------------------
127
+ =======================
128
+
129
+ 2025-12-03 16:43:38 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
130
+ 2025-12-03 16:43:38 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
131
+ 2025-12-03 16:43:38 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
132
+ 2025-12-03 16:43:38 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
133
+ Augmented 628 quotes: Counter({'„”': 76, '„“': 72, '””': 70, '«»': 69, '″″': 68, '""': 64, '「」': 59, '““': 55, '»«': 50, '《》': 45})
134
+ 2025-12-03 16:43:45 INFO: Running tagger to retag /local/tmp.5440223/tmpaf_ddzqh/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
135
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmpaf_ddzqh/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
136
+ 2025-12-03 16:43:45 INFO: Running tagger in predict mode
137
+ 2025-12-03 16:43:45 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
138
+ 2025-12-03 16:43:47 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
139
+ 2025-12-03 16:43:47 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
140
+ 2025-12-03 16:43:47 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
141
+ 2025-12-03 16:43:47 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
142
+ 2025-12-03 16:43:47 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
143
+ 2025-12-03 16:43:50 INFO: Loading data with batch size 250...
144
+ 2025-12-03 16:45:08 INFO: Start evaluation...
145
+ 2025-12-03 16:51:48 INFO: UPOS XPOS UFeats AllTags
146
+ 2025-12-03 16:51:48 INFO: 67.24 45.44 61.68 45.04
147
+ 2025-12-03 16:51:48 INFO: POS Tagger score: sv_diachronic 45.04
148
+ 2025-12-03 16:51:49 INFO: Running tagger to retag /local/tmp.5440223/tmpaf_ddzqh/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
149
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmpaf_ddzqh/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
150
+ 2025-12-03 16:51:49 INFO: Running tagger in predict mode
151
+ 2025-12-03 16:51:49 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
152
+ 2025-12-03 16:51:51 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
153
+ 2025-12-03 16:51:51 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
154
+ 2025-12-03 16:51:51 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
155
+ 2025-12-03 16:51:51 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
156
+ 2025-12-03 16:51:51 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
157
+ 2025-12-03 16:51:51 INFO: Loading data with batch size 250...
158
+ 2025-12-03 16:51:51 INFO: Start evaluation...
159
+ 2025-12-03 16:51:52 INFO: UPOS XPOS UFeats AllTags
160
+ 2025-12-03 16:51:52 INFO: 93.32 90.84 93.32 85.64
161
+ 2025-12-03 16:51:52 INFO: POS Tagger score: sv_diachronic 85.64
162
+ 2025-12-03 16:51:52 INFO: Running tagger to retag /local/tmp.5440223/tmpaf_ddzqh/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
163
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmpaf_ddzqh/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
164
+ 2025-12-03 16:51:52 INFO: Running tagger in predict mode
165
+ 2025-12-03 16:51:52 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
166
+ 2025-12-03 16:51:53 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
167
+ 2025-12-03 16:51:53 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
168
+ 2025-12-03 16:51:53 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
169
+ 2025-12-03 16:51:53 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
170
+ 2025-12-03 16:51:54 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
171
+ 2025-12-03 16:51:54 INFO: Loading data with batch size 250...
172
+ 2025-12-03 16:51:54 INFO: Start evaluation...
173
+ 2025-12-03 16:51:54 INFO: UPOS XPOS UFeats AllTags
174
+ 2025-12-03 16:51:54 INFO: 93.14 96.78 95.32 90.28
175
+ 2025-12-03 16:51:54 INFO: POS Tagger score: sv_diachronic 90.28
176
+ Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
177
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5440223/tmpaf_ddzqh/sv_diachronic.train.gold.conllu
178
+ Swapped 'w1, w2' for 'w1 ,w2' 552 times
179
+ Added 780 new sentences with asdf, zzzz -> asdf,zzzz
180
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5440223/tmpaf_ddzqh/sv_diachronic.dev.gold.conllu
181
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5440223/tmpaf_ddzqh/sv_diachronic.test.gold.conllu
182
+ Running stanza dependency parser training…
183
+ 2025-12-03 16:52:18 INFO: Training program called with:
184
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt --batch_size 32 --dropout 0.33
185
+ 2025-12-03 16:52:18 DEBUG: UD_Swedish-diachronic: sv_diachronic
186
+ 2025-12-03 16:52:18 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
187
+ 2025-12-03 16:52:18 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
188
+ 2025-12-03 16:52:18 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
189
+ 2025-12-03 16:52:18 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
190
+ 2025-12-03 16:52:18 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
191
+ 2025-12-03 16:52:18 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
192
+ 2025-12-03 16:52:18 INFO: Running parser in train mode
193
+ 2025-12-03 16:52:18 INFO: Using pretrained contextualized char embedding
194
+ 2025-12-03 16:52:18 INFO: Loading data with batch size 32...
195
+ 2025-12-03 16:52:33 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 112061
196
+ 2025-12-03 16:52:33 INFO: Original data size: 112061
197
+ 2025-12-03 16:52:35 INFO: Augmented data size: 116654
198
+ 2025-12-03 16:53:16 WARNING: sv_diachronic is not a known dataset. Examining the data to choose which xpos vocab to use
199
+ 2025-12-03 16:53:16 INFO: Original length = 116654
200
+ 2025-12-03 16:53:16 INFO: Filtered length = 116654
201
+ 2025-12-03 16:53:49 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
202
+ 2025-12-03 16:53:59 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
203
+ 2025-12-03 16:54:26 DEBUG: 75073 batches created.
204
+ 2025-12-03 16:54:26 DEBUG: 9 batches created.
205
+ 2025-12-03 16:54:26 INFO: Training parser...
206
+ 2025-12-03 16:54:26 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
207
+ 2025-12-03 16:54:26 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
208
+ 2025-12-03 16:54:26 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
209
+ 2025-12-03 16:54:27 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
210
+ 2025-12-03 16:54:31 INFO: Finished STEP 20/50000, loss = 0.000000 (0.036 sec/batch), lr: 0.003000
211
+ 2025-12-03 16:54:31 INFO: Finished STEP 40/50000, loss = 0.000000 (0.034 sec/batch), lr: 0.003000
212
+ 2025-12-03 16:54:32 INFO: Finished STEP 60/50000, loss = 40756.410156 (0.033 sec/batch), lr: 0.003000
213
+ 2025-12-03 16:54:33 INFO: Finished STEP 80/50000, loss = 0.000000 (0.034 sec/batch), lr: 0.003000
214
+ 2025-12-03 16:54:33 INFO: Finished STEP 100/50000, loss = 1.985795 (0.033 sec/batch), lr: 0.003000
215
+ 2025-12-03 16:54:33 INFO: Evaluating on dev set...
216
+ 2025-12-03 16:54:34 INFO: LAS MLAS BLEX
217
+ 2025-12-03 16:54:34 INFO: 2.23 0.82 1.64
218
+ 2025-12-03 16:54:34 INFO: step 100: train_loss = 973.415672, dev_score = 0.0223
219
+ 2025-12-03 16:54:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
220
+ 2025-12-03 16:54:34 INFO: new best model saved.
221
+ 2025-12-03 16:54:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
222
+ 2025-12-03 16:54:35 INFO: new model checkpoint saved.
223
+ 2025-12-03 16:54:35 INFO: Finished STEP 120/50000, loss = 1.061848 (0.034 sec/batch), lr: 0.003000
224
+ 2025-12-03 16:54:36 INFO: Finished STEP 140/50000, loss = 1.775142 (0.034 sec/batch), lr: 0.003000
225
+ 2025-12-03 16:54:37 INFO: Finished STEP 160/50000, loss = 0.801000 (0.032 sec/batch), lr: 0.003000
226
+ 2025-12-03 16:54:38 INFO: Finished STEP 180/50000, loss = 0.948637 (0.032 sec/batch), lr: 0.003000
227
+ 2025-12-03 16:54:38 INFO: Finished STEP 200/50000, loss = 0.647672 (0.034 sec/batch), lr: 0.003000
228
+ 2025-12-03 16:54:38 INFO: Evaluating on dev set...
229
+ 2025-12-03 16:54:39 INFO: LAS MLAS BLEX
230
+ 2025-12-03 16:54:39 INFO: 2.23 0.51 1.02
231
+ 2025-12-03 16:54:39 INFO: step 200: train_loss = 1.180679, dev_score = 0.0223
232
+ 2025-12-03 16:54:39 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
233
+ 2025-12-03 16:54:39 INFO: new best model saved.
234
+ 2025-12-03 16:54:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
235
+ 2025-12-03 16:54:40 INFO: new model checkpoint saved.
236
+ 2025-12-03 16:54:40 INFO: Finished STEP 220/50000, loss = 1.209441 (0.033 sec/batch), lr: 0.003000
237
+ 2025-12-03 16:54:41 INFO: Finished STEP 240/50000, loss = 0.651829 (0.033 sec/batch), lr: 0.003000
238
+ 2025-12-03 16:54:42 INFO: Finished STEP 260/50000, loss = 1.160513 (0.036 sec/batch), lr: 0.003000
239
+ 2025-12-03 16:54:43 INFO: Finished STEP 280/50000, loss = 0.839569 (0.035 sec/batch), lr: 0.003000
240
+ 2025-12-03 16:54:43 INFO: Finished STEP 300/50000, loss = 2.961314 (0.034 sec/batch), lr: 0.003000
241
+ 2025-12-03 16:54:43 INFO: Evaluating on dev set...
242
+ 2025-12-03 16:54:44 INFO: LAS MLAS BLEX
243
+ 2025-12-03 16:54:44 INFO: 2.23 0.34 0.69
244
+ 2025-12-03 16:54:44 INFO: step 300: train_loss = 1.561353, dev_score = 0.0223
245
+ 2025-12-03 16:54:44 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
246
+ 2025-12-03 16:54:44 INFO: new best model saved.
247
+ 2025-12-03 16:54:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
248
+ 2025-12-03 16:54:45 INFO: new model checkpoint saved.
249
+ 2025-12-03 16:54:45 INFO: Finished STEP 320/50000, loss = 1.887101 (0.034 sec/batch), lr: 0.003000
250
+ 2025-12-03 16:54:46 INFO: Finished STEP 340/50000, loss = 1.244869 (0.033 sec/batch), lr: 0.003000
251
+ 2025-12-03 16:54:47 INFO: Finished STEP 360/50000, loss = 1.526011 (0.039 sec/batch), lr: 0.003000
252
+ 2025-12-03 16:54:48 INFO: Finished STEP 380/50000, loss = 1.167714 (0.033 sec/batch), lr: 0.003000
253
+ 2025-12-03 16:54:48 INFO: Finished STEP 400/50000, loss = 1.603913 (0.034 sec/batch), lr: 0.003000
254
+ 2025-12-03 16:54:48 INFO: Evaluating on dev set...
255
+ 2025-12-03 16:54:49 INFO: LAS MLAS BLEX
256
+ 2025-12-03 16:54:49 INFO: 6.93 5.33 5.74
257
+ 2025-12-03 16:54:49 INFO: step 400: train_loss = 1.735054, dev_score = 0.0693
258
+ 2025-12-03 16:54:49 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
259
+ 2025-12-03 16:54:49 INFO: new best model saved.
260
+ 2025-12-03 16:54:50 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
261
+ 2025-12-03 16:54:50 INFO: new model checkpoint saved.
262
+ 2025-12-03 16:54:51 INFO: Finished STEP 420/50000, loss = 1.496035 (0.036 sec/batch), lr: 0.003000
263
+ 2025-12-03 16:54:51 INFO: Finished STEP 440/50000, loss = 1.360082 (0.034 sec/batch), lr: 0.003000
264
+ 2025-12-03 16:54:52 INFO: Finished STEP 460/50000, loss = 1.136768 (0.033 sec/batch), lr: 0.003000
265
+ 2025-12-03 16:54:53 INFO: Finished STEP 480/50000, loss = 1.158461 (0.033 sec/batch), lr: 0.003000
266
+ 2025-12-03 16:54:53 INFO: Finished STEP 500/50000, loss = 1.181382 (0.033 sec/batch), lr: 0.003000
267
+ 2025-12-03 16:54:53 INFO: Evaluating on dev set...
268
+ 2025-12-03 16:54:54 INFO: LAS MLAS BLEX
269
+ 2025-12-03 16:54:54 INFO: 5.69 3.75 4.10
270
+ 2025-12-03 16:54:54 INFO: step 500: train_loss = 1.528116, dev_score = 0.0569
271
+ 2025-12-03 16:54:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
272
+ 2025-12-03 16:54:54 INFO: new model checkpoint saved.
273
+ 2025-12-03 16:54:55 INFO: Finished STEP 520/50000, loss = 2.018800 (0.035 sec/batch), lr: 0.003000
274
+ 2025-12-03 16:54:56 INFO: Finished STEP 540/50000, loss = 1.521049 (0.037 sec/batch), lr: 0.003000
275
+ 2025-12-03 16:54:57 INFO: Finished STEP 560/50000, loss = 1.107444 (0.034 sec/batch), lr: 0.003000
276
+ 2025-12-03 16:54:57 INFO: Finished STEP 580/50000, loss = 1.231447 (0.039 sec/batch), lr: 0.003000
277
+ 2025-12-03 16:54:58 INFO: Finished STEP 600/50000, loss = 0.928001 (0.034 sec/batch), lr: 0.003000
278
+ 2025-12-03 16:54:58 INFO: Evaluating on dev set...
279
+ 2025-12-03 16:54:59 INFO: LAS MLAS BLEX
280
+ 2025-12-03 16:54:59 INFO: 18.32 9.04 10.13
281
+ 2025-12-03 16:54:59 INFO: step 600: train_loss = 1.459400, dev_score = 0.1832
282
+ 2025-12-03 16:54:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
283
+ 2025-12-03 16:54:59 INFO: new best model saved.
284
+ 2025-12-03 16:55:00 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
285
+ 2025-12-03 16:55:00 INFO: new model checkpoint saved.
286
+ 2025-12-03 16:55:00 INFO: Finished STEP 620/50000, loss = 2.224530 (0.034 sec/batch), lr: 0.003000
287
+ 2025-12-03 16:55:01 INFO: Finished STEP 640/50000, loss = 2.260295 (0.035 sec/batch), lr: 0.003000
288
+ 2025-12-03 16:55:02 INFO: Finished STEP 660/50000, loss = 2.073965 (0.037 sec/batch), lr: 0.003000
289
+ 2025-12-03 16:55:02 INFO: Finished STEP 680/50000, loss = 2.039528 (0.036 sec/batch), lr: 0.003000
290
+ 2025-12-03 16:55:03 INFO: Finished STEP 700/50000, loss = 2.508281 (0.035 sec/batch), lr: 0.003000
291
+ 2025-12-03 16:55:03 INFO: Evaluating on dev set...
292
+ 2025-12-03 16:55:04 INFO: LAS MLAS BLEX
293
+ 2025-12-03 16:55:04 INFO: 17.33 5.14 5.50
294
+ 2025-12-03 16:55:04 INFO: step 700: train_loss = 2.354263, dev_score = 0.1733
295
+ 2025-12-03 16:55:04 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
296
+ 2025-12-03 16:55:04 INFO: new model checkpoint saved.
297
+ 2025-12-03 16:55:05 INFO: Finished STEP 720/50000, loss = 1.973946 (0.036 sec/batch), lr: 0.003000
298
+ 2025-12-03 16:55:06 INFO: Finished STEP 740/50000, loss = 1.847519 (0.035 sec/batch), lr: 0.003000
299
+ 2025-12-03 16:55:07 INFO: Finished STEP 760/50000, loss = 2.336380 (0.035 sec/batch), lr: 0.003000
300
+ 2025-12-03 16:55:07 INFO: Finished STEP 780/50000, loss = 1.968456 (0.034 sec/batch), lr: 0.003000
301
+ 2025-12-03 16:55:08 INFO: Finished STEP 800/50000, loss = 2.258449 (0.036 sec/batch), lr: 0.003000
302
+ 2025-12-03 16:55:08 INFO: Evaluating on dev set...
303
+ 2025-12-03 16:55:09 INFO: LAS MLAS BLEX
304
+ 2025-12-03 16:55:09 INFO: 16.58 6.40 7.16
305
+ 2025-12-03 16:55:09 INFO: step 800: train_loss = 2.111074, dev_score = 0.1658
306
+ 2025-12-03 16:55:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
307
+ 2025-12-03 16:55:09 INFO: new model checkpoint saved.
308
+ 2025-12-03 16:55:10 INFO: Finished STEP 820/50000, loss = 2.679285 (0.036 sec/batch), lr: 0.003000
309
+ 2025-12-03 16:55:11 INFO: Finished STEP 840/50000, loss = 1.837710 (0.035 sec/batch), lr: 0.003000
310
+ 2025-12-03 16:55:11 INFO: Finished STEP 860/50000, loss = 1.455105 (0.035 sec/batch), lr: 0.003000
311
+ 2025-12-03 16:55:12 INFO: Finished STEP 880/50000, loss = 1.590474 (0.036 sec/batch), lr: 0.003000
312
+ 2025-12-03 16:55:13 INFO: Finished STEP 900/50000, loss = 1.833156 (0.034 sec/batch), lr: 0.003000
313
+ 2025-12-03 16:55:13 INFO: Evaluating on dev set...
314
+ 2025-12-03 16:55:14 INFO: LAS MLAS BLEX
315
+ 2025-12-03 16:55:14 INFO: 20.54 5.35 5.74
316
+ 2025-12-03 16:55:14 INFO: step 900: train_loss = 1.979503, dev_score = 0.2054
317
+ 2025-12-03 16:55:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
318
+ 2025-12-03 16:55:14 INFO: new best model saved.
319
+ 2025-12-03 16:55:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
320
+ 2025-12-03 16:55:14 INFO: new model checkpoint saved.
321
+ 2025-12-03 16:55:15 INFO: Finished STEP 920/50000, loss = 1.408719 (0.035 sec/batch), lr: 0.003000
322
+ 2025-12-03 16:55:16 INFO: Finished STEP 940/50000, loss = 2.133405 (0.034 sec/batch), lr: 0.003000
323
+ 2025-12-03 16:55:17 INFO: Finished STEP 960/50000, loss = 1.914302 (0.034 sec/batch), lr: 0.003000
324
+ 2025-12-03 16:55:17 INFO: Finished STEP 980/50000, loss = 2.111408 (0.034 sec/batch), lr: 0.003000
325
+ 2025-12-03 16:55:18 INFO: Finished STEP 1000/50000, loss = 2.271649 (0.036 sec/batch), lr: 0.003000
326
+ 2025-12-03 16:55:18 INFO: Evaluating on dev set...
327
+ 2025-12-03 16:55:19 INFO: LAS MLAS BLEX
328
+ 2025-12-03 16:55:19 INFO: 16.09 9.06 10.57
329
+ 2025-12-03 16:55:19 INFO: step 1000: train_loss = 1.958875, dev_score = 0.1609
330
+ 2025-12-03 16:55:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
331
+ 2025-12-03 16:55:19 INFO: new model checkpoint saved.
332
+ 2025-12-03 16:55:20 INFO: Finished STEP 1020/50000, loss = 1.779836 (0.035 sec/batch), lr: 0.003000
333
+ 2025-12-03 16:55:21 INFO: Finished STEP 1040/50000, loss = 2.605173 (0.038 sec/batch), lr: 0.003000
334
+ 2025-12-03 16:55:21 INFO: Finished STEP 1060/50000, loss = 1.898196 (0.038 sec/batch), lr: 0.003000
335
+ 2025-12-03 16:55:22 INFO: Finished STEP 1080/50000, loss = 1.787909 (0.034 sec/batch), lr: 0.003000
336
+ 2025-12-03 16:55:23 INFO: Finished STEP 1100/50000, loss = 1.449179 (0.034 sec/batch), lr: 0.003000
337
+ 2025-12-03 16:55:23 INFO: Evaluating on dev set...
338
+ 2025-12-03 16:55:23 INFO: LAS MLAS BLEX
339
+ 2025-12-03 16:55:23 INFO: 23.51 11.35 12.52
340
+ 2025-12-03 16:55:23 INFO: step 1100: train_loss = 1.964432, dev_score = 0.2351
341
+ 2025-12-03 16:55:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
342
+ 2025-12-03 16:55:24 INFO: new best model saved.
343
+ 2025-12-03 16:55:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
344
+ 2025-12-03 16:55:24 INFO: new model checkpoint saved.
345
+ 2025-12-03 16:55:25 INFO: Finished STEP 1120/50000, loss = 2.237896 (0.037 sec/batch), lr: 0.003000
346
+ 2025-12-03 16:55:26 INFO: Finished STEP 1140/50000, loss = 2.699857 (0.037 sec/batch), lr: 0.003000
347
+ 2025-12-03 16:55:27 INFO: Finished STEP 1160/50000, loss = 1.600584 (0.036 sec/batch), lr: 0.003000
348
+ 2025-12-03 16:55:27 INFO: Finished STEP 1180/50000, loss = 1.557849 (0.037 sec/batch), lr: 0.003000
349
+ 2025-12-03 16:55:28 INFO: Finished STEP 1200/50000, loss = 2.624414 (0.037 sec/batch), lr: 0.003000
350
+ 2025-12-03 16:55:28 INFO: Evaluating on dev set...
351
+ 2025-12-03 16:55:29 INFO: LAS MLAS BLEX
352
+ 2025-12-03 16:55:29 INFO: 27.97 12.30 14.29
353
+ 2025-12-03 16:55:29 INFO: step 1200: train_loss = 2.363800, dev_score = 0.2797
354
+ 2025-12-03 16:55:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
355
+ 2025-12-03 16:55:29 INFO: new best model saved.
356
+ 2025-12-03 16:55:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
357
+ 2025-12-03 16:55:30 INFO: new model checkpoint saved.
358
+ 2025-12-03 16:55:30 INFO: Finished STEP 1220/50000, loss = 1.706681 (0.037 sec/batch), lr: 0.003000
359
+ 2025-12-03 16:55:31 INFO: Finished STEP 1240/50000, loss = 1.934880 (0.036 sec/batch), lr: 0.003000
360
+ 2025-12-03 16:55:32 INFO: Finished STEP 1260/50000, loss = 2.524523 (0.037 sec/batch), lr: 0.003000
361
+ 2025-12-03 16:55:33 INFO: Finished STEP 1280/50000, loss = 2.054730 (0.036 sec/batch), lr: 0.003000
362
+ 2025-12-03 16:55:33 INFO: Finished STEP 1300/50000, loss = 2.073152 (0.036 sec/batch), lr: 0.003000
363
+ 2025-12-03 16:55:33 INFO: Evaluating on dev set...
364
+ 2025-12-03 16:55:34 INFO: LAS MLAS BLEX
365
+ 2025-12-03 16:55:34 INFO: 32.67 12.73 14.37
366
+ 2025-12-03 16:55:34 INFO: step 1300: train_loss = 2.413291, dev_score = 0.3267
367
+ 2025-12-03 16:55:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
368
+ 2025-12-03 16:55:34 INFO: new best model saved.
369
+ 2025-12-03 16:55:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
370
+ 2025-12-03 16:55:35 INFO: new model checkpoint saved.
371
+ 2025-12-03 16:55:36 INFO: Finished STEP 1320/50000, loss = 1.758738 (0.035 sec/batch), lr: 0.003000
372
+ 2025-12-03 16:55:36 INFO: Finished STEP 1340/50000, loss = 2.503182 (0.035 sec/batch), lr: 0.003000
373
+ 2025-12-03 16:55:37 INFO: Finished STEP 1360/50000, loss = 1.706251 (0.036 sec/batch), lr: 0.003000
374
+ 2025-12-03 16:55:38 INFO: Finished STEP 1380/50000, loss = 2.229746 (0.037 sec/batch), lr: 0.003000
375
+ 2025-12-03 16:55:39 INFO: Finished STEP 1400/50000, loss = 2.225713 (0.036 sec/batch), lr: 0.003000
376
+ 2025-12-03 16:55:39 INFO: Evaluating on dev set...
377
+ 2025-12-03 16:55:39 INFO: LAS MLAS BLEX
378
+ 2025-12-03 16:55:39 INFO: 31.44 10.55 12.66
379
+ 2025-12-03 16:55:39 INFO: step 1400: train_loss = 2.246589, dev_score = 0.3144
380
+ 2025-12-03 16:55:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
381
+ 2025-12-03 16:55:40 INFO: new model checkpoint saved.
382
+ 2025-12-03 16:55:41 INFO: Finished STEP 1420/50000, loss = 2.337524 (0.036 sec/batch), lr: 0.003000
383
+ 2025-12-03 16:55:41 INFO: Finished STEP 1440/50000, loss = 2.199153 (0.035 sec/batch), lr: 0.003000
384
+ 2025-12-03 16:55:42 INFO: Finished STEP 1460/50000, loss = 2.390700 (0.036 sec/batch), lr: 0.003000
385
+ 2025-12-03 16:55:43 INFO: Finished STEP 1480/50000, loss = 2.104558 (0.036 sec/batch), lr: 0.003000
386
+ 2025-12-03 16:55:44 INFO: Finished STEP 1500/50000, loss = 2.326790 (0.035 sec/batch), lr: 0.003000
387
+ 2025-12-03 16:55:44 INFO: Evaluating on dev set...
388
+ 2025-12-03 16:55:44 INFO: LAS MLAS BLEX
389
+ 2025-12-03 16:55:44 INFO: 33.17 15.32 16.15
390
+ 2025-12-03 16:55:44 INFO: step 1500: train_loss = 2.308034, dev_score = 0.3317
391
+ 2025-12-03 16:55:44 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
392
+ 2025-12-03 16:55:44 INFO: new best model saved.
393
+ 2025-12-03 16:55:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
394
+ 2025-12-03 16:55:45 INFO: new model checkpoint saved.
395
+ 2025-12-03 16:55:46 INFO: Finished STEP 1520/50000, loss = 3.384029 (0.036 sec/batch), lr: 0.003000
396
+ 2025-12-03 16:55:47 INFO: Finished STEP 1540/50000, loss = 2.028357 (0.035 sec/batch), lr: 0.003000
397
+ 2025-12-03 16:55:47 INFO: Finished STEP 1560/50000, loss = 2.677545 (0.037 sec/batch), lr: 0.003000
398
+ 2025-12-03 16:55:48 INFO: Finished STEP 1580/50000, loss = 2.399013 (0.036 sec/batch), lr: 0.003000
399
+ 2025-12-03 16:55:49 INFO: Finished STEP 1600/50000, loss = 2.604865 (0.035 sec/batch), lr: 0.003000
400
+ 2025-12-03 16:55:49 INFO: Evaluating on dev set...
401
+ 2025-12-03 16:55:49 INFO: LAS MLAS BLEX
402
+ 2025-12-03 16:55:49 INFO: 32.18 15.29 17.71
403
+ 2025-12-03 16:55:49 INFO: step 1600: train_loss = 2.230821, dev_score = 0.3218
404
+ 2025-12-03 16:55:50 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
405
+ 2025-12-03 16:55:50 INFO: new model checkpoint saved.
406
+ 2025-12-03 16:55:51 INFO: Finished STEP 1620/50000, loss = 2.209406 (0.035 sec/batch), lr: 0.003000
407
+ 2025-12-03 16:55:51 INFO: Finished STEP 1640/50000, loss = 2.806011 (0.035 sec/batch), lr: 0.003000
408
+ 2025-12-03 16:55:52 INFO: Finished STEP 1660/50000, loss = 1.941491 (0.037 sec/batch), lr: 0.003000
409
+ 2025-12-03 16:55:53 INFO: Finished STEP 1680/50000, loss = 1.985415 (0.038 sec/batch), lr: 0.003000
410
+ 2025-12-03 16:55:54 INFO: Finished STEP 1700/50000, loss = 2.132746 (0.036 sec/batch), lr: 0.003000
411
+ 2025-12-03 16:55:54 INFO: Evaluating on dev set...
412
+ 2025-12-03 16:55:54 INFO: LAS MLAS BLEX
413
+ 2025-12-03 16:55:54 INFO: 32.67 18.30 19.15
414
+ 2025-12-03 16:55:54 INFO: step 1700: train_loss = 2.178722, dev_score = 0.3267
415
+ 2025-12-03 16:55:55 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
416
+ 2025-12-03 16:55:55 INFO: new model checkpoint saved.
417
+ 2025-12-03 16:55:56 INFO: Finished STEP 1720/50000, loss = 2.132302 (0.037 sec/batch), lr: 0.003000
418
+ 2025-12-03 16:55:56 INFO: Finished STEP 1740/50000, loss = 1.370333 (0.039 sec/batch), lr: 0.003000
419
+ 2025-12-03 16:55:57 INFO: Finished STEP 1760/50000, loss = 1.751216 (0.038 sec/batch), lr: 0.003000
420
+ 2025-12-03 16:55:58 INFO: Finished STEP 1780/50000, loss = 1.879204 (0.039 sec/batch), lr: 0.003000
421
+ 2025-12-03 16:55:59 INFO: Finished STEP 1800/50000, loss = 2.273577 (0.037 sec/batch), lr: 0.003000
422
+ 2025-12-03 16:55:59 INFO: Evaluating on dev set...
423
+ 2025-12-03 16:55:59 INFO: LAS MLAS BLEX
424
+ 2025-12-03 16:55:59 INFO: 41.83 25.16 26.83
425
+ 2025-12-03 16:55:59 INFO: step 1800: train_loss = 2.245952, dev_score = 0.4183
426
+ 2025-12-03 16:56:00 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
427
+ 2025-12-03 16:56:00 INFO: new best model saved.
428
+ 2025-12-03 16:56:00 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
429
+ 2025-12-03 16:56:00 INFO: new model checkpoint saved.
430
+ 2025-12-03 16:56:01 INFO: Finished STEP 1820/50000, loss = 1.773414 (0.036 sec/batch), lr: 0.003000
431
+ 2025-12-03 16:56:02 INFO: Finished STEP 1840/50000, loss = 2.022996 (0.036 sec/batch), lr: 0.003000
432
+ 2025-12-03 16:56:03 INFO: Finished STEP 1860/50000, loss = 1.621772 (0.035 sec/batch), lr: 0.003000
433
+ 2025-12-03 16:56:03 INFO: Finished STEP 1880/50000, loss = 1.765723 (0.036 sec/batch), lr: 0.003000
434
+ 2025-12-03 16:56:04 INFO: Finished STEP 1900/50000, loss = 2.141253 (0.036 sec/batch), lr: 0.003000
435
+ 2025-12-03 16:56:04 INFO: Evaluating on dev set...
436
+ 2025-12-03 16:56:05 INFO: LAS MLAS BLEX
437
+ 2025-12-03 16:56:05 INFO: 27.48 21.03 23.61
438
+ 2025-12-03 16:56:05 INFO: step 1900: train_loss = 2.128514, dev_score = 0.2748
439
+ 2025-12-03 16:56:05 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
440
+ 2025-12-03 16:56:05 INFO: new model checkpoint saved.
441
+ 2025-12-03 16:56:06 INFO: Finished STEP 1920/50000, loss = 1.849969 (0.036 sec/batch), lr: 0.003000
442
+ 2025-12-03 16:56:07 INFO: Finished STEP 1940/50000, loss = 1.526435 (0.036 sec/batch), lr: 0.003000
443
+ 2025-12-03 16:56:07 INFO: Finished STEP 1960/50000, loss = 2.894019 (0.039 sec/batch), lr: 0.003000
444
+ 2025-12-03 16:56:08 INFO: Finished STEP 1980/50000, loss = 3.949821 (0.039 sec/batch), lr: 0.003000
445
+ 2025-12-03 16:56:09 INFO: Finished STEP 2000/50000, loss = 1.907542 (0.037 sec/batch), lr: 0.003000
446
+ 2025-12-03 16:56:09 INFO: Evaluating on dev set...
447
+ 2025-12-03 16:56:09 INFO: LAS MLAS BLEX
448
+ 2025-12-03 16:56:09 INFO: 41.58 26.22 29.60
449
+ 2025-12-03 16:56:09 INFO: step 2000: train_loss = 2.400021, dev_score = 0.4158
450
+ 2025-12-03 16:56:10 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
451
+ 2025-12-03 16:56:10 INFO: new model checkpoint saved.
452
+ 2025-12-03 16:56:11 INFO: Finished STEP 2020/50000, loss = 1.865386 (0.035 sec/batch), lr: 0.003000
453
+ 2025-12-03 16:56:12 INFO: Finished STEP 2040/50000, loss = 2.452761 (0.035 sec/batch), lr: 0.003000
454
+ 2025-12-03 16:56:12 INFO: Finished STEP 2060/50000, loss = 2.583950 (0.036 sec/batch), lr: 0.003000
455
+ 2025-12-03 16:56:13 INFO: Finished STEP 2080/50000, loss = 3.295110 (0.035 sec/batch), lr: 0.003000
456
+ 2025-12-03 16:56:14 INFO: Finished STEP 2100/50000, loss = 2.720312 (0.038 sec/batch), lr: 0.003000
457
+ 2025-12-03 16:56:14 INFO: Evaluating on dev set...
458
+ 2025-12-03 16:56:14 INFO: LAS MLAS BLEX
459
+ 2025-12-03 16:56:14 INFO: 43.07 27.48 30.87
460
+ 2025-12-03 16:56:14 INFO: step 2100: train_loss = 2.531355, dev_score = 0.4307
461
+ 2025-12-03 16:56:15 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
462
+ 2025-12-03 16:56:15 INFO: new best model saved.
463
+ 2025-12-03 16:56:15 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
464
+ 2025-12-03 16:56:15 INFO: new model checkpoint saved.
465
+ 2025-12-03 16:56:16 INFO: Finished STEP 2120/50000, loss = 2.415540 (0.036 sec/batch), lr: 0.003000
466
+ 2025-12-03 16:56:17 INFO: Finished STEP 2140/50000, loss = 1.821096 (0.035 sec/batch), lr: 0.003000
467
+ 2025-12-03 16:56:17 INFO: Finished STEP 2160/50000, loss = 1.902365 (0.038 sec/batch), lr: 0.003000
468
+ 2025-12-03 16:56:18 INFO: Finished STEP 2180/50000, loss = 2.743435 (0.038 sec/batch), lr: 0.003000
469
+ 2025-12-03 16:56:19 INFO: Finished STEP 2200/50000, loss = 2.577427 (0.037 sec/batch), lr: 0.003000
470
+ 2025-12-03 16:56:19 INFO: Evaluating on dev set...
471
+ 2025-12-03 16:56:20 INFO: LAS MLAS BLEX
472
+ 2025-12-03 16:56:20 INFO: 40.59 26.83 29.35
473
+ 2025-12-03 16:56:20 INFO: step 2200: train_loss = 2.488173, dev_score = 0.4059
474
+ 2025-12-03 16:56:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
475
+ 2025-12-03 16:56:20 INFO: new model checkpoint saved.
476
+ 2025-12-03 16:56:21 INFO: Finished STEP 2220/50000, loss = 2.148330 (0.040 sec/batch), lr: 0.003000
477
+ 2025-12-03 16:56:22 INFO: Finished STEP 2240/50000, loss = 3.634011 (0.038 sec/batch), lr: 0.003000
478
+ 2025-12-03 16:56:22 INFO: Finished STEP 2260/50000, loss = 2.595282 (0.037 sec/batch), lr: 0.003000
479
+ 2025-12-03 16:56:23 INFO: Finished STEP 2280/50000, loss = 3.153853 (0.035 sec/batch), lr: 0.003000
480
+ 2025-12-03 16:56:24 INFO: Finished STEP 2300/50000, loss = 2.201555 (0.037 sec/batch), lr: 0.003000
481
+ 2025-12-03 16:56:24 INFO: Evaluating on dev set...
482
+ 2025-12-03 16:56:24 INFO: LAS MLAS BLEX
483
+ 2025-12-03 16:56:24 INFO: 38.86 20.81 25.05
484
+ 2025-12-03 16:56:24 INFO: step 2300: train_loss = 2.601222, dev_score = 0.3886
485
+ 2025-12-03 16:56:25 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
486
+ 2025-12-03 16:56:25 INFO: new model checkpoint saved.
487
+ 2025-12-03 16:56:26 INFO: Finished STEP 2320/50000, loss = 7.203435 (0.037 sec/batch), lr: 0.003000
488
+ 2025-12-03 16:56:27 INFO: Finished STEP 2340/50000, loss = 1.763414 (0.036 sec/batch), lr: 0.003000
489
+ 2025-12-03 16:56:27 INFO: Finished STEP 2360/50000, loss = 2.256621 (0.037 sec/batch), lr: 0.003000
490
+ 2025-12-03 16:56:28 INFO: Finished STEP 2380/50000, loss = 2.804393 (0.036 sec/batch), lr: 0.003000
491
+ 2025-12-03 16:56:29 INFO: Finished STEP 2400/50000, loss = 1.426095 (0.039 sec/batch), lr: 0.003000
492
+ 2025-12-03 16:56:29 INFO: Evaluating on dev set...
493
+ 2025-12-03 16:56:29 INFO: LAS MLAS BLEX
494
+ 2025-12-03 16:56:29 INFO: 44.06 28.69 31.22
495
+ 2025-12-03 16:56:29 INFO: step 2400: train_loss = 2.516615, dev_score = 0.4406
496
+ 2025-12-03 16:56:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
497
+ 2025-12-03 16:56:30 INFO: new best model saved.
498
+ 2025-12-03 16:56:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
499
+ 2025-12-03 16:56:30 INFO: new model checkpoint saved.
500
+ 2025-12-03 16:56:31 INFO: Finished STEP 2420/50000, loss = 1.585505 (0.036 sec/batch), lr: 0.003000
501
+ 2025-12-03 16:56:32 INFO: Finished STEP 2440/50000, loss = 2.617753 (0.037 sec/batch), lr: 0.003000
502
+ 2025-12-03 16:56:33 INFO: Finished STEP 2460/50000, loss = 2.995166 (0.036 sec/batch), lr: 0.003000
503
+ 2025-12-03 16:56:33 INFO: Finished STEP 2480/50000, loss = 1.925491 (0.036 sec/batch), lr: 0.003000
504
+ 2025-12-03 16:56:34 INFO: Finished STEP 2500/50000, loss = 2.408792 (0.036 sec/batch), lr: 0.003000
505
+ 2025-12-03 16:56:34 INFO: Evaluating on dev set...
506
+ 2025-12-03 16:56:35 INFO: LAS MLAS BLEX
507
+ 2025-12-03 16:56:35 INFO: 43.81 27.43 32.07
508
+ 2025-12-03 16:56:35 INFO: step 2500: train_loss = 2.470480, dev_score = 0.4381
509
+ 2025-12-03 16:56:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
510
+ 2025-12-03 16:56:35 INFO: new model checkpoint saved.
511
+ 2025-12-03 16:56:36 INFO: Finished STEP 2520/50000, loss = 2.627738 (0.037 sec/batch), lr: 0.003000
512
+ 2025-12-03 16:56:37 INFO: Finished STEP 2540/50000, loss = 2.610071 (0.037 sec/batch), lr: 0.003000
513
+ 2025-12-03 16:56:37 INFO: Finished STEP 2560/50000, loss = 2.468114 (0.038 sec/batch), lr: 0.003000
514
+ 2025-12-03 16:56:38 INFO: Finished STEP 2580/50000, loss = 2.554904 (0.037 sec/batch), lr: 0.003000
515
+ 2025-12-03 16:56:39 INFO: Finished STEP 2600/50000, loss = 3.268888 (0.035 sec/batch), lr: 0.003000
516
+ 2025-12-03 16:56:39 INFO: Evaluating on dev set...
517
+ 2025-12-03 16:56:39 INFO: LAS MLAS BLEX
518
+ 2025-12-03 16:56:39 INFO: 43.56 28.57 31.98
519
+ 2025-12-03 16:56:39 INFO: step 2600: train_loss = 2.428570, dev_score = 0.4356
520
+ 2025-12-03 16:56:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
521
+ 2025-12-03 16:56:40 INFO: new model checkpoint saved.
522
+ 2025-12-03 16:56:41 INFO: Finished STEP 2620/50000, loss = 2.626238 (0.036 sec/batch), lr: 0.003000
523
+ 2025-12-03 16:56:42 INFO: Finished STEP 2640/50000, loss = 2.904417 (0.040 sec/batch), lr: 0.003000
524
+ 2025-12-03 16:56:42 INFO: Finished STEP 2660/50000, loss = 1.967971 (0.036 sec/batch), lr: 0.003000
525
+ 2025-12-03 16:56:43 INFO: Finished STEP 2680/50000, loss = 2.958992 (0.036 sec/batch), lr: 0.003000
526
+ 2025-12-03 16:56:44 INFO: Finished STEP 2700/50000, loss = 2.882056 (0.035 sec/batch), lr: 0.003000
527
+ 2025-12-03 16:56:44 INFO: Evaluating on dev set...
528
+ 2025-12-03 16:56:44 INFO: LAS MLAS BLEX
529
+ 2025-12-03 16:56:44 INFO: 43.81 27.80 30.29
530
+ 2025-12-03 16:56:44 INFO: step 2700: train_loss = 2.538486, dev_score = 0.4381
531
+ 2025-12-03 16:56:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
532
+ 2025-12-03 16:56:45 INFO: new model checkpoint saved.
533
+ 2025-12-03 16:56:46 INFO: Finished STEP 2720/50000, loss = 2.539391 (0.038 sec/batch), lr: 0.003000
534
+ 2025-12-03 16:56:47 INFO: Finished STEP 2740/50000, loss = 3.078291 (0.037 sec/batch), lr: 0.003000
535
+ 2025-12-03 16:56:47 INFO: Finished STEP 2760/50000, loss = 1.980047 (0.036 sec/batch), lr: 0.003000
536
+ 2025-12-03 16:56:48 INFO: Finished STEP 2780/50000, loss = 3.076151 (0.037 sec/batch), lr: 0.003000
537
+ 2025-12-03 16:56:49 INFO: Finished STEP 2800/50000, loss = 2.814807 (0.036 sec/batch), lr: 0.003000
538
+ 2025-12-03 16:56:49 INFO: Evaluating on dev set...
539
+ 2025-12-03 16:56:49 INFO: LAS MLAS BLEX
540
+ 2025-12-03 16:56:49 INFO: 40.10 28.45 32.70
541
+ 2025-12-03 16:56:49 INFO: step 2800: train_loss = 2.424600, dev_score = 0.4010
542
+ 2025-12-03 16:56:50 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
543
+ 2025-12-03 16:56:50 INFO: new model checkpoint saved.
544
+ 2025-12-03 16:56:51 INFO: Finished STEP 2820/50000, loss = 3.120166 (0.039 sec/batch), lr: 0.003000
545
+ 2025-12-03 16:56:52 INFO: Finished STEP 2840/50000, loss = 1.625387 (0.037 sec/batch), lr: 0.003000
546
+ 2025-12-03 16:56:52 INFO: Finished STEP 2860/50000, loss = 3.716342 (0.039 sec/batch), lr: 0.003000
547
+ 2025-12-03 16:56:53 INFO: Finished STEP 2880/50000, loss = 1.852068 (0.037 sec/batch), lr: 0.003000
548
+ 2025-12-03 16:56:54 INFO: Finished STEP 2900/50000, loss = 1.801724 (0.035 sec/batch), lr: 0.003000
549
+ 2025-12-03 16:56:54 INFO: Evaluating on dev set...
550
+ 2025-12-03 16:56:54 INFO: LAS MLAS BLEX
551
+ 2025-12-03 16:56:54 INFO: 42.08 27.91 32.56
552
+ 2025-12-03 16:56:54 INFO: step 2900: train_loss = 2.439244, dev_score = 0.4208
553
+ 2025-12-03 16:56:55 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
554
+ 2025-12-03 16:56:55 INFO: new model checkpoint saved.
555
+ 2025-12-03 16:56:56 INFO: Finished STEP 2920/50000, loss = 1.574109 (0.036 sec/batch), lr: 0.003000
556
+ 2025-12-03 16:56:57 INFO: Finished STEP 2940/50000, loss = 1.749939 (0.035 sec/batch), lr: 0.003000
557
+ 2025-12-03 16:56:57 INFO: Finished STEP 2960/50000, loss = 2.008373 (0.036 sec/batch), lr: 0.003000
558
+ 2025-12-03 16:56:58 INFO: Finished STEP 2980/50000, loss = 3.268587 (0.035 sec/batch), lr: 0.003000
559
+ 2025-12-03 16:56:59 INFO: Finished STEP 3000/50000, loss = 2.530843 (0.037 sec/batch), lr: 0.003000
560
+ 2025-12-03 16:56:59 INFO: Evaluating on dev set...
561
+ 2025-12-03 16:56:59 INFO: LAS MLAS BLEX
562
+ 2025-12-03 16:56:59 INFO: 44.80 28.09 31.87
563
+ 2025-12-03 16:56:59 INFO: step 3000: train_loss = 2.402534, dev_score = 0.4480
564
+ 2025-12-03 16:57:00 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
565
+ 2025-12-03 16:57:00 INFO: new best model saved.
566
+ 2025-12-03 16:57:00 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
567
+ 2025-12-03 16:57:00 INFO: new model checkpoint saved.
568
+ 2025-12-03 16:57:01 INFO: Finished STEP 3020/50000, loss = 1.958558 (0.036 sec/batch), lr: 0.003000
569
+ 2025-12-03 16:57:02 INFO: Finished STEP 3040/50000, loss = 2.543922 (0.037 sec/batch), lr: 0.003000
570
+ 2025-12-03 16:57:03 INFO: Finished STEP 3060/50000, loss = 3.399988 (0.036 sec/batch), lr: 0.003000
571
+ 2025-12-03 16:57:03 INFO: Finished STEP 3080/50000, loss = 1.684248 (0.035 sec/batch), lr: 0.003000
572
+ 2025-12-03 16:57:04 INFO: Finished STEP 3100/50000, loss = 2.111065 (0.035 sec/batch), lr: 0.003000
573
+ 2025-12-03 16:57:04 INFO: Evaluating on dev set...
574
+ 2025-12-03 16:57:05 INFO: LAS MLAS BLEX
575
+ 2025-12-03 16:57:05 INFO: 38.37 25.21 30.67
576
+ 2025-12-03 16:57:05 INFO: step 3100: train_loss = 2.495950, dev_score = 0.3837
577
+ 2025-12-03 16:57:05 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
578
+ 2025-12-03 16:57:05 INFO: new model checkpoint saved.
579
+ 2025-12-03 16:57:06 INFO: Finished STEP 3120/50000, loss = 2.607497 (0.037 sec/batch), lr: 0.003000
580
+ 2025-12-03 16:57:07 INFO: Finished STEP 3140/50000, loss = 2.633862 (0.036 sec/batch), lr: 0.003000
581
+ 2025-12-03 16:57:08 INFO: Finished STEP 3160/50000, loss = 2.089188 (0.036 sec/batch), lr: 0.003000
582
+ 2025-12-03 16:57:08 INFO: Finished STEP 3180/50000, loss = 1.868492 (0.035 sec/batch), lr: 0.003000
583
+ 2025-12-03 16:57:09 INFO: Finished STEP 3200/50000, loss = 1.850400 (0.037 sec/batch), lr: 0.003000
584
+ 2025-12-03 16:57:09 INFO: Evaluating on dev set...
585
+ 2025-12-03 16:57:10 INFO: LAS MLAS BLEX
586
+ 2025-12-03 16:57:10 INFO: 36.88 24.05 29.96
587
+ 2025-12-03 16:57:10 INFO: step 3200: train_loss = 2.353660, dev_score = 0.3688
588
+ 2025-12-03 16:57:10 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
589
+ 2025-12-03 16:57:10 INFO: new model checkpoint saved.
590
+ 2025-12-03 16:57:11 INFO: Finished STEP 3220/50000, loss = 2.261243 (0.035 sec/batch), lr: 0.003000
591
+ 2025-12-03 16:57:12 INFO: Finished STEP 3240/50000, loss = 1.947779 (0.036 sec/batch), lr: 0.003000
592
+ 2025-12-03 16:57:12 INFO: Finished STEP 3260/50000, loss = 1.686020 (0.035 sec/batch), lr: 0.003000
593
+ 2025-12-03 16:57:13 INFO: Finished STEP 3280/50000, loss = 1.550403 (0.036 sec/batch), lr: 0.003000
594
+ 2025-12-03 16:57:14 INFO: Finished STEP 3300/50000, loss = 1.477478 (0.039 sec/batch), lr: 0.003000
595
+ 2025-12-03 16:57:14 INFO: Evaluating on dev set...
596
+ 2025-12-03 16:57:14 INFO: LAS MLAS BLEX
597
+ 2025-12-03 16:57:14 INFO: 31.19 22.13 25.89
598
+ 2025-12-03 16:57:14 INFO: step 3300: train_loss = 2.430038, dev_score = 0.3119
599
+ 2025-12-03 16:57:15 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
600
+ 2025-12-03 16:57:15 INFO: new model checkpoint saved.
601
+ 2025-12-03 16:57:16 INFO: Finished STEP 3320/50000, loss = 2.528574 (0.035 sec/batch), lr: 0.003000
602
+ 2025-12-03 16:57:17 INFO: Finished STEP 3340/50000, loss = 1.504506 (0.037 sec/batch), lr: 0.003000
603
+ 2025-12-03 16:57:17 INFO: Finished STEP 3360/50000, loss = 2.166523 (0.035 sec/batch), lr: 0.003000
604
+ 2025-12-03 16:57:18 INFO: Finished STEP 3380/50000, loss = 2.198786 (0.036 sec/batch), lr: 0.003000
605
+ 2025-12-03 16:57:19 INFO: Finished STEP 3400/50000, loss = 2.534374 (0.036 sec/batch), lr: 0.003000
606
+ 2025-12-03 16:57:19 INFO: Evaluating on dev set...
607
+ 2025-12-03 16:57:20 INFO: LAS MLAS BLEX
608
+ 2025-12-03 16:57:20 INFO: 44.80 29.24 33.47
609
+ 2025-12-03 16:57:20 INFO: step 3400: train_loss = 2.590817, dev_score = 0.4480
610
+ 2025-12-03 16:57:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
611
+ 2025-12-03 16:57:20 INFO: new best model saved.
612
+ 2025-12-03 16:57:21 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
613
+ 2025-12-03 16:57:21 INFO: new model checkpoint saved.
614
+ 2025-12-03 16:57:21 INFO: Finished STEP 3420/50000, loss = 3.138576 (0.036 sec/batch), lr: 0.003000
615
+ 2025-12-03 16:57:22 INFO: Finished STEP 3440/50000, loss = 2.721511 (0.036 sec/batch), lr: 0.003000
616
+ 2025-12-03 16:57:23 INFO: Finished STEP 3460/50000, loss = 2.431372 (0.036 sec/batch), lr: 0.003000
617
+ 2025-12-03 16:57:24 INFO: Finished STEP 3480/50000, loss = 3.378656 (0.037 sec/batch), lr: 0.003000
618
+ 2025-12-03 16:57:24 INFO: Finished STEP 3500/50000, loss = 3.028686 (0.039 sec/batch), lr: 0.003000
619
+ 2025-12-03 16:57:24 INFO: Evaluating on dev set...
620
+ 2025-12-03 16:57:25 INFO: LAS MLAS BLEX
621
+ 2025-12-03 16:57:25 INFO: 39.36 32.49 37.97
622
+ 2025-12-03 16:57:25 INFO: step 3500: train_loss = 2.838540, dev_score = 0.3936
623
+ 2025-12-03 16:57:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
624
+ 2025-12-03 16:57:26 INFO: new model checkpoint saved.
625
+ 2025-12-03 16:57:26 INFO: Finished STEP 3520/50000, loss = 2.257903 (0.039 sec/batch), lr: 0.003000
626
+ 2025-12-03 16:57:27 INFO: Finished STEP 3540/50000, loss = 2.388372 (0.036 sec/batch), lr: 0.003000
627
+ 2025-12-03 16:57:28 INFO: Finished STEP 3560/50000, loss = 2.756948 (0.036 sec/batch), lr: 0.003000
628
+ 2025-12-03 16:57:29 INFO: Finished STEP 3580/50000, loss = 3.632342 (0.039 sec/batch), lr: 0.003000
629
+ 2025-12-03 16:57:29 INFO: Finished STEP 3600/50000, loss = 2.432103 (0.035 sec/batch), lr: 0.003000
630
+ 2025-12-03 16:57:29 INFO: Evaluating on dev set...
631
+ 2025-12-03 16:57:30 INFO: LAS MLAS BLEX
632
+ 2025-12-03 16:57:30 INFO: 45.05 31.16 34.53
633
+ 2025-12-03 16:57:30 INFO: step 3600: train_loss = 2.653434, dev_score = 0.4505
634
+ 2025-12-03 16:57:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
635
+ 2025-12-03 16:57:30 INFO: new best model saved.
636
+ 2025-12-03 16:57:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
637
+ 2025-12-03 16:57:31 INFO: new model checkpoint saved.
638
+ 2025-12-03 16:57:32 INFO: Finished STEP 3620/50000, loss = 2.430850 (0.038 sec/batch), lr: 0.003000
639
+ 2025-12-03 16:57:32 INFO: Finished STEP 3640/50000, loss = 2.130368 (0.037 sec/batch), lr: 0.003000
640
+ 2025-12-03 16:57:33 INFO: Finished STEP 3660/50000, loss = 2.547717 (0.036 sec/batch), lr: 0.003000
641
+ 2025-12-03 16:57:34 INFO: Finished STEP 3680/50000, loss = 3.377545 (0.038 sec/batch), lr: 0.003000
642
+ 2025-12-03 16:57:35 INFO: Finished STEP 3700/50000, loss = 2.476320 (0.038 sec/batch), lr: 0.003000
643
+ 2025-12-03 16:57:35 INFO: Evaluating on dev set...
644
+ 2025-12-03 16:57:35 INFO: LAS MLAS BLEX
645
+ 2025-12-03 16:57:35 INFO: 48.76 33.96 38.57
646
+ 2025-12-03 16:57:35 INFO: step 3700: train_loss = 2.687569, dev_score = 0.4876
647
+ 2025-12-03 16:57:36 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
648
+ 2025-12-03 16:57:36 INFO: new best model saved.
649
+ 2025-12-03 16:57:36 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
650
+ 2025-12-03 16:57:36 INFO: new model checkpoint saved.
651
+ 2025-12-03 16:57:37 INFO: Finished STEP 3720/50000, loss = 2.905651 (0.037 sec/batch), lr: 0.003000
652
+ 2025-12-03 16:57:38 INFO: Finished STEP 3740/50000, loss = 3.174312 (0.038 sec/batch), lr: 0.003000
653
+ 2025-12-03 16:57:39 INFO: Finished STEP 3760/50000, loss = 2.021272 (0.037 sec/batch), lr: 0.003000
654
+ 2025-12-03 16:57:39 INFO: Finished STEP 3780/50000, loss = 2.649514 (0.036 sec/batch), lr: 0.003000
655
+ 2025-12-03 16:57:40 INFO: Finished STEP 3800/50000, loss = 2.542420 (0.035 sec/batch), lr: 0.003000
656
+ 2025-12-03 16:57:40 INFO: Evaluating on dev set...
657
+ 2025-12-03 16:57:41 INFO: LAS MLAS BLEX
658
+ 2025-12-03 16:57:41 INFO: 48.02 31.73 35.07
659
+ 2025-12-03 16:57:41 INFO: step 3800: train_loss = 2.677523, dev_score = 0.4802
660
+ 2025-12-03 16:57:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
661
+ 2025-12-03 16:57:41 INFO: new model checkpoint saved.
662
+ 2025-12-03 16:57:42 INFO: Finished STEP 3820/50000, loss = 3.016644 (0.037 sec/batch), lr: 0.003000
663
+ 2025-12-03 16:57:43 INFO: Finished STEP 3840/50000, loss = 2.505940 (0.035 sec/batch), lr: 0.003000
664
+ 2025-12-03 16:57:44 INFO: Finished STEP 3860/50000, loss = 2.399217 (0.038 sec/batch), lr: 0.003000
665
+ 2025-12-03 16:57:44 INFO: Finished STEP 3880/50000, loss = 2.742128 (0.037 sec/batch), lr: 0.003000
666
+ 2025-12-03 16:57:45 INFO: Finished STEP 3900/50000, loss = 1.561072 (0.038 sec/batch), lr: 0.003000
667
+ 2025-12-03 16:57:45 INFO: Evaluating on dev set...
668
+ 2025-12-03 16:57:46 INFO: LAS MLAS BLEX
669
+ 2025-12-03 16:57:46 INFO: 47.03 32.14 35.94
670
+ 2025-12-03 16:57:46 INFO: step 3900: train_loss = 2.663757, dev_score = 0.4703
671
+ 2025-12-03 16:57:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
672
+ 2025-12-03 16:57:46 INFO: new model checkpoint saved.
673
+ 2025-12-03 16:57:47 INFO: Finished STEP 3920/50000, loss = 3.016428 (0.039 sec/batch), lr: 0.003000
674
+ 2025-12-03 16:57:48 INFO: Finished STEP 3940/50000, loss = 2.364043 (0.038 sec/batch), lr: 0.003000
675
+ 2025-12-03 16:57:49 INFO: Finished STEP 3960/50000, loss = 1.340432 (0.037 sec/batch), lr: 0.003000
676
+ 2025-12-03 16:57:49 INFO: Finished STEP 3980/50000, loss = 2.343470 (0.037 sec/batch), lr: 0.003000
677
+ 2025-12-03 16:57:50 INFO: Finished STEP 4000/50000, loss = 3.607741 (0.038 sec/batch), lr: 0.003000
678
+ 2025-12-03 16:57:50 INFO: Evaluating on dev set...
679
+ 2025-12-03 16:57:51 INFO: LAS MLAS BLEX
680
+ 2025-12-03 16:57:51 INFO: 48.76 33.54 37.32
681
+ 2025-12-03 16:57:51 INFO: step 4000: train_loss = 2.675620, dev_score = 0.4876
682
+ 2025-12-03 16:57:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
683
+ 2025-12-03 16:57:51 INFO: new best model saved.
684
+ 2025-12-03 16:57:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
685
+ 2025-12-03 16:57:52 INFO: new model checkpoint saved.
686
+ 2025-12-03 16:57:52 INFO: Finished STEP 4020/50000, loss = 2.606227 (0.039 sec/batch), lr: 0.003000
687
+ 2025-12-03 16:57:53 INFO: Finished STEP 4040/50000, loss = 2.095632 (0.037 sec/batch), lr: 0.003000
688
+ 2025-12-03 16:57:54 INFO: Finished STEP 4060/50000, loss = 2.065331 (0.038 sec/batch), lr: 0.003000
689
+ 2025-12-03 16:57:55 INFO: Finished STEP 4080/50000, loss = 1.945910 (0.039 sec/batch), lr: 0.003000
690
+ 2025-12-03 16:57:56 INFO: Finished STEP 4100/50000, loss = 2.439826 (0.036 sec/batch), lr: 0.003000
691
+ 2025-12-03 16:57:56 INFO: Evaluating on dev set...
692
+ 2025-12-03 16:57:56 INFO: LAS MLAS BLEX
693
+ 2025-12-03 16:57:56 INFO: 43.56 28.87 33.47
694
+ 2025-12-03 16:57:56 INFO: step 4100: train_loss = 2.684596, dev_score = 0.4356
695
+ 2025-12-03 16:57:57 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
696
+ 2025-12-03 16:57:57 INFO: new model checkpoint saved.
697
+ 2025-12-03 16:57:58 INFO: Finished STEP 4120/50000, loss = 3.153655 (0.037 sec/batch), lr: 0.003000
698
+ 2025-12-03 16:57:58 INFO: Finished STEP 4140/50000, loss = 2.724533 (0.037 sec/batch), lr: 0.003000
699
+ 2025-12-03 16:57:59 INFO: Finished STEP 4160/50000, loss = 3.230067 (0.040 sec/batch), lr: 0.003000
700
+ 2025-12-03 16:58:00 INFO: Finished STEP 4180/50000, loss = 4.007288 (0.036 sec/batch), lr: 0.003000
701
+ 2025-12-03 16:58:01 INFO: Finished STEP 4200/50000, loss = 2.615037 (0.036 sec/batch), lr: 0.003000
702
+ 2025-12-03 16:58:01 INFO: Evaluating on dev set...
703
+ 2025-12-03 16:58:01 INFO: LAS MLAS BLEX
704
+ 2025-12-03 16:58:01 INFO: 48.51 33.97 37.37
705
+ 2025-12-03 16:58:01 INFO: step 4200: train_loss = 2.689222, dev_score = 0.4851
706
+ 2025-12-03 16:58:02 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
707
+ 2025-12-03 16:58:02 INFO: new model checkpoint saved.
708
+ 2025-12-03 16:58:03 INFO: Finished STEP 4220/50000, loss = 2.533058 (0.037 sec/batch), lr: 0.003000
709
+ 2025-12-03 16:58:03 INFO: Finished STEP 4240/50000, loss = 2.750103 (0.038 sec/batch), lr: 0.003000
710
+ 2025-12-03 16:58:04 INFO: Finished STEP 4260/50000, loss = 2.774418 (0.039 sec/batch), lr: 0.003000
711
+ 2025-12-03 16:58:05 INFO: Finished STEP 4280/50000, loss = 2.897849 (0.038 sec/batch), lr: 0.003000
712
+ 2025-12-03 16:58:06 INFO: Finished STEP 4300/50000, loss = 2.617096 (0.041 sec/batch), lr: 0.003000
713
+ 2025-12-03 16:58:06 INFO: Evaluating on dev set...
714
+ 2025-12-03 16:58:06 INFO: LAS MLAS BLEX
715
+ 2025-12-03 16:58:06 INFO: 43.81 28.57 32.35
716
+ 2025-12-03 16:58:06 INFO: step 4300: train_loss = 2.673825, dev_score = 0.4381
717
+ 2025-12-03 16:58:07 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
718
+ 2025-12-03 16:58:07 INFO: new model checkpoint saved.
719
+ 2025-12-03 16:58:08 INFO: Finished STEP 4320/50000, loss = 2.601406 (0.035 sec/batch), lr: 0.003000
720
+ 2025-12-03 16:58:09 INFO: Finished STEP 4340/50000, loss = 1.767829 (0.036 sec/batch), lr: 0.003000
721
+ 2025-12-03 16:58:09 INFO: Finished STEP 4360/50000, loss = 4.286842 (0.037 sec/batch), lr: 0.003000
722
+ 2025-12-03 16:58:10 INFO: Finished STEP 4380/50000, loss = 2.548105 (0.037 sec/batch), lr: 0.003000
723
+ 2025-12-03 16:58:11 INFO: Finished STEP 4400/50000, loss = 2.197368 (0.037 sec/batch), lr: 0.003000
724
+ 2025-12-03 16:58:11 INFO: Evaluating on dev set...
725
+ 2025-12-03 16:58:11 INFO: LAS MLAS BLEX
726
+ 2025-12-03 16:58:11 INFO: 42.57 27.37 34.11
727
+ 2025-12-03 16:58:11 INFO: step 4400: train_loss = 2.806391, dev_score = 0.4257
728
+ 2025-12-03 16:58:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
729
+ 2025-12-03 16:58:12 INFO: new model checkpoint saved.
730
+ 2025-12-03 16:58:13 INFO: Finished STEP 4420/50000, loss = 2.152693 (0.039 sec/batch), lr: 0.003000
731
+ 2025-12-03 16:58:14 INFO: Finished STEP 4440/50000, loss = 2.775348 (0.040 sec/batch), lr: 0.003000
732
+ 2025-12-03 16:58:14 INFO: Finished STEP 4460/50000, loss = 2.762387 (0.036 sec/batch), lr: 0.003000
733
+ 2025-12-03 16:58:15 INFO: Finished STEP 4480/50000, loss = 2.136549 (0.037 sec/batch), lr: 0.003000
734
+ 2025-12-03 16:58:16 INFO: Finished STEP 4500/50000, loss = 2.218260 (0.037 sec/batch), lr: 0.003000
735
+ 2025-12-03 16:58:16 INFO: Evaluating on dev set...
736
+ 2025-12-03 16:58:16 INFO: LAS MLAS BLEX
737
+ 2025-12-03 16:58:16 INFO: 39.60 26.16 32.07
738
+ 2025-12-03 16:58:16 INFO: step 4500: train_loss = 2.590793, dev_score = 0.3960
739
+ 2025-12-03 16:58:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
740
+ 2025-12-03 16:58:17 INFO: new model checkpoint saved.
741
+ 2025-12-03 16:58:18 INFO: Finished STEP 4520/50000, loss = 2.940754 (0.036 sec/batch), lr: 0.003000
742
+ 2025-12-03 16:58:19 INFO: Finished STEP 4540/50000, loss = 3.413527 (0.037 sec/batch), lr: 0.003000
743
+ 2025-12-03 16:58:19 INFO: Finished STEP 4560/50000, loss = 2.152988 (0.036 sec/batch), lr: 0.003000
744
+ 2025-12-03 16:58:20 INFO: Finished STEP 4580/50000, loss = 3.002730 (0.040 sec/batch), lr: 0.003000
745
+ 2025-12-03 16:58:21 INFO: Finished STEP 4600/50000, loss = 2.575607 (0.037 sec/batch), lr: 0.003000
746
+ 2025-12-03 16:58:21 INFO: Evaluating on dev set...
747
+ 2025-12-03 16:58:21 INFO: LAS MLAS BLEX
748
+ 2025-12-03 16:58:21 INFO: 46.04 32.63 36.44
749
+ 2025-12-03 16:58:21 INFO: step 4600: train_loss = 2.624743, dev_score = 0.4604
750
+ 2025-12-03 16:58:22 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
751
+ 2025-12-03 16:58:22 INFO: new model checkpoint saved.
752
+ 2025-12-03 16:58:23 INFO: Finished STEP 4620/50000, loss = 1.642988 (0.036 sec/batch), lr: 0.003000
753
+ 2025-12-03 16:58:24 INFO: Finished STEP 4640/50000, loss = 1.793095 (0.037 sec/batch), lr: 0.003000
754
+ 2025-12-03 16:58:24 INFO: Finished STEP 4660/50000, loss = 2.214316 (0.037 sec/batch), lr: 0.003000
755
+ 2025-12-03 16:58:25 INFO: Finished STEP 4680/50000, loss = 4.035345 (0.037 sec/batch), lr: 0.003000
756
+ 2025-12-03 16:58:26 INFO: Finished STEP 4700/50000, loss = 2.505473 (0.037 sec/batch), lr: 0.003000
757
+ 2025-12-03 16:58:26 INFO: Evaluating on dev set...
758
+ 2025-12-03 16:58:26 INFO: LAS MLAS BLEX
759
+ 2025-12-03 16:58:26 INFO: 45.54 30.38 35.02
760
+ 2025-12-03 16:58:26 INFO: step 4700: train_loss = 2.544833, dev_score = 0.4554
761
+ 2025-12-03 16:58:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
762
+ 2025-12-03 16:58:27 INFO: new model checkpoint saved.
763
+ 2025-12-03 16:58:28 INFO: Finished STEP 4720/50000, loss = 2.611087 (0.036 sec/batch), lr: 0.003000
764
+ 2025-12-03 16:58:29 INFO: Finished STEP 4740/50000, loss = 2.382816 (0.038 sec/batch), lr: 0.003000
765
+ 2025-12-03 16:58:29 INFO: Finished STEP 4760/50000, loss = 4.116714 (0.038 sec/batch), lr: 0.003000
766
+ 2025-12-03 16:58:30 INFO: Finished STEP 4780/50000, loss = 3.737564 (0.037 sec/batch), lr: 0.003000
767
+ 2025-12-03 16:58:31 INFO: Finished STEP 4800/50000, loss = 3.283456 (0.040 sec/batch), lr: 0.003000
768
+ 2025-12-03 16:58:31 INFO: Evaluating on dev set...
769
+ 2025-12-03 16:58:32 INFO: LAS MLAS BLEX
770
+ 2025-12-03 16:58:32 INFO: 47.03 31.32 36.33
771
+ 2025-12-03 16:58:32 INFO: step 4800: train_loss = 2.612103, dev_score = 0.4703
772
+ 2025-12-03 16:58:32 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
773
+ 2025-12-03 16:58:32 INFO: new model checkpoint saved.
774
+ 2025-12-03 16:58:33 INFO: Finished STEP 4820/50000, loss = 2.108931 (0.038 sec/batch), lr: 0.003000
775
+ 2025-12-03 16:58:34 INFO: Finished STEP 4840/50000, loss = 3.035322 (0.037 sec/batch), lr: 0.003000
776
+ 2025-12-03 16:58:35 INFO: Finished STEP 4860/50000, loss = 2.147128 (0.037 sec/batch), lr: 0.003000
777
+ 2025-12-03 16:58:35 INFO: Finished STEP 4880/50000, loss = 2.547672 (0.040 sec/batch), lr: 0.003000
778
+ 2025-12-03 16:58:36 INFO: Finished STEP 4900/50000, loss = 2.383922 (0.039 sec/batch), lr: 0.003000
779
+ 2025-12-03 16:58:36 INFO: Evaluating on dev set...
780
+ 2025-12-03 16:58:37 INFO: LAS MLAS BLEX
781
+ 2025-12-03 16:58:37 INFO: 43.56 28.57 33.19
782
+ 2025-12-03 16:58:37 INFO: step 4900: train_loss = 2.794512, dev_score = 0.4356
783
+ 2025-12-03 16:58:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
784
+ 2025-12-03 16:58:37 INFO: new model checkpoint saved.
785
+ 2025-12-03 16:58:38 INFO: Finished STEP 4920/50000, loss = 2.903427 (0.036 sec/batch), lr: 0.003000
786
+ 2025-12-03 16:58:39 INFO: Finished STEP 4940/50000, loss = 2.407435 (0.037 sec/batch), lr: 0.003000
787
+ 2025-12-03 16:58:40 INFO: Finished STEP 4960/50000, loss = 2.236732 (0.039 sec/batch), lr: 0.003000
788
+ 2025-12-03 16:58:40 INFO: Finished STEP 4980/50000, loss = 3.158165 (0.039 sec/batch), lr: 0.003000
789
+ 2025-12-03 16:58:41 INFO: Finished STEP 5000/50000, loss = 1.865582 (0.037 sec/batch), lr: 0.003000
790
+ 2025-12-03 16:58:41 INFO: Evaluating on dev set...
791
+ 2025-12-03 16:58:42 INFO: LAS MLAS BLEX
792
+ 2025-12-03 16:58:42 INFO: 46.29 30.02 33.40
793
+ 2025-12-03 16:58:42 INFO: step 5000: train_loss = 3.121457, dev_score = 0.4629
794
+ 2025-12-03 16:58:42 INFO: Training ended with 5000 steps.
795
+ 2025-12-03 16:58:42 INFO: Best dev F1 = 48.76, at iteration = 3700
796
+ 2025-12-03 16:58:44 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
797
+ 2025-12-03 16:58:44 INFO: Running parser in predict mode
798
+ 2025-12-03 16:58:44 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
799
+ 2025-12-03 16:58:46 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
800
+ 2025-12-03 16:58:46 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
801
+ 2025-12-03 16:58:46 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
802
+ 2025-12-03 16:58:46 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
803
+ 2025-12-03 16:58:46 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
804
+ 2025-12-03 16:58:46 INFO: Loading data with batch size 32...
805
+ 2025-12-03 16:58:46 DEBUG: 9 batches created.
806
+ 2025-12-03 16:58:47 INFO: F1 scores for each dependency:
807
+ Note that unlabeled attachment errors hurt the labeled attachment scores
808
+ acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
809
+ acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
810
+ advcl: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
811
+ advmod: p 0.4848 r 0.6400 f1 0.5517 (25 actual)
812
+ amod: p 0.7586 r 0.7097 f1 0.7333 (31 actual)
813
+ appos: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
814
+ aux: p 0.7500 r 0.8182 f1 0.7826 (11 actual)
815
+ case: p 0.8036 r 0.8036 f1 0.8036 (56 actual)
816
+ cc: p 0.5714 r 0.6154 f1 0.5926 (13 actual)
817
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
818
+ conj: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
819
+ cop: p 1.0000 r 0.3333 f1 0.5000 (3 actual)
820
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
821
+ det: p 0.8696 r 0.9091 f1 0.8889 (22 actual)
822
+ expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
823
+ iobj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
824
+ mark: p 0.5000 r 0.1667 f1 0.2500 (12 actual)
825
+ nmod: p 0.0000 r 0.0000 f1 0.0000 (15 actual)
826
+ nmod:poss: p 0.8947 r 0.8947 f1 0.8947 (19 actual)
827
+ nsubj: p 0.4231 r 0.6471 f1 0.5116 (17 actual)
828
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
829
+ obj: p 0.0000 r 0.0000 f1 0.0000 (22 actual)
830
+ obl: p 0.2143 r 0.4390 f1 0.2880 (41 actual)
831
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
832
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
833
+ parataxis: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
834
+ punct: p 0.4423 r 0.4423 f1 0.4423 (52 actual)
835
+ root: p 0.4444 r 0.4444 f1 0.4444 (9 actual)
836
+ xcomp: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
837
+ 2025-12-03 16:58:47 INFO: LAS MLAS BLEX
838
+ 2025-12-03 16:58:47 INFO: 48.76 33.54 37.32
839
+ 2025-12-03 16:58:47 INFO: Parser score:
840
+ 2025-12-03 16:58:47 INFO: sv_diachronic 48.76
841
+ 2025-12-03 16:58:47 INFO: Finished running dev set on
842
+ UD_Swedish-diachronic
843
+ UAS LAS CLAS MLAS BLEX
844
+ 57.18 48.76 37.32 33.54 37.32
845
+ 2025-12-03 16:58:47 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
846
+ 2025-12-03 16:58:47 INFO: Running parser in predict mode
847
+ 2025-12-03 16:58:47 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
848
+ 2025-12-03 16:58:49 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
849
+ 2025-12-03 16:58:49 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
850
+ 2025-12-03 16:58:49 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
851
+ 2025-12-03 16:58:49 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
852
+ 2025-12-03 16:58:49 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
853
+ 2025-12-03 16:58:49 INFO: Loading data with batch size 32...
854
+ 2025-12-03 16:58:49 DEBUG: 93 batches created.
855
+ 2025-12-03 16:58:54 INFO: F1 scores for each dependency:
856
+ Note that unlabeled attachment errors hurt the labeled attachment scores
857
+ acl: p 0.0000 r 0.0000 f1 0.0000 (32 actual)
858
+ acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
859
+ acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (75 actual)
860
+ advcl: p 0.0000 r 0.0000 f1 0.0000 (60 actual)
861
+ advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
862
+ advmod: p 0.4360 r 0.5597 f1 0.4902 (268 actual)
863
+ amod: p 0.7222 r 0.7913 f1 0.7552 (230 actual)
864
+ appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
865
+ aux: p 0.7053 r 0.7976 f1 0.7486 (84 actual)
866
+ aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
867
+ case: p 0.7571 r 0.7185 f1 0.7373 (373 actual)
868
+ cc: p 0.5155 r 0.5355 f1 0.5253 (155 actual)
869
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
870
+ compound:prt: p 0.0000 r 0.0000 f1 0.0000 (21 actual)
871
+ conj: p 0.0476 r 0.0127 f1 0.0200 (158 actual)
872
+ cop: p 0.6500 r 0.2826 f1 0.3939 (46 actual)
873
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
874
+ dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
875
+ det: p 0.7970 r 0.7740 f1 0.7854 (208 actual)
876
+ discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
877
+ dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
878
+ expl: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
879
+ expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
880
+ fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
881
+ flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
882
+ flat:name: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
883
+ goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
884
+ iobj: p 0.0000 r 0.0000 f1 0.0000 (14 actual)
885
+ mark: p 0.6437 r 0.3660 f1 0.4667 (153 actual)
886
+ nmod: p 0.0909 r 0.0588 f1 0.0714 (102 actual)
887
+ nmod:poss: p 0.8026 r 0.8592 f1 0.8299 (142 actual)
888
+ nsubj: p 0.4277 r 0.4857 f1 0.4548 (280 actual)
889
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
890
+ nummod: p 0.0000 r 0.0000 f1 0.0000 (10 actual)
891
+ obj: p 0.1739 r 0.0219 f1 0.0388 (183 actual)
892
+ obl: p 0.1360 r 0.3561 f1 0.1968 (278 actual)
893
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
894
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
895
+ parataxis: p 0.0000 r 0.0000 f1 0.0000 (18 actual)
896
+ punct: p 0.3560 r 0.3576 f1 0.3568 (425 actual)
897
+ reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
898
+ root: p 0.4040 r 0.4040 f1 0.4040 (99 actual)
899
+ vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
900
+ xcomp: p 0.0198 r 0.0667 f1 0.0306 (75 actual)
901
+ 2025-12-03 16:58:54 INFO: LAS MLAS BLEX
902
+ 2025-12-03 16:58:54 INFO: 42.84 30.44 33.75
903
+ 2025-12-03 16:58:54 INFO: Parser score:
904
+ 2025-12-03 16:58:54 INFO: sv_diachronic 42.84
905
+ 2025-12-03 16:58:54 INFO: Finished running test set on
906
+ UD_Swedish-diachronic
907
+ UAS LAS CLAS MLAS BLEX
908
+ 53.77 42.84 33.75 30.44 33.75
909
+ DONE.
910
+ Full log saved to: logs/log_conll17.pt_sv_diachron_is_20251203_164326.txt
911
+ Symlink updated: logs/latest.txt → log_conll17.pt_sv_diachron_is_20251203_164326.txt
logs/log_conll17.pt_sv_diachron_is_20251203_221228.txt ADDED
@@ -0,0 +1,147 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ === LOGFILE: logs/log_conll17.pt_sv_diachron_is_20251203_221228.txt ===
2
+ Language codes: sv diachron is
3
+ Using pretrained model: conll17.pt
4
+
5
+ Running: python prepare-train-val-test.py sv diachron is
6
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
7
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
8
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
9
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
10
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
11
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
12
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
13
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
14
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
15
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
16
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
17
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_gc-ud-dev.conllu
18
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_modern-ud-train.conllu
19
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_pud-ud-test.conllu
20
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_modern-ud-test.conllu
21
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_gc-ud-train.conllu
22
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_gc-ud-test.conllu
23
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_modern-ud-dev.conllu
24
+ Including DigPhil MACHINE in TRAIN (minus gold)…
25
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
26
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
27
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
28
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
29
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
30
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
31
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
32
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
33
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
34
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
35
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
36
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
37
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
38
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
39
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
40
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
41
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
42
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
43
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
44
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
45
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
46
+ Cleaning TRAIN...
47
+ [REMOVED] sent_id=6 ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
48
+ [REMOVED] sent_id=7_8 ERRORS=['Multiple roots found: [5, 10]']
49
+ [REMOVED] sent_id=30_31 ERRORS=['Multiple roots found: [3, 18]']
50
+ [REMOVED] sent_id=35 ERRORS=['Line 36: Invalid token ID or head']
51
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [1, 5]']
52
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [7, 20]']
53
+ [REMOVED] sent_id=8_9 ERRORS=['Multiple roots found: [24, 57]']
54
+ [REMOVED] sent_id=12_13 ERRORS=['Multiple roots found: [11, 16]']
55
+ [REMOVED] sent_id=124_split2 ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
56
+ [REMOVED] sent_id=396 ERRORS=['Token 2: Missing form']
57
+ [REMOVED] sent_id=416 ERRORS=['Token 2: Missing form']
58
+ [REMOVED] sent_id=589 ERRORS=['Token 2: Missing form']
59
+ [REMOVED] sent_id=909 ERRORS=['Token 2: Missing form']
60
+ [REMOVED] sent_id=912 ERRORS=['Token 2: Missing form']
61
+ [REMOVED] sent_id=3_split1 ERRORS=['Multiple roots found: [4, 15, 17]']
62
+ [REMOVED] sent_id=3_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
63
+ [REMOVED] sent_id=3_4 ERRORS=['Multiple roots found: [1, 5]']
64
+ [REMOVED] sent_id=5_6 ERRORS=['Multiple roots found: [3, 24]']
65
+ [REMOVED] sent_id=11_12_13 ERRORS=['Multiple roots found: [5, 17, 25]']
66
+ [REMOVED] sent_id=119 ERRORS=['Token 2: Missing form']
67
+ [REMOVED] sent_id=179 ERRORS=['Token 2: Missing form']
68
+ [REMOVED] sent_id=188 ERRORS=['Token 2: Missing form']
69
+ [REMOVED] sent_id=223 ERRORS=['Token 2: Missing form']
70
+ [REMOVED] sent_id=268 ERRORS=['Token 2: Missing form']
71
+ [REMOVED] sent_id=325 ERRORS=['Token 2: Missing form']
72
+ [REMOVED] sent_id=388 ERRORS=['Token 2: Missing form']
73
+ [REMOVED] sent_id=399 ERRORS=['Token 2: Missing form']
74
+ [REMOVED] sent_id=475 ERRORS=['Token 2: Missing form']
75
+ [REMOVED] sent_id=505 ERRORS=['Token 2: Missing form']
76
+ [REMOVED] sent_id=520 ERRORS=['Token 2: Missing form']
77
+ [REMOVED] sent_id=562 ERRORS=['Token 2: Missing form']
78
+ [REMOVED] sent_id=669 ERRORS=['Token 2: Missing form']
79
+ [REMOVED] sent_id=711 ERRORS=['Token 2: Missing form']
80
+ [REMOVED] sent_id=731 ERRORS=['Token 2: Missing form']
81
+ [REMOVED] sent_id=867 ERRORS=['Token 2: Missing form']
82
+ [REMOVED] sent_id=884 ERRORS=['Token 2: Missing form']
83
+ [REMOVED] sent_id=923 ERRORS=['Token 2: Missing form']
84
+ [REMOVED] sent_id=939 ERRORS=['Token 2: Missing form']
85
+ [REMOVED] sent_id=1086 ERRORS=['Token 2: Missing form']
86
+ [REMOVED] sent_id=1179 ERRORS=['Token 2: Missing form']
87
+ [REMOVED] sent_id=1251 ERRORS=['Token 2: Missing form']
88
+ [REMOVED] sent_id=1345 ERRORS=['Token 2: Missing form']
89
+ [REMOVED] sent_id=1459 ERRORS=['Token 2: Missing form']
90
+ [REMOVED] sent_id=1656 ERRORS=['Token 2: Missing form']
91
+ [REMOVED] sent_id=1669 ERRORS=['Token 2: Missing form']
92
+ [REMOVED] sent_id=87_88 ERRORS=['Multiple roots found: [3, 6]']
93
+ [REMOVED] sent_id=65_split2_66_split2 ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
94
+ [REMOVED] sent_id=25 ERRORS=['Token 2: Missing form']
95
+ [REMOVED] sent_id=136 ERRORS=['Token 2: Missing form']
96
+ [REMOVED] sent_id=208 ERRORS=['Token 2: Missing form']
97
+ [REMOVED] sent_id=230 ERRORS=['Token 2: Missing form']
98
+ [REMOVED] sent_id=245 ERRORS=['Token 2: Missing form']
99
+ [REMOVED] sent_id=276 ERRORS=['Token 2: Missing form']
100
+ [REMOVED] sent_id=320 ERRORS=['Token 2: Missing form']
101
+ [REMOVED] sent_id=366 ERRORS=['Token 2: Missing form']
102
+ [REMOVED] sent_id=519 ERRORS=['Token 2: Missing form']
103
+ [REMOVED] sent_id=569 ERRORS=['Token 2: Missing form']
104
+ [REMOVED] sent_id=50_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
105
+ [REMOVED] sent_id=53_54 ERRORS=['Multiple roots found: [27, 91]']
106
+ [REMOVED] sent_id=55_56_57 ERRORS=['Multiple roots found: [2, 4, 13]']
107
+ [REMOVED] sent_id=17_split1 ERRORS=['Multiple roots found: [2, 14, 17]']
108
+ [REMOVED] sent_id=17_split2 ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
109
+ [REMOVED] sent_id=19_split1 ERRORS=['Multiple roots found: [3, 31]']
110
+ Cleaning DEV...
111
+ [REMOVED] sent_id=33 ERRORS=['Token 15: Missing deprel']
112
+ Cleaning TEST...
113
+ Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (75787 valid sentences)
114
+ Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
115
+ Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
116
+ Done.
117
+ Sourcing scripts/config_alvis.sh
118
+ Running stanza dataset preparation…
119
+ 2025-12-03 22:12:37 INFO: Datasets program called with:
120
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
121
+ 2025-12-03 22:12:37 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
122
+
123
+ 2025-12-03 22:12:37 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
124
+ 2025-12-03 22:12:37 DEBUG: Processing parameter "processors"...
125
+ 2025-12-03 22:12:37 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
126
+ 2025-12-03 22:12:37 INFO: Downloading these customized packages for language: sv (Swedish)...
127
+ =======================
128
+ | Processor | Package |
129
+ -----------------------
130
+ =======================
131
+
132
+ 2025-12-03 22:12:37 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
133
+ 2025-12-03 22:12:37 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
134
+ 2025-12-03 22:12:37 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
135
+ 2025-12-03 22:12:37 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
136
+ Augmented 191 quotes: Counter({'„”': 29, '»«': 25, '「」': 25, '″″': 20, '«»': 17, '„“': 16, '””': 16, '""': 15, '《》': 15, '““': 13})
137
+ 2025-12-03 22:12:41 INFO: Running tagger to retag /local/tmp.5441282/tmp2whacqzu/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
138
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmp2whacqzu/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
139
+ 2025-12-03 22:12:41 INFO: Running tagger in predict mode
140
+ 2025-12-03 22:12:41 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
141
+ 2025-12-03 22:12:43 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
142
+ 2025-12-03 22:12:43 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
143
+ 2025-12-03 22:12:43 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
144
+ 2025-12-03 22:12:43 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
145
+ 2025-12-03 22:12:44 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
146
+ 2025-12-03 22:12:46 INFO: Loading data with batch size 250...
147
+ 2025-12-03 22:13:33 INFO: Start evaluation...
logs/log_conll17.pt_sv_diachron_nn_20251203_160846.txt ADDED
@@ -0,0 +1,1146 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ === LOGFILE: logs/log_conll17.pt_sv_diachron_nn_20251203_160846.txt ===
2
+ Language codes: sv diachron nn
3
+ Using pretrained model: conll17.pt
4
+
5
+ Running: python prepare-train-val-test.py sv diachron nn
6
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
7
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
8
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
9
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
10
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
11
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
12
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
13
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
14
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
15
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
16
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
17
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-nn/no_nynorsk-ud-test.conllu
18
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-nn/no_nynorsk-ud-train.conllu
19
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-nn/no_nynorsk-ud-dev.conllu
20
+ Including DigPhil MACHINE in TRAIN (minus gold)…
21
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
22
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
23
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
24
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
25
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
26
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
27
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
28
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
29
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
30
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
31
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
32
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
33
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
34
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
35
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
36
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
37
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
38
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
39
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
40
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
41
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
42
+ Cleaning TRAIN...
43
+ [REMOVED] sent_id=6 ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
44
+ [REMOVED] sent_id=7_8 ERRORS=['Multiple roots found: [5, 10]']
45
+ [REMOVED] sent_id=30_31 ERRORS=['Multiple roots found: [3, 18]']
46
+ [REMOVED] sent_id=35 ERRORS=['Line 36: Invalid token ID or head']
47
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [1, 5]']
48
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [7, 20]']
49
+ [REMOVED] sent_id=8_9 ERRORS=['Multiple roots found: [24, 57]']
50
+ [REMOVED] sent_id=12_13 ERRORS=['Multiple roots found: [11, 16]']
51
+ [REMOVED] sent_id=124_split2 ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
52
+ [REMOVED] sent_id=396 ERRORS=['Token 2: Missing form']
53
+ [REMOVED] sent_id=416 ERRORS=['Token 2: Missing form']
54
+ [REMOVED] sent_id=589 ERRORS=['Token 2: Missing form']
55
+ [REMOVED] sent_id=909 ERRORS=['Token 2: Missing form']
56
+ [REMOVED] sent_id=912 ERRORS=['Token 2: Missing form']
57
+ [REMOVED] sent_id=3_split1 ERRORS=['Multiple roots found: [4, 15, 17]']
58
+ [REMOVED] sent_id=3_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
59
+ [REMOVED] sent_id=3_4 ERRORS=['Multiple roots found: [1, 5]']
60
+ [REMOVED] sent_id=5_6 ERRORS=['Multiple roots found: [3, 24]']
61
+ [REMOVED] sent_id=11_12_13 ERRORS=['Multiple roots found: [5, 17, 25]']
62
+ [REMOVED] sent_id=119 ERRORS=['Token 2: Missing form']
63
+ [REMOVED] sent_id=179 ERRORS=['Token 2: Missing form']
64
+ [REMOVED] sent_id=188 ERRORS=['Token 2: Missing form']
65
+ [REMOVED] sent_id=223 ERRORS=['Token 2: Missing form']
66
+ [REMOVED] sent_id=268 ERRORS=['Token 2: Missing form']
67
+ [REMOVED] sent_id=325 ERRORS=['Token 2: Missing form']
68
+ [REMOVED] sent_id=388 ERRORS=['Token 2: Missing form']
69
+ [REMOVED] sent_id=399 ERRORS=['Token 2: Missing form']
70
+ [REMOVED] sent_id=475 ERRORS=['Token 2: Missing form']
71
+ [REMOVED] sent_id=505 ERRORS=['Token 2: Missing form']
72
+ [REMOVED] sent_id=520 ERRORS=['Token 2: Missing form']
73
+ [REMOVED] sent_id=562 ERRORS=['Token 2: Missing form']
74
+ [REMOVED] sent_id=669 ERRORS=['Token 2: Missing form']
75
+ [REMOVED] sent_id=711 ERRORS=['Token 2: Missing form']
76
+ [REMOVED] sent_id=731 ERRORS=['Token 2: Missing form']
77
+ [REMOVED] sent_id=867 ERRORS=['Token 2: Missing form']
78
+ [REMOVED] sent_id=884 ERRORS=['Token 2: Missing form']
79
+ [REMOVED] sent_id=923 ERRORS=['Token 2: Missing form']
80
+ [REMOVED] sent_id=939 ERRORS=['Token 2: Missing form']
81
+ [REMOVED] sent_id=1086 ERRORS=['Token 2: Missing form']
82
+ [REMOVED] sent_id=1179 ERRORS=['Token 2: Missing form']
83
+ [REMOVED] sent_id=1251 ERRORS=['Token 2: Missing form']
84
+ [REMOVED] sent_id=1345 ERRORS=['Token 2: Missing form']
85
+ [REMOVED] sent_id=1459 ERRORS=['Token 2: Missing form']
86
+ [REMOVED] sent_id=1656 ERRORS=['Token 2: Missing form']
87
+ [REMOVED] sent_id=1669 ERRORS=['Token 2: Missing form']
88
+ [REMOVED] sent_id=87_88 ERRORS=['Multiple roots found: [3, 6]']
89
+ [REMOVED] sent_id=65_split2_66_split2 ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
90
+ [REMOVED] sent_id=25 ERRORS=['Token 2: Missing form']
91
+ [REMOVED] sent_id=136 ERRORS=['Token 2: Missing form']
92
+ [REMOVED] sent_id=208 ERRORS=['Token 2: Missing form']
93
+ [REMOVED] sent_id=230 ERRORS=['Token 2: Missing form']
94
+ [REMOVED] sent_id=245 ERRORS=['Token 2: Missing form']
95
+ [REMOVED] sent_id=276 ERRORS=['Token 2: Missing form']
96
+ [REMOVED] sent_id=320 ERRORS=['Token 2: Missing form']
97
+ [REMOVED] sent_id=366 ERRORS=['Token 2: Missing form']
98
+ [REMOVED] sent_id=519 ERRORS=['Token 2: Missing form']
99
+ [REMOVED] sent_id=569 ERRORS=['Token 2: Missing form']
100
+ [REMOVED] sent_id=50_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
101
+ [REMOVED] sent_id=53_54 ERRORS=['Multiple roots found: [27, 91]']
102
+ [REMOVED] sent_id=55_56_57 ERRORS=['Multiple roots found: [2, 4, 13]']
103
+ [REMOVED] sent_id=17_split1 ERRORS=['Multiple roots found: [2, 14, 17]']
104
+ [REMOVED] sent_id=17_split2 ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
105
+ [REMOVED] sent_id=19_split1 ERRORS=['Multiple roots found: [3, 31]']
106
+ Cleaning DEV...
107
+ [REMOVED] sent_id=33 ERRORS=['Token 15: Missing deprel']
108
+ Cleaning TEST...
109
+ Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (83827 valid sentences)
110
+ Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
111
+ Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
112
+ Done.
113
+ Sourcing scripts/config_alvis.sh
114
+ Running stanza dataset preparation…
115
+ 2025-12-03 16:08:56 INFO: Datasets program called with:
116
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
117
+ 2025-12-03 16:08:56 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
118
+
119
+ 2025-12-03 16:08:56 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
120
+ 2025-12-03 16:08:56 DEBUG: Processing parameter "processors"...
121
+ 2025-12-03 16:08:56 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
122
+ 2025-12-03 16:08:56 INFO: Downloading these customized packages for language: sv (Swedish)...
123
+ =======================
124
+ | Processor | Package |
125
+ -----------------------
126
+ =======================
127
+
128
+ 2025-12-03 16:08:56 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
129
+ 2025-12-03 16:08:56 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
130
+ 2025-12-03 16:08:56 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
131
+ 2025-12-03 16:08:56 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
132
+ Augmented 291 quotes: Counter({'″″': 35, '„”': 34, '「」': 33, '””': 32, '«»': 30, '»«': 29, '《》': 25, '„“': 25, '""': 24, '““': 24})
133
+ 2025-12-03 16:09:00 INFO: Running tagger to retag /local/tmp.5440223/tmpb29e30zw/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
134
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmpb29e30zw/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
135
+ 2025-12-03 16:09:00 INFO: Running tagger in predict mode
136
+ 2025-12-03 16:09:00 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
137
+ 2025-12-03 16:09:02 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
138
+ 2025-12-03 16:09:02 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
139
+ 2025-12-03 16:09:02 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
140
+ 2025-12-03 16:09:02 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
141
+ 2025-12-03 16:09:02 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
142
+ 2025-12-03 16:09:05 INFO: Loading data with batch size 250...
143
+ 2025-12-03 16:09:55 INFO: Start evaluation...
144
+ 2025-12-03 16:14:23 INFO: UPOS XPOS UFeats AllTags
145
+ 2025-12-03 16:14:23 INFO: 91.40 68.78 83.33 68.18
146
+ 2025-12-03 16:14:23 INFO: POS Tagger score: sv_diachronic 68.18
147
+ 2025-12-03 16:14:24 INFO: Running tagger to retag /local/tmp.5440223/tmpb29e30zw/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
148
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmpb29e30zw/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
149
+ 2025-12-03 16:14:24 INFO: Running tagger in predict mode
150
+ 2025-12-03 16:14:24 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
151
+ 2025-12-03 16:14:25 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
152
+ 2025-12-03 16:14:25 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
153
+ 2025-12-03 16:14:25 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
154
+ 2025-12-03 16:14:26 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
155
+ 2025-12-03 16:14:26 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
156
+ 2025-12-03 16:14:26 INFO: Loading data with batch size 250...
157
+ 2025-12-03 16:14:26 INFO: Start evaluation...
158
+ 2025-12-03 16:14:26 INFO: UPOS XPOS UFeats AllTags
159
+ 2025-12-03 16:14:26 INFO: 93.32 90.84 93.32 85.64
160
+ 2025-12-03 16:14:26 INFO: POS Tagger score: sv_diachronic 85.64
161
+ 2025-12-03 16:14:26 INFO: Running tagger to retag /local/tmp.5440223/tmpb29e30zw/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
162
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmpb29e30zw/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
163
+ 2025-12-03 16:14:26 INFO: Running tagger in predict mode
164
+ 2025-12-03 16:14:26 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
165
+ 2025-12-03 16:14:28 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
166
+ 2025-12-03 16:14:28 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
167
+ 2025-12-03 16:14:28 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
168
+ 2025-12-03 16:14:28 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
169
+ 2025-12-03 16:14:28 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
170
+ 2025-12-03 16:14:28 INFO: Loading data with batch size 250...
171
+ 2025-12-03 16:14:28 INFO: Start evaluation...
172
+ 2025-12-03 16:14:29 INFO: UPOS XPOS UFeats AllTags
173
+ 2025-12-03 16:14:29 INFO: 93.14 96.78 95.32 90.28
174
+ 2025-12-03 16:14:29 INFO: POS Tagger score: sv_diachronic 90.28
175
+ Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
176
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5440223/tmpb29e30zw/sv_diachronic.train.gold.conllu
177
+ Swapped 'w1, w2' for 'w1 ,w2' 287 times
178
+ Added 675 new sentences with asdf, zzzz -> asdf,zzzz
179
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5440223/tmpb29e30zw/sv_diachronic.dev.gold.conllu
180
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5440223/tmpb29e30zw/sv_diachronic.test.gold.conllu
181
+ Running stanza dependency parser training…
182
+ 2025-12-03 16:14:46 INFO: Training program called with:
183
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt --batch_size 32 --dropout 0.33
184
+ 2025-12-03 16:14:46 DEBUG: UD_Swedish-diachronic: sv_diachronic
185
+ 2025-12-03 16:14:46 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
186
+ 2025-12-03 16:14:46 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
187
+ 2025-12-03 16:14:46 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
188
+ 2025-12-03 16:14:46 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
189
+ 2025-12-03 16:14:46 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
190
+ 2025-12-03 16:14:46 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
191
+ 2025-12-03 16:14:46 INFO: Running parser in train mode
192
+ 2025-12-03 16:14:46 INFO: Using pretrained contextualized char embedding
193
+ 2025-12-03 16:14:46 INFO: Loading data with batch size 32...
194
+ 2025-12-03 16:14:55 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 84502
195
+ 2025-12-03 16:14:55 INFO: Original data size: 84502
196
+ 2025-12-03 16:14:56 INFO: Augmented data size: 84681
197
+ 2025-12-03 16:15:19 WARNING: sv_diachronic is not a known dataset. Examining the data to choose which xpos vocab to use
198
+ 2025-12-03 16:15:19 INFO: Original length = 84681
199
+ 2025-12-03 16:15:20 INFO: Filtered length = 84681
200
+ 2025-12-03 16:15:39 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
201
+ 2025-12-03 16:15:48 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
202
+ 2025-12-03 16:16:06 DEBUG: 50979 batches created.
203
+ 2025-12-03 16:16:06 DEBUG: 9 batches created.
204
+ 2025-12-03 16:16:06 INFO: Training parser...
205
+ 2025-12-03 16:16:06 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
206
+ 2025-12-03 16:16:06 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
207
+ 2025-12-03 16:16:06 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
208
+ 2025-12-03 16:16:07 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
209
+ 2025-12-03 16:16:10 INFO: Finished STEP 20/50000, loss = 0.000000 (0.034 sec/batch), lr: 0.003000
210
+ 2025-12-03 16:16:11 INFO: Finished STEP 40/50000, loss = 0.000000 (0.034 sec/batch), lr: 0.003000
211
+ 2025-12-03 16:16:12 INFO: Finished STEP 60/50000, loss = 0.000000 (0.034 sec/batch), lr: 0.003000
212
+ 2025-12-03 16:16:12 INFO: Finished STEP 80/50000, loss = 0.000000 (0.034 sec/batch), lr: 0.003000
213
+ 2025-12-03 16:16:13 INFO: Finished STEP 100/50000, loss = 2.544647 (0.032 sec/batch), lr: 0.003000
214
+ 2025-12-03 16:16:13 INFO: Evaluating on dev set...
215
+ 2025-12-03 16:16:13 INFO: LAS MLAS BLEX
216
+ 2025-12-03 16:16:13 INFO: 0.50 0.31 0.63
217
+ 2025-12-03 16:16:13 INFO: step 100: train_loss = 1630.446468, dev_score = 0.0050
218
+ 2025-12-03 16:16:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
219
+ 2025-12-03 16:16:14 INFO: new best model saved.
220
+ 2025-12-03 16:16:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
221
+ 2025-12-03 16:16:14 INFO: new model checkpoint saved.
222
+ 2025-12-03 16:16:15 INFO: Finished STEP 120/50000, loss = 1.238500 (0.034 sec/batch), lr: 0.003000
223
+ 2025-12-03 16:16:16 INFO: Finished STEP 140/50000, loss = 0.948354 (0.032 sec/batch), lr: 0.003000
224
+ 2025-12-03 16:16:16 INFO: Finished STEP 160/50000, loss = 0.951914 (0.034 sec/batch), lr: 0.003000
225
+ 2025-12-03 16:16:17 INFO: Finished STEP 180/50000, loss = 0.857833 (0.035 sec/batch), lr: 0.003000
226
+ 2025-12-03 16:16:18 INFO: Finished STEP 200/50000, loss = 0.565075 (0.033 sec/batch), lr: 0.003000
227
+ 2025-12-03 16:16:18 INFO: Evaluating on dev set...
228
+ 2025-12-03 16:16:18 INFO: LAS MLAS BLEX
229
+ 2025-12-03 16:16:18 INFO: 1.98 0.74 1.47
230
+ 2025-12-03 16:16:18 INFO: step 200: train_loss = 1.330910, dev_score = 0.0198
231
+ 2025-12-03 16:16:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
232
+ 2025-12-03 16:16:18 INFO: new best model saved.
233
+ 2025-12-03 16:16:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
234
+ 2025-12-03 16:16:19 INFO: new model checkpoint saved.
235
+ 2025-12-03 16:16:20 INFO: Finished STEP 220/50000, loss = 0.848067 (0.036 sec/batch), lr: 0.003000
236
+ 2025-12-03 16:16:20 INFO: Finished STEP 240/50000, loss = 0.882011 (0.033 sec/batch), lr: 0.003000
237
+ 2025-12-03 16:16:21 INFO: Finished STEP 260/50000, loss = 1.041156 (0.033 sec/batch), lr: 0.003000
238
+ 2025-12-03 16:16:22 INFO: Finished STEP 280/50000, loss = 0.680701 (0.034 sec/batch), lr: 0.003000
239
+ 2025-12-03 16:16:22 INFO: Finished STEP 300/50000, loss = 1.099256 (0.033 sec/batch), lr: 0.003000
240
+ 2025-12-03 16:16:22 INFO: Evaluating on dev set...
241
+ 2025-12-03 16:16:23 INFO: LAS MLAS BLEX
242
+ 2025-12-03 16:16:23 INFO: 2.23 0.78 1.17
243
+ 2025-12-03 16:16:23 INFO: step 300: train_loss = 1.027015, dev_score = 0.0223
244
+ 2025-12-03 16:16:23 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
245
+ 2025-12-03 16:16:23 INFO: new best model saved.
246
+ 2025-12-03 16:16:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
247
+ 2025-12-03 16:16:24 INFO: new model checkpoint saved.
248
+ 2025-12-03 16:16:24 INFO: Finished STEP 320/50000, loss = 2.176608 (0.038 sec/batch), lr: 0.003000
249
+ 2025-12-03 16:16:25 INFO: Finished STEP 340/50000, loss = 1.932453 (0.034 sec/batch), lr: 0.003000
250
+ 2025-12-03 16:16:26 INFO: Finished STEP 360/50000, loss = 1.561046 (0.033 sec/batch), lr: 0.003000
251
+ 2025-12-03 16:16:26 INFO: Finished STEP 380/50000, loss = 1.603219 (0.033 sec/batch), lr: 0.003000
252
+ 2025-12-03 16:16:27 INFO: Finished STEP 400/50000, loss = 1.655808 (0.032 sec/batch), lr: 0.003000
253
+ 2025-12-03 16:16:27 INFO: Evaluating on dev set...
254
+ 2025-12-03 16:16:28 INFO: LAS MLAS BLEX
255
+ 2025-12-03 16:16:28 INFO: 4.95 3.76 3.76
256
+ 2025-12-03 16:16:28 INFO: step 400: train_loss = 2.058403, dev_score = 0.0495
257
+ 2025-12-03 16:16:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
258
+ 2025-12-03 16:16:28 INFO: new best model saved.
259
+ 2025-12-03 16:16:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
260
+ 2025-12-03 16:16:29 INFO: new model checkpoint saved.
261
+ 2025-12-03 16:16:29 INFO: Finished STEP 420/50000, loss = 1.380271 (0.035 sec/batch), lr: 0.003000
262
+ 2025-12-03 16:16:30 INFO: Finished STEP 440/50000, loss = 1.612059 (0.033 sec/batch), lr: 0.003000
263
+ 2025-12-03 16:16:31 INFO: Finished STEP 460/50000, loss = 1.926968 (0.038 sec/batch), lr: 0.003000
264
+ 2025-12-03 16:16:31 INFO: Finished STEP 480/50000, loss = 1.468068 (0.034 sec/batch), lr: 0.003000
265
+ 2025-12-03 16:16:32 INFO: Finished STEP 500/50000, loss = 1.142256 (0.034 sec/batch), lr: 0.003000
266
+ 2025-12-03 16:16:32 INFO: Evaluating on dev set...
267
+ 2025-12-03 16:16:33 INFO: LAS MLAS BLEX
268
+ 2025-12-03 16:16:33 INFO: 11.14 2.25 3.60
269
+ 2025-12-03 16:16:33 INFO: step 500: train_loss = 1.588690, dev_score = 0.1114
270
+ 2025-12-03 16:16:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
271
+ 2025-12-03 16:16:33 INFO: new best model saved.
272
+ 2025-12-03 16:16:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
273
+ 2025-12-03 16:16:33 INFO: new model checkpoint saved.
274
+ 2025-12-03 16:16:34 INFO: Finished STEP 520/50000, loss = 1.300849 (0.034 sec/batch), lr: 0.003000
275
+ 2025-12-03 16:16:35 INFO: Finished STEP 540/50000, loss = 1.461207 (0.033 sec/batch), lr: 0.003000
276
+ 2025-12-03 16:16:36 INFO: Finished STEP 560/50000, loss = 1.139107 (0.034 sec/batch), lr: 0.003000
277
+ 2025-12-03 16:16:36 INFO: Finished STEP 580/50000, loss = 1.904085 (0.033 sec/batch), lr: 0.003000
278
+ 2025-12-03 16:16:37 INFO: Finished STEP 600/50000, loss = 1.216272 (0.033 sec/batch), lr: 0.003000
279
+ 2025-12-03 16:16:37 INFO: Evaluating on dev set...
280
+ 2025-12-03 16:16:38 INFO: LAS MLAS BLEX
281
+ 2025-12-03 16:16:38 INFO: 14.36 6.17 6.61
282
+ 2025-12-03 16:16:38 INFO: step 600: train_loss = 1.484879, dev_score = 0.1436
283
+ 2025-12-03 16:16:38 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
284
+ 2025-12-03 16:16:38 INFO: new best model saved.
285
+ 2025-12-03 16:16:38 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
286
+ 2025-12-03 16:16:38 INFO: new model checkpoint saved.
287
+ 2025-12-03 16:16:39 INFO: Finished STEP 620/50000, loss = 0.721617 (0.033 sec/batch), lr: 0.003000
288
+ 2025-12-03 16:16:40 INFO: Finished STEP 640/50000, loss = 1.909156 (0.035 sec/batch), lr: 0.003000
289
+ 2025-12-03 16:16:41 INFO: Finished STEP 660/50000, loss = 0.978125 (0.034 sec/batch), lr: 0.003000
290
+ 2025-12-03 16:16:41 INFO: Finished STEP 680/50000, loss = 2.391089 (0.036 sec/batch), lr: 0.003000
291
+ 2025-12-03 16:16:42 INFO: Finished STEP 700/50000, loss = 2.533999 (0.034 sec/batch), lr: 0.003000
292
+ 2025-12-03 16:16:42 INFO: Evaluating on dev set...
293
+ 2025-12-03 16:16:43 INFO: LAS MLAS BLEX
294
+ 2025-12-03 16:16:43 INFO: 8.66 5.80 6.83
295
+ 2025-12-03 16:16:43 INFO: step 700: train_loss = 1.751765, dev_score = 0.0866
296
+ 2025-12-03 16:16:43 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
297
+ 2025-12-03 16:16:43 INFO: new model checkpoint saved.
298
+ 2025-12-03 16:16:44 INFO: Finished STEP 720/50000, loss = 2.019368 (0.035 sec/batch), lr: 0.003000
299
+ 2025-12-03 16:16:45 INFO: Finished STEP 740/50000, loss = 2.141022 (0.038 sec/batch), lr: 0.003000
300
+ 2025-12-03 16:16:45 INFO: Finished STEP 760/50000, loss = 2.237685 (0.036 sec/batch), lr: 0.003000
301
+ 2025-12-03 16:16:46 INFO: Finished STEP 780/50000, loss = 1.872372 (0.033 sec/batch), lr: 0.003000
302
+ 2025-12-03 16:16:47 INFO: Finished STEP 800/50000, loss = 2.076525 (0.035 sec/batch), lr: 0.003000
303
+ 2025-12-03 16:16:47 INFO: Evaluating on dev set...
304
+ 2025-12-03 16:16:48 INFO: LAS MLAS BLEX
305
+ 2025-12-03 16:16:48 INFO: 15.10 5.44 7.38
306
+ 2025-12-03 16:16:48 INFO: step 800: train_loss = 2.225672, dev_score = 0.1510
307
+ 2025-12-03 16:16:48 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
308
+ 2025-12-03 16:16:48 INFO: new best model saved.
309
+ 2025-12-03 16:16:49 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
310
+ 2025-12-03 16:16:49 INFO: new model checkpoint saved.
311
+ 2025-12-03 16:16:49 INFO: Finished STEP 820/50000, loss = 3.131036 (0.035 sec/batch), lr: 0.003000
312
+ 2025-12-03 16:16:50 INFO: Finished STEP 840/50000, loss = 1.733034 (0.035 sec/batch), lr: 0.003000
313
+ 2025-12-03 16:16:51 INFO: Finished STEP 860/50000, loss = 2.062121 (0.036 sec/batch), lr: 0.003000
314
+ 2025-12-03 16:16:51 INFO: Finished STEP 880/50000, loss = 2.321656 (0.038 sec/batch), lr: 0.003000
315
+ 2025-12-03 16:16:52 INFO: Finished STEP 900/50000, loss = 2.393223 (0.034 sec/batch), lr: 0.003000
316
+ 2025-12-03 16:16:52 INFO: Evaluating on dev set...
317
+ 2025-12-03 16:16:53 INFO: LAS MLAS BLEX
318
+ 2025-12-03 16:16:53 INFO: 12.13 6.62 7.35
319
+ 2025-12-03 16:16:53 INFO: step 900: train_loss = 2.193318, dev_score = 0.1213
320
+ 2025-12-03 16:16:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
321
+ 2025-12-03 16:16:53 INFO: new model checkpoint saved.
322
+ 2025-12-03 16:16:54 INFO: Finished STEP 920/50000, loss = 1.626200 (0.035 sec/batch), lr: 0.003000
323
+ 2025-12-03 16:16:55 INFO: Finished STEP 940/50000, loss = 1.677986 (0.035 sec/batch), lr: 0.003000
324
+ 2025-12-03 16:16:55 INFO: Finished STEP 960/50000, loss = 2.482033 (0.035 sec/batch), lr: 0.003000
325
+ 2025-12-03 16:16:56 INFO: Finished STEP 980/50000, loss = 2.240081 (0.036 sec/batch), lr: 0.003000
326
+ 2025-12-03 16:16:57 INFO: Finished STEP 1000/50000, loss = 1.593451 (0.034 sec/batch), lr: 0.003000
327
+ 2025-12-03 16:16:57 INFO: Evaluating on dev set...
328
+ 2025-12-03 16:16:58 INFO: LAS MLAS BLEX
329
+ 2025-12-03 16:16:58 INFO: 16.09 6.51 9.11
330
+ 2025-12-03 16:16:58 INFO: step 1000: train_loss = 2.112306, dev_score = 0.1609
331
+ 2025-12-03 16:16:58 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
332
+ 2025-12-03 16:16:58 INFO: new best model saved.
333
+ 2025-12-03 16:16:58 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
334
+ 2025-12-03 16:16:58 INFO: new model checkpoint saved.
335
+ 2025-12-03 16:16:59 INFO: Finished STEP 1020/50000, loss = 1.938198 (0.036 sec/batch), lr: 0.003000
336
+ 2025-12-03 16:17:00 INFO: Finished STEP 1040/50000, loss = 2.565099 (0.037 sec/batch), lr: 0.003000
337
+ 2025-12-03 16:17:01 INFO: Finished STEP 1060/50000, loss = 1.819492 (0.035 sec/batch), lr: 0.003000
338
+ 2025-12-03 16:17:01 INFO: Finished STEP 1080/50000, loss = 2.274921 (0.037 sec/batch), lr: 0.003000
339
+ 2025-12-03 16:17:02 INFO: Finished STEP 1100/50000, loss = 1.612033 (0.034 sec/batch), lr: 0.003000
340
+ 2025-12-03 16:17:02 INFO: Evaluating on dev set...
341
+ 2025-12-03 16:17:02 INFO: LAS MLAS BLEX
342
+ 2025-12-03 16:17:02 INFO: 17.33 6.94 9.26
343
+ 2025-12-03 16:17:02 INFO: step 1100: train_loss = 2.158392, dev_score = 0.1733
344
+ 2025-12-03 16:17:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
345
+ 2025-12-03 16:17:03 INFO: new best model saved.
346
+ 2025-12-03 16:17:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
347
+ 2025-12-03 16:17:03 INFO: new model checkpoint saved.
348
+ 2025-12-03 16:17:04 INFO: Finished STEP 1120/50000, loss = 1.828422 (0.035 sec/batch), lr: 0.003000
349
+ 2025-12-03 16:17:05 INFO: Finished STEP 1140/50000, loss = 2.403217 (0.034 sec/batch), lr: 0.003000
350
+ 2025-12-03 16:17:05 INFO: Finished STEP 1160/50000, loss = 1.956016 (0.037 sec/batch), lr: 0.003000
351
+ 2025-12-03 16:17:06 INFO: Finished STEP 1180/50000, loss = 2.927859 (0.037 sec/batch), lr: 0.003000
352
+ 2025-12-03 16:17:07 INFO: Finished STEP 1200/50000, loss = 3.263619 (0.036 sec/batch), lr: 0.003000
353
+ 2025-12-03 16:17:07 INFO: Evaluating on dev set...
354
+ 2025-12-03 16:17:07 INFO: LAS MLAS BLEX
355
+ 2025-12-03 16:17:07 INFO: 17.33 7.08 8.02
356
+ 2025-12-03 16:17:07 INFO: step 1200: train_loss = 2.337974, dev_score = 0.1733
357
+ 2025-12-03 16:17:08 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
358
+ 2025-12-03 16:17:08 INFO: new best model saved.
359
+ 2025-12-03 16:17:08 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
360
+ 2025-12-03 16:17:08 INFO: new model checkpoint saved.
361
+ 2025-12-03 16:17:09 INFO: Finished STEP 1220/50000, loss = 2.337204 (0.038 sec/batch), lr: 0.003000
362
+ 2025-12-03 16:17:10 INFO: Finished STEP 1240/50000, loss = 2.144393 (0.038 sec/batch), lr: 0.003000
363
+ 2025-12-03 16:17:10 INFO: Finished STEP 1260/50000, loss = 1.892079 (0.036 sec/batch), lr: 0.003000
364
+ 2025-12-03 16:17:11 INFO: Finished STEP 1280/50000, loss = 3.527288 (0.036 sec/batch), lr: 0.003000
365
+ 2025-12-03 16:17:12 INFO: Finished STEP 1300/50000, loss = 2.162596 (0.036 sec/batch), lr: 0.003000
366
+ 2025-12-03 16:17:12 INFO: Evaluating on dev set...
367
+ 2025-12-03 16:17:12 INFO: LAS MLAS BLEX
368
+ 2025-12-03 16:17:12 INFO: 20.54 4.91 6.05
369
+ 2025-12-03 16:17:12 INFO: step 1300: train_loss = 2.585007, dev_score = 0.2054
370
+ 2025-12-03 16:17:13 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
371
+ 2025-12-03 16:17:13 INFO: new best model saved.
372
+ 2025-12-03 16:17:13 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
373
+ 2025-12-03 16:17:13 INFO: new model checkpoint saved.
374
+ 2025-12-03 16:17:14 INFO: Finished STEP 1320/50000, loss = 1.637469 (0.036 sec/batch), lr: 0.003000
375
+ 2025-12-03 16:17:15 INFO: Finished STEP 1340/50000, loss = 2.203557 (0.036 sec/batch), lr: 0.003000
376
+ 2025-12-03 16:17:15 INFO: Finished STEP 1360/50000, loss = 2.725776 (0.037 sec/batch), lr: 0.003000
377
+ 2025-12-03 16:17:16 INFO: Finished STEP 1380/50000, loss = 3.045333 (0.036 sec/batch), lr: 0.003000
378
+ 2025-12-03 16:17:17 INFO: Finished STEP 1400/50000, loss = 2.141542 (0.035 sec/batch), lr: 0.003000
379
+ 2025-12-03 16:17:17 INFO: Evaluating on dev set...
380
+ 2025-12-03 16:17:17 INFO: LAS MLAS BLEX
381
+ 2025-12-03 16:17:17 INFO: 17.08 6.56 8.49
382
+ 2025-12-03 16:17:17 INFO: step 1400: train_loss = 2.520364, dev_score = 0.1708
383
+ 2025-12-03 16:17:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
384
+ 2025-12-03 16:17:18 INFO: new model checkpoint saved.
385
+ 2025-12-03 16:17:19 INFO: Finished STEP 1420/50000, loss = 2.897650 (0.037 sec/batch), lr: 0.003000
386
+ 2025-12-03 16:17:19 INFO: Finished STEP 1440/50000, loss = 2.777742 (0.036 sec/batch), lr: 0.003000
387
+ 2025-12-03 16:17:20 INFO: Finished STEP 1460/50000, loss = 2.882655 (0.036 sec/batch), lr: 0.003000
388
+ 2025-12-03 16:17:21 INFO: Finished STEP 1480/50000, loss = 3.077738 (0.035 sec/batch), lr: 0.003000
389
+ 2025-12-03 16:17:22 INFO: Finished STEP 1500/50000, loss = 2.251906 (0.037 sec/batch), lr: 0.003000
390
+ 2025-12-03 16:17:22 INFO: Evaluating on dev set...
391
+ 2025-12-03 16:17:22 INFO: LAS MLAS BLEX
392
+ 2025-12-03 16:17:22 INFO: 18.56 4.53 6.42
393
+ 2025-12-03 16:17:22 INFO: step 1500: train_loss = 2.564763, dev_score = 0.1856
394
+ 2025-12-03 16:17:23 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
395
+ 2025-12-03 16:17:23 INFO: new model checkpoint saved.
396
+ 2025-12-03 16:17:23 INFO: Finished STEP 1520/50000, loss = 2.730649 (0.035 sec/batch), lr: 0.003000
397
+ 2025-12-03 16:17:24 INFO: Finished STEP 1540/50000, loss = 2.161053 (0.034 sec/batch), lr: 0.003000
398
+ 2025-12-03 16:17:25 INFO: Finished STEP 1560/50000, loss = 2.637189 (0.035 sec/batch), lr: 0.003000
399
+ 2025-12-03 16:17:26 INFO: Finished STEP 1580/50000, loss = 3.734738 (0.036 sec/batch), lr: 0.003000
400
+ 2025-12-03 16:17:26 INFO: Finished STEP 1600/50000, loss = 2.531164 (0.035 sec/batch), lr: 0.003000
401
+ 2025-12-03 16:17:26 INFO: Evaluating on dev set...
402
+ 2025-12-03 16:17:27 INFO: LAS MLAS BLEX
403
+ 2025-12-03 16:17:27 INFO: 15.84 6.92 7.29
404
+ 2025-12-03 16:17:27 INFO: step 1600: train_loss = 2.502230, dev_score = 0.1584
405
+ 2025-12-03 16:17:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
406
+ 2025-12-03 16:17:28 INFO: new model checkpoint saved.
407
+ 2025-12-03 16:17:28 INFO: Finished STEP 1620/50000, loss = 2.637498 (0.035 sec/batch), lr: 0.003000
408
+ 2025-12-03 16:17:29 INFO: Finished STEP 1640/50000, loss = 2.271445 (0.035 sec/batch), lr: 0.003000
409
+ 2025-12-03 16:17:30 INFO: Finished STEP 1660/50000, loss = 2.368207 (0.036 sec/batch), lr: 0.003000
410
+ 2025-12-03 16:17:30 INFO: Finished STEP 1680/50000, loss = 1.976211 (0.035 sec/batch), lr: 0.003000
411
+ 2025-12-03 16:17:31 INFO: Finished STEP 1700/50000, loss = 3.195338 (0.035 sec/batch), lr: 0.003000
412
+ 2025-12-03 16:17:31 INFO: Evaluating on dev set...
413
+ 2025-12-03 16:17:32 INFO: LAS MLAS BLEX
414
+ 2025-12-03 16:17:32 INFO: 17.82 5.46 6.63
415
+ 2025-12-03 16:17:32 INFO: step 1700: train_loss = 2.483364, dev_score = 0.1782
416
+ 2025-12-03 16:17:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
417
+ 2025-12-03 16:17:33 INFO: new model checkpoint saved.
418
+ 2025-12-03 16:17:33 INFO: Finished STEP 1720/50000, loss = 2.068750 (0.035 sec/batch), lr: 0.003000
419
+ 2025-12-03 16:17:34 INFO: Finished STEP 1740/50000, loss = 2.067800 (0.036 sec/batch), lr: 0.003000
420
+ 2025-12-03 16:17:35 INFO: Finished STEP 1760/50000, loss = 1.790439 (0.037 sec/batch), lr: 0.003000
421
+ 2025-12-03 16:17:36 INFO: Finished STEP 1780/50000, loss = 2.936446 (0.036 sec/batch), lr: 0.003000
422
+ 2025-12-03 16:17:36 INFO: Finished STEP 1800/50000, loss = 2.655161 (0.036 sec/batch), lr: 0.003000
423
+ 2025-12-03 16:17:36 INFO: Evaluating on dev set...
424
+ 2025-12-03 16:17:37 INFO: LAS MLAS BLEX
425
+ 2025-12-03 16:17:37 INFO: 19.80 4.00 5.60
426
+ 2025-12-03 16:17:37 INFO: step 1800: train_loss = 2.423402, dev_score = 0.1980
427
+ 2025-12-03 16:17:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
428
+ 2025-12-03 16:17:37 INFO: new model checkpoint saved.
429
+ 2025-12-03 16:17:38 INFO: Finished STEP 1820/50000, loss = 2.177969 (0.035 sec/batch), lr: 0.003000
430
+ 2025-12-03 16:17:39 INFO: Finished STEP 1840/50000, loss = 2.328897 (0.035 sec/batch), lr: 0.003000
431
+ 2025-12-03 16:17:40 INFO: Finished STEP 1860/50000, loss = 2.487621 (0.036 sec/batch), lr: 0.003000
432
+ 2025-12-03 16:17:40 INFO: Finished STEP 1880/50000, loss = 2.681947 (0.037 sec/batch), lr: 0.003000
433
+ 2025-12-03 16:17:41 INFO: Finished STEP 1900/50000, loss = 2.722781 (0.035 sec/batch), lr: 0.003000
434
+ 2025-12-03 16:17:41 INFO: Evaluating on dev set...
435
+ 2025-12-03 16:17:42 INFO: LAS MLAS BLEX
436
+ 2025-12-03 16:17:42 INFO: 14.60 2.63 3.76
437
+ 2025-12-03 16:17:42 INFO: step 1900: train_loss = 2.752183, dev_score = 0.1460
438
+ 2025-12-03 16:17:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
439
+ 2025-12-03 16:17:42 INFO: new model checkpoint saved.
440
+ 2025-12-03 16:17:43 INFO: Finished STEP 1920/50000, loss = 4.953551 (0.037 sec/batch), lr: 0.003000
441
+ 2025-12-03 16:17:44 INFO: Finished STEP 1940/50000, loss = 2.656226 (0.035 sec/batch), lr: 0.003000
442
+ 2025-12-03 16:17:44 INFO: Finished STEP 1960/50000, loss = 3.238807 (0.036 sec/batch), lr: 0.003000
443
+ 2025-12-03 16:17:45 INFO: Finished STEP 1980/50000, loss = 2.610945 (0.035 sec/batch), lr: 0.003000
444
+ 2025-12-03 16:17:46 INFO: Finished STEP 2000/50000, loss = 3.019795 (0.036 sec/batch), lr: 0.003000
445
+ 2025-12-03 16:17:46 INFO: Evaluating on dev set...
446
+ 2025-12-03 16:17:46 INFO: LAS MLAS BLEX
447
+ 2025-12-03 16:17:46 INFO: 21.04 5.64 6.77
448
+ 2025-12-03 16:17:46 INFO: step 2000: train_loss = 2.768040, dev_score = 0.2104
449
+ 2025-12-03 16:17:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
450
+ 2025-12-03 16:17:47 INFO: new best model saved.
451
+ 2025-12-03 16:17:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
452
+ 2025-12-03 16:17:47 INFO: new model checkpoint saved.
453
+ 2025-12-03 16:17:48 INFO: Finished STEP 2020/50000, loss = 2.635247 (0.036 sec/batch), lr: 0.003000
454
+ 2025-12-03 16:17:49 INFO: Finished STEP 2040/50000, loss = 6.767557 (0.038 sec/batch), lr: 0.003000
455
+ 2025-12-03 16:17:49 INFO: Finished STEP 2060/50000, loss = 3.098332 (0.036 sec/batch), lr: 0.003000
456
+ 2025-12-03 16:17:50 INFO: Finished STEP 2080/50000, loss = 2.531297 (0.035 sec/batch), lr: 0.003000
457
+ 2025-12-03 16:17:51 INFO: Finished STEP 2100/50000, loss = 2.883096 (0.035 sec/batch), lr: 0.003000
458
+ 2025-12-03 16:17:51 INFO: Evaluating on dev set...
459
+ 2025-12-03 16:17:51 INFO: LAS MLAS BLEX
460
+ 2025-12-03 16:17:51 INFO: 18.81 5.77 7.31
461
+ 2025-12-03 16:17:51 INFO: step 2100: train_loss = 2.786879, dev_score = 0.1881
462
+ 2025-12-03 16:17:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
463
+ 2025-12-03 16:17:52 INFO: new model checkpoint saved.
464
+ 2025-12-03 16:17:53 INFO: Finished STEP 2120/50000, loss = 2.654868 (0.037 sec/batch), lr: 0.003000
465
+ 2025-12-03 16:17:53 INFO: Finished STEP 2140/50000, loss = 2.848173 (0.036 sec/batch), lr: 0.003000
466
+ 2025-12-03 16:17:54 INFO: Finished STEP 2160/50000, loss = 2.908323 (0.035 sec/batch), lr: 0.003000
467
+ 2025-12-03 16:17:55 INFO: Finished STEP 2180/50000, loss = 2.927492 (0.037 sec/batch), lr: 0.003000
468
+ 2025-12-03 16:17:56 INFO: Finished STEP 2200/50000, loss = 2.714343 (0.034 sec/batch), lr: 0.003000
469
+ 2025-12-03 16:17:56 INFO: Evaluating on dev set...
470
+ 2025-12-03 16:17:56 INFO: LAS MLAS BLEX
471
+ 2025-12-03 16:17:56 INFO: 17.82 5.39 6.55
472
+ 2025-12-03 16:17:56 INFO: step 2200: train_loss = 2.822561, dev_score = 0.1782
473
+ 2025-12-03 16:17:57 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
474
+ 2025-12-03 16:17:57 INFO: new model checkpoint saved.
475
+ 2025-12-03 16:17:57 INFO: Finished STEP 2220/50000, loss = 2.536452 (0.037 sec/batch), lr: 0.003000
476
+ 2025-12-03 16:17:58 INFO: Finished STEP 2240/50000, loss = 2.872534 (0.035 sec/batch), lr: 0.003000
477
+ 2025-12-03 16:17:59 INFO: Finished STEP 2260/50000, loss = 2.345134 (0.034 sec/batch), lr: 0.003000
478
+ 2025-12-03 16:18:00 INFO: Finished STEP 2280/50000, loss = 2.463192 (0.036 sec/batch), lr: 0.003000
479
+ 2025-12-03 16:18:00 INFO: Finished STEP 2300/50000, loss = 3.496889 (0.035 sec/batch), lr: 0.003000
480
+ 2025-12-03 16:18:00 INFO: Evaluating on dev set...
481
+ 2025-12-03 16:18:01 INFO: LAS MLAS BLEX
482
+ 2025-12-03 16:18:01 INFO: 19.80 7.98 9.51
483
+ 2025-12-03 16:18:01 INFO: step 2300: train_loss = 2.812180, dev_score = 0.1980
484
+ 2025-12-03 16:18:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
485
+ 2025-12-03 16:18:01 INFO: new model checkpoint saved.
486
+ 2025-12-03 16:18:02 INFO: Finished STEP 2320/50000, loss = 2.366118 (0.036 sec/batch), lr: 0.003000
487
+ 2025-12-03 16:18:03 INFO: Finished STEP 2340/50000, loss = 2.439280 (0.035 sec/batch), lr: 0.003000
488
+ 2025-12-03 16:18:04 INFO: Finished STEP 2360/50000, loss = 2.759614 (0.036 sec/batch), lr: 0.003000
489
+ 2025-12-03 16:18:04 INFO: Finished STEP 2380/50000, loss = 2.453089 (0.037 sec/batch), lr: 0.003000
490
+ 2025-12-03 16:18:05 INFO: Finished STEP 2400/50000, loss = 3.469812 (0.037 sec/batch), lr: 0.003000
491
+ 2025-12-03 16:18:05 INFO: Evaluating on dev set...
492
+ 2025-12-03 16:18:06 INFO: LAS MLAS BLEX
493
+ 2025-12-03 16:18:06 INFO: 22.52 8.08 12.69
494
+ 2025-12-03 16:18:06 INFO: step 2400: train_loss = 2.901111, dev_score = 0.2252
495
+ 2025-12-03 16:18:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
496
+ 2025-12-03 16:18:06 INFO: new best model saved.
497
+ 2025-12-03 16:18:07 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
498
+ 2025-12-03 16:18:07 INFO: new model checkpoint saved.
499
+ 2025-12-03 16:18:07 INFO: Finished STEP 2420/50000, loss = 2.935807 (0.035 sec/batch), lr: 0.003000
500
+ 2025-12-03 16:18:08 INFO: Finished STEP 2440/50000, loss = 2.855149 (0.036 sec/batch), lr: 0.003000
501
+ 2025-12-03 16:18:09 INFO: Finished STEP 2460/50000, loss = 2.437950 (0.037 sec/batch), lr: 0.003000
502
+ 2025-12-03 16:18:09 INFO: Finished STEP 2480/50000, loss = 2.520339 (0.037 sec/batch), lr: 0.003000
503
+ 2025-12-03 16:18:10 INFO: Finished STEP 2500/50000, loss = 3.398178 (0.035 sec/batch), lr: 0.003000
504
+ 2025-12-03 16:18:10 INFO: Evaluating on dev set...
505
+ 2025-12-03 16:18:11 INFO: LAS MLAS BLEX
506
+ 2025-12-03 16:18:11 INFO: 24.26 7.00 8.95
507
+ 2025-12-03 16:18:11 INFO: step 2500: train_loss = 2.718300, dev_score = 0.2426
508
+ 2025-12-03 16:18:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
509
+ 2025-12-03 16:18:11 INFO: new best model saved.
510
+ 2025-12-03 16:18:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
511
+ 2025-12-03 16:18:12 INFO: new model checkpoint saved.
512
+ 2025-12-03 16:18:12 INFO: Finished STEP 2520/50000, loss = 2.441736 (0.036 sec/batch), lr: 0.003000
513
+ 2025-12-03 16:18:13 INFO: Finished STEP 2540/50000, loss = 2.816937 (0.035 sec/batch), lr: 0.003000
514
+ 2025-12-03 16:18:14 INFO: Finished STEP 2560/50000, loss = 2.678046 (0.037 sec/batch), lr: 0.003000
515
+ 2025-12-03 16:18:15 INFO: Finished STEP 2580/50000, loss = 3.707354 (0.036 sec/batch), lr: 0.003000
516
+ 2025-12-03 16:18:15 INFO: Finished STEP 2600/50000, loss = 3.227098 (0.035 sec/batch), lr: 0.003000
517
+ 2025-12-03 16:18:15 INFO: Evaluating on dev set...
518
+ 2025-12-03 16:18:16 INFO: LAS MLAS BLEX
519
+ 2025-12-03 16:18:16 INFO: 32.67 18.97 23.51
520
+ 2025-12-03 16:18:16 INFO: step 2600: train_loss = 2.653443, dev_score = 0.3267
521
+ 2025-12-03 16:18:16 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
522
+ 2025-12-03 16:18:16 INFO: new best model saved.
523
+ 2025-12-03 16:18:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
524
+ 2025-12-03 16:18:17 INFO: new model checkpoint saved.
525
+ 2025-12-03 16:18:17 INFO: Finished STEP 2620/50000, loss = 2.361423 (0.034 sec/batch), lr: 0.003000
526
+ 2025-12-03 16:18:18 INFO: Finished STEP 2640/50000, loss = 3.119617 (0.035 sec/batch), lr: 0.003000
527
+ 2025-12-03 16:18:19 INFO: Finished STEP 2660/50000, loss = 2.850250 (0.035 sec/batch), lr: 0.003000
528
+ 2025-12-03 16:18:19 INFO: Finished STEP 2680/50000, loss = 1.845776 (0.035 sec/batch), lr: 0.003000
529
+ 2025-12-03 16:18:20 INFO: Finished STEP 2700/50000, loss = 2.429141 (0.038 sec/batch), lr: 0.003000
530
+ 2025-12-03 16:18:20 INFO: Evaluating on dev set...
531
+ 2025-12-03 16:18:21 INFO: LAS MLAS BLEX
532
+ 2025-12-03 16:18:21 INFO: 26.98 8.94 11.38
533
+ 2025-12-03 16:18:21 INFO: step 2700: train_loss = 2.652709, dev_score = 0.2698
534
+ 2025-12-03 16:18:21 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
535
+ 2025-12-03 16:18:21 INFO: new model checkpoint saved.
536
+ 2025-12-03 16:18:22 INFO: Finished STEP 2720/50000, loss = 2.742228 (0.035 sec/batch), lr: 0.003000
537
+ 2025-12-03 16:18:23 INFO: Finished STEP 2740/50000, loss = 2.495540 (0.035 sec/batch), lr: 0.003000
538
+ 2025-12-03 16:18:23 INFO: Finished STEP 2760/50000, loss = 3.289120 (0.038 sec/batch), lr: 0.003000
539
+ 2025-12-03 16:18:24 INFO: Finished STEP 2780/50000, loss = 5.089163 (0.035 sec/batch), lr: 0.003000
540
+ 2025-12-03 16:18:25 INFO: Finished STEP 2800/50000, loss = 2.216514 (0.036 sec/batch), lr: 0.003000
541
+ 2025-12-03 16:18:25 INFO: Evaluating on dev set...
542
+ 2025-12-03 16:18:25 INFO: LAS MLAS BLEX
543
+ 2025-12-03 16:18:25 INFO: 28.71 12.65 15.10
544
+ 2025-12-03 16:18:25 INFO: step 2800: train_loss = 2.792307, dev_score = 0.2871
545
+ 2025-12-03 16:18:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
546
+ 2025-12-03 16:18:26 INFO: new model checkpoint saved.
547
+ 2025-12-03 16:18:27 INFO: Finished STEP 2820/50000, loss = 3.050649 (0.035 sec/batch), lr: 0.003000
548
+ 2025-12-03 16:18:27 INFO: Finished STEP 2840/50000, loss = 2.810336 (0.035 sec/batch), lr: 0.003000
549
+ 2025-12-03 16:18:28 INFO: Finished STEP 2860/50000, loss = 3.588280 (0.035 sec/batch), lr: 0.003000
550
+ 2025-12-03 16:18:29 INFO: Finished STEP 2880/50000, loss = 2.568533 (0.035 sec/batch), lr: 0.003000
551
+ 2025-12-03 16:18:30 INFO: Finished STEP 2900/50000, loss = 2.860657 (0.036 sec/batch), lr: 0.003000
552
+ 2025-12-03 16:18:30 INFO: Evaluating on dev set...
553
+ 2025-12-03 16:18:30 INFO: LAS MLAS BLEX
554
+ 2025-12-03 16:18:30 INFO: 38.61 25.41 28.69
555
+ 2025-12-03 16:18:30 INFO: step 2900: train_loss = 2.718064, dev_score = 0.3861
556
+ 2025-12-03 16:18:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
557
+ 2025-12-03 16:18:30 INFO: new best model saved.
558
+ 2025-12-03 16:18:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
559
+ 2025-12-03 16:18:31 INFO: new model checkpoint saved.
560
+ 2025-12-03 16:18:32 INFO: Finished STEP 2920/50000, loss = 2.412151 (0.037 sec/batch), lr: 0.003000
561
+ 2025-12-03 16:18:32 INFO: Finished STEP 2940/50000, loss = 2.437013 (0.037 sec/batch), lr: 0.003000
562
+ 2025-12-03 16:18:33 INFO: Finished STEP 2960/50000, loss = 2.900063 (0.040 sec/batch), lr: 0.003000
563
+ 2025-12-03 16:18:34 INFO: Finished STEP 2980/50000, loss = 2.077594 (0.036 sec/batch), lr: 0.003000
564
+ 2025-12-03 16:18:35 INFO: Finished STEP 3000/50000, loss = 3.412100 (0.040 sec/batch), lr: 0.003000
565
+ 2025-12-03 16:18:35 INFO: Evaluating on dev set...
566
+ 2025-12-03 16:18:35 INFO: LAS MLAS BLEX
567
+ 2025-12-03 16:18:35 INFO: 39.11 28.34 31.58
568
+ 2025-12-03 16:18:35 INFO: step 3000: train_loss = 2.965783, dev_score = 0.3911
569
+ 2025-12-03 16:18:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
570
+ 2025-12-03 16:18:35 INFO: new best model saved.
571
+ 2025-12-03 16:18:36 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
572
+ 2025-12-03 16:18:36 INFO: new model checkpoint saved.
573
+ 2025-12-03 16:18:37 INFO: Finished STEP 3020/50000, loss = 2.790848 (0.037 sec/batch), lr: 0.003000
574
+ 2025-12-03 16:18:38 INFO: Finished STEP 3040/50000, loss = 2.839850 (0.038 sec/batch), lr: 0.003000
575
+ 2025-12-03 16:18:38 INFO: Finished STEP 3060/50000, loss = 3.927093 (0.036 sec/batch), lr: 0.003000
576
+ 2025-12-03 16:18:39 INFO: Finished STEP 3080/50000, loss = 2.426868 (0.037 sec/batch), lr: 0.003000
577
+ 2025-12-03 16:18:40 INFO: Finished STEP 3100/50000, loss = 3.054374 (0.036 sec/batch), lr: 0.003000
578
+ 2025-12-03 16:18:40 INFO: Evaluating on dev set...
579
+ 2025-12-03 16:18:40 INFO: LAS MLAS BLEX
580
+ 2025-12-03 16:18:40 INFO: 44.31 26.46 32.10
581
+ 2025-12-03 16:18:40 INFO: step 3100: train_loss = 2.805694, dev_score = 0.4431
582
+ 2025-12-03 16:18:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
583
+ 2025-12-03 16:18:41 INFO: new best model saved.
584
+ 2025-12-03 16:18:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
585
+ 2025-12-03 16:18:41 INFO: new model checkpoint saved.
586
+ 2025-12-03 16:18:42 INFO: Finished STEP 3120/50000, loss = 3.185657 (0.037 sec/batch), lr: 0.003000
587
+ 2025-12-03 16:18:43 INFO: Finished STEP 3140/50000, loss = 2.043775 (0.036 sec/batch), lr: 0.003000
588
+ 2025-12-03 16:18:43 INFO: Finished STEP 3160/50000, loss = 3.097543 (0.036 sec/batch), lr: 0.003000
589
+ 2025-12-03 16:18:44 INFO: Finished STEP 3180/50000, loss = 2.425006 (0.038 sec/batch), lr: 0.003000
590
+ 2025-12-03 16:18:45 INFO: Finished STEP 3200/50000, loss = 2.603312 (0.037 sec/batch), lr: 0.003000
591
+ 2025-12-03 16:18:45 INFO: Evaluating on dev set...
592
+ 2025-12-03 16:18:45 INFO: LAS MLAS BLEX
593
+ 2025-12-03 16:18:45 INFO: 37.87 28.63 33.47
594
+ 2025-12-03 16:18:45 INFO: step 3200: train_loss = 2.828842, dev_score = 0.3787
595
+ 2025-12-03 16:18:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
596
+ 2025-12-03 16:18:46 INFO: new model checkpoint saved.
597
+ 2025-12-03 16:18:47 INFO: Finished STEP 3220/50000, loss = 2.627999 (0.037 sec/batch), lr: 0.003000
598
+ 2025-12-03 16:18:48 INFO: Finished STEP 3240/50000, loss = 2.886214 (0.038 sec/batch), lr: 0.003000
599
+ 2025-12-03 16:18:48 INFO: Finished STEP 3260/50000, loss = 3.165301 (0.036 sec/batch), lr: 0.003000
600
+ 2025-12-03 16:18:49 INFO: Finished STEP 3280/50000, loss = 3.070810 (0.036 sec/batch), lr: 0.003000
601
+ 2025-12-03 16:18:50 INFO: Finished STEP 3300/50000, loss = 2.911757 (0.037 sec/batch), lr: 0.003000
602
+ 2025-12-03 16:18:50 INFO: Evaluating on dev set...
603
+ 2025-12-03 16:18:50 INFO: LAS MLAS BLEX
604
+ 2025-12-03 16:18:50 INFO: 43.32 29.81 36.02
605
+ 2025-12-03 16:18:50 INFO: step 3300: train_loss = 2.809183, dev_score = 0.4332
606
+ 2025-12-03 16:18:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
607
+ 2025-12-03 16:18:51 INFO: new model checkpoint saved.
608
+ 2025-12-03 16:18:52 INFO: Finished STEP 3320/50000, loss = 3.703699 (0.036 sec/batch), lr: 0.003000
609
+ 2025-12-03 16:18:52 INFO: Finished STEP 3340/50000, loss = 2.229795 (0.037 sec/batch), lr: 0.003000
610
+ 2025-12-03 16:18:53 INFO: Finished STEP 3360/50000, loss = 2.598623 (0.037 sec/batch), lr: 0.003000
611
+ 2025-12-03 16:18:54 INFO: Finished STEP 3380/50000, loss = 3.340048 (0.036 sec/batch), lr: 0.003000
612
+ 2025-12-03 16:18:55 INFO: Finished STEP 3400/50000, loss = 3.125400 (0.036 sec/batch), lr: 0.003000
613
+ 2025-12-03 16:18:55 INFO: Evaluating on dev set...
614
+ 2025-12-03 16:18:55 INFO: LAS MLAS BLEX
615
+ 2025-12-03 16:18:55 INFO: 41.34 27.96 30.97
616
+ 2025-12-03 16:18:55 INFO: step 3400: train_loss = 2.884862, dev_score = 0.4134
617
+ 2025-12-03 16:18:56 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
618
+ 2025-12-03 16:18:56 INFO: new model checkpoint saved.
619
+ 2025-12-03 16:18:57 INFO: Finished STEP 3420/50000, loss = 3.178128 (0.037 sec/batch), lr: 0.003000
620
+ 2025-12-03 16:18:57 INFO: Finished STEP 3440/50000, loss = 3.063135 (0.038 sec/batch), lr: 0.003000
621
+ 2025-12-03 16:18:58 INFO: Finished STEP 3460/50000, loss = 3.864271 (0.038 sec/batch), lr: 0.003000
622
+ 2025-12-03 16:18:59 INFO: Finished STEP 3480/50000, loss = 4.055978 (0.036 sec/batch), lr: 0.003000
623
+ 2025-12-03 16:19:00 INFO: Finished STEP 3500/50000, loss = 3.362248 (0.037 sec/batch), lr: 0.003000
624
+ 2025-12-03 16:19:00 INFO: Evaluating on dev set...
625
+ 2025-12-03 16:19:00 INFO: LAS MLAS BLEX
626
+ 2025-12-03 16:19:00 INFO: 25.74 16.70 20.99
627
+ 2025-12-03 16:19:00 INFO: step 3500: train_loss = 3.113213, dev_score = 0.2574
628
+ 2025-12-03 16:19:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
629
+ 2025-12-03 16:19:01 INFO: new model checkpoint saved.
630
+ 2025-12-03 16:19:02 INFO: Finished STEP 3520/50000, loss = 3.160386 (0.038 sec/batch), lr: 0.003000
631
+ 2025-12-03 16:19:02 INFO: Finished STEP 3540/50000, loss = 2.799130 (0.036 sec/batch), lr: 0.003000
632
+ 2025-12-03 16:19:03 INFO: Finished STEP 3560/50000, loss = 2.893708 (0.038 sec/batch), lr: 0.003000
633
+ 2025-12-03 16:19:04 INFO: Finished STEP 3580/50000, loss = 3.332929 (0.037 sec/batch), lr: 0.003000
634
+ 2025-12-03 16:19:05 INFO: Finished STEP 3600/50000, loss = 3.308331 (0.037 sec/batch), lr: 0.003000
635
+ 2025-12-03 16:19:05 INFO: Evaluating on dev set...
636
+ 2025-12-03 16:19:05 INFO: LAS MLAS BLEX
637
+ 2025-12-03 16:19:05 INFO: 24.01 17.04 19.07
638
+ 2025-12-03 16:19:05 INFO: step 3600: train_loss = 2.976930, dev_score = 0.2401
639
+ 2025-12-03 16:19:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
640
+ 2025-12-03 16:19:06 INFO: new model checkpoint saved.
641
+ 2025-12-03 16:19:06 INFO: Finished STEP 3620/50000, loss = 2.658552 (0.037 sec/batch), lr: 0.003000
642
+ 2025-12-03 16:19:07 INFO: Finished STEP 3640/50000, loss = 2.503973 (0.036 sec/batch), lr: 0.003000
643
+ 2025-12-03 16:19:08 INFO: Finished STEP 3660/50000, loss = 2.589127 (0.036 sec/batch), lr: 0.003000
644
+ 2025-12-03 16:19:09 INFO: Finished STEP 3680/50000, loss = 1.667207 (0.036 sec/batch), lr: 0.003000
645
+ 2025-12-03 16:19:09 INFO: Finished STEP 3700/50000, loss = 2.403793 (0.037 sec/batch), lr: 0.003000
646
+ 2025-12-03 16:19:09 INFO: Evaluating on dev set...
647
+ 2025-12-03 16:19:10 INFO: LAS MLAS BLEX
648
+ 2025-12-03 16:19:10 INFO: 36.88 20.04 23.03
649
+ 2025-12-03 16:19:10 INFO: step 3700: train_loss = 2.761385, dev_score = 0.3688
650
+ 2025-12-03 16:19:10 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
651
+ 2025-12-03 16:19:10 INFO: new model checkpoint saved.
652
+ 2025-12-03 16:19:11 INFO: Finished STEP 3720/50000, loss = 2.459764 (0.036 sec/batch), lr: 0.003000
653
+ 2025-12-03 16:19:12 INFO: Finished STEP 3740/50000, loss = 2.585299 (0.036 sec/batch), lr: 0.003000
654
+ 2025-12-03 16:19:13 INFO: Finished STEP 3760/50000, loss = 3.539890 (0.039 sec/batch), lr: 0.003000
655
+ 2025-12-03 16:19:14 INFO: Finished STEP 3780/50000, loss = 2.995441 (0.037 sec/batch), lr: 0.003000
656
+ 2025-12-03 16:19:14 INFO: Finished STEP 3800/50000, loss = 4.468492 (0.037 sec/batch), lr: 0.003000
657
+ 2025-12-03 16:19:14 INFO: Evaluating on dev set...
658
+ 2025-12-03 16:19:15 INFO: LAS MLAS BLEX
659
+ 2025-12-03 16:19:15 INFO: 44.55 30.67 36.13
660
+ 2025-12-03 16:19:15 INFO: step 3800: train_loss = 2.753659, dev_score = 0.4455
661
+ 2025-12-03 16:19:15 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
662
+ 2025-12-03 16:19:15 INFO: new best model saved.
663
+ 2025-12-03 16:19:16 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
664
+ 2025-12-03 16:19:16 INFO: new model checkpoint saved.
665
+ 2025-12-03 16:19:16 INFO: Finished STEP 3820/50000, loss = 2.647424 (0.037 sec/batch), lr: 0.003000
666
+ 2025-12-03 16:19:17 INFO: Finished STEP 3840/50000, loss = 1.907927 (0.039 sec/batch), lr: 0.003000
667
+ 2025-12-03 16:19:18 INFO: Finished STEP 3860/50000, loss = 4.335546 (0.038 sec/batch), lr: 0.003000
668
+ 2025-12-03 16:19:19 INFO: Finished STEP 3880/50000, loss = 2.569001 (0.036 sec/batch), lr: 0.003000
669
+ 2025-12-03 16:19:19 INFO: Finished STEP 3900/50000, loss = 2.036506 (0.037 sec/batch), lr: 0.003000
670
+ 2025-12-03 16:19:19 INFO: Evaluating on dev set...
671
+ 2025-12-03 16:19:20 INFO: LAS MLAS BLEX
672
+ 2025-12-03 16:19:20 INFO: 47.52 30.74 37.47
673
+ 2025-12-03 16:19:20 INFO: step 3900: train_loss = 2.869117, dev_score = 0.4752
674
+ 2025-12-03 16:19:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
675
+ 2025-12-03 16:19:20 INFO: new best model saved.
676
+ 2025-12-03 16:19:21 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
677
+ 2025-12-03 16:19:21 INFO: new model checkpoint saved.
678
+ 2025-12-03 16:19:22 INFO: Finished STEP 3920/50000, loss = 2.400958 (0.037 sec/batch), lr: 0.003000
679
+ 2025-12-03 16:19:22 INFO: Finished STEP 3940/50000, loss = 2.911517 (0.039 sec/batch), lr: 0.003000
680
+ 2025-12-03 16:19:23 INFO: Finished STEP 3960/50000, loss = 2.343444 (0.036 sec/batch), lr: 0.003000
681
+ 2025-12-03 16:19:24 INFO: Finished STEP 3980/50000, loss = 3.208296 (0.036 sec/batch), lr: 0.003000
682
+ 2025-12-03 16:19:25 INFO: Finished STEP 4000/50000, loss = 2.644890 (0.037 sec/batch), lr: 0.003000
683
+ 2025-12-03 16:19:25 INFO: Evaluating on dev set...
684
+ 2025-12-03 16:19:25 INFO: LAS MLAS BLEX
685
+ 2025-12-03 16:19:25 INFO: 41.83 27.56 30.48
686
+ 2025-12-03 16:19:25 INFO: step 4000: train_loss = 2.723412, dev_score = 0.4183
687
+ 2025-12-03 16:19:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
688
+ 2025-12-03 16:19:26 INFO: new model checkpoint saved.
689
+ 2025-12-03 16:19:26 INFO: Finished STEP 4020/50000, loss = 2.191464 (0.037 sec/batch), lr: 0.003000
690
+ 2025-12-03 16:19:27 INFO: Finished STEP 4040/50000, loss = 2.476584 (0.037 sec/batch), lr: 0.003000
691
+ 2025-12-03 16:19:28 INFO: Finished STEP 4060/50000, loss = 1.470418 (0.036 sec/batch), lr: 0.003000
692
+ 2025-12-03 16:19:29 INFO: Finished STEP 4080/50000, loss = 3.445699 (0.039 sec/batch), lr: 0.003000
693
+ 2025-12-03 16:19:29 INFO: Finished STEP 4100/50000, loss = 2.292026 (0.038 sec/batch), lr: 0.003000
694
+ 2025-12-03 16:19:29 INFO: Evaluating on dev set...
695
+ 2025-12-03 16:19:30 INFO: LAS MLAS BLEX
696
+ 2025-12-03 16:19:30 INFO: 43.07 25.42 32.20
697
+ 2025-12-03 16:19:30 INFO: step 4100: train_loss = 2.871121, dev_score = 0.4307
698
+ 2025-12-03 16:19:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
699
+ 2025-12-03 16:19:31 INFO: new model checkpoint saved.
700
+ 2025-12-03 16:19:31 INFO: Finished STEP 4120/50000, loss = 2.831055 (0.037 sec/batch), lr: 0.003000
701
+ 2025-12-03 16:19:32 INFO: Finished STEP 4140/50000, loss = 2.560626 (0.039 sec/batch), lr: 0.003000
702
+ 2025-12-03 16:19:33 INFO: Finished STEP 4160/50000, loss = 3.727808 (0.037 sec/batch), lr: 0.003000
703
+ 2025-12-03 16:19:34 INFO: Finished STEP 4180/50000, loss = 2.795270 (0.037 sec/batch), lr: 0.003000
704
+ 2025-12-03 16:19:34 INFO: Finished STEP 4200/50000, loss = 2.281801 (0.037 sec/batch), lr: 0.003000
705
+ 2025-12-03 16:19:34 INFO: Evaluating on dev set...
706
+ 2025-12-03 16:19:35 INFO: LAS MLAS BLEX
707
+ 2025-12-03 16:19:35 INFO: 47.52 33.40 38.90
708
+ 2025-12-03 16:19:35 INFO: step 4200: train_loss = 2.974722, dev_score = 0.4752
709
+ 2025-12-03 16:19:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
710
+ 2025-12-03 16:19:35 INFO: new best model saved.
711
+ 2025-12-03 16:19:36 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
712
+ 2025-12-03 16:19:36 INFO: new model checkpoint saved.
713
+ 2025-12-03 16:19:36 INFO: Finished STEP 4220/50000, loss = 3.340445 (0.038 sec/batch), lr: 0.003000
714
+ 2025-12-03 16:19:37 INFO: Finished STEP 4240/50000, loss = 2.860859 (0.036 sec/batch), lr: 0.003000
715
+ 2025-12-03 16:19:38 INFO: Finished STEP 4260/50000, loss = 3.377073 (0.039 sec/batch), lr: 0.003000
716
+ 2025-12-03 16:19:39 INFO: Finished STEP 4280/50000, loss = 3.502925 (0.039 sec/batch), lr: 0.003000
717
+ 2025-12-03 16:19:40 INFO: Finished STEP 4300/50000, loss = 3.385393 (0.038 sec/batch), lr: 0.003000
718
+ 2025-12-03 16:19:40 INFO: Evaluating on dev set...
719
+ 2025-12-03 16:19:40 INFO: LAS MLAS BLEX
720
+ 2025-12-03 16:19:40 INFO: 48.51 33.05 39.41
721
+ 2025-12-03 16:19:40 INFO: step 4300: train_loss = 3.251988, dev_score = 0.4851
722
+ 2025-12-03 16:19:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
723
+ 2025-12-03 16:19:40 INFO: new best model saved.
724
+ 2025-12-03 16:19:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
725
+ 2025-12-03 16:19:41 INFO: new model checkpoint saved.
726
+ 2025-12-03 16:19:42 INFO: Finished STEP 4320/50000, loss = 2.750068 (0.037 sec/batch), lr: 0.003000
727
+ 2025-12-03 16:19:43 INFO: Finished STEP 4340/50000, loss = 3.086794 (0.039 sec/batch), lr: 0.003000
728
+ 2025-12-03 16:19:43 INFO: Finished STEP 4360/50000, loss = 2.528754 (0.039 sec/batch), lr: 0.003000
729
+ 2025-12-03 16:19:44 INFO: Finished STEP 4380/50000, loss = 2.073762 (0.038 sec/batch), lr: 0.003000
730
+ 2025-12-03 16:19:45 INFO: Finished STEP 4400/50000, loss = 2.995987 (0.039 sec/batch), lr: 0.003000
731
+ 2025-12-03 16:19:45 INFO: Evaluating on dev set...
732
+ 2025-12-03 16:19:45 INFO: LAS MLAS BLEX
733
+ 2025-12-03 16:19:45 INFO: 48.51 32.20 37.71
734
+ 2025-12-03 16:19:45 INFO: step 4400: train_loss = 2.948329, dev_score = 0.4851
735
+ 2025-12-03 16:19:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
736
+ 2025-12-03 16:19:46 INFO: new best model saved.
737
+ 2025-12-03 16:19:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
738
+ 2025-12-03 16:19:46 INFO: new model checkpoint saved.
739
+ 2025-12-03 16:19:47 INFO: Finished STEP 4420/50000, loss = 3.022907 (0.037 sec/batch), lr: 0.003000
740
+ 2025-12-03 16:19:48 INFO: Finished STEP 4440/50000, loss = 2.624864 (0.038 sec/batch), lr: 0.003000
741
+ 2025-12-03 16:19:48 INFO: Finished STEP 4460/50000, loss = 2.730788 (0.037 sec/batch), lr: 0.003000
742
+ 2025-12-03 16:19:49 INFO: Finished STEP 4480/50000, loss = 2.909690 (0.038 sec/batch), lr: 0.003000
743
+ 2025-12-03 16:19:50 INFO: Finished STEP 4500/50000, loss = 4.465760 (0.038 sec/batch), lr: 0.003000
744
+ 2025-12-03 16:19:50 INFO: Evaluating on dev set...
745
+ 2025-12-03 16:19:51 INFO: LAS MLAS BLEX
746
+ 2025-12-03 16:19:51 INFO: 47.52 32.35 38.24
747
+ 2025-12-03 16:19:51 INFO: step 4500: train_loss = 3.059535, dev_score = 0.4752
748
+ 2025-12-03 16:19:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
749
+ 2025-12-03 16:19:51 INFO: new model checkpoint saved.
750
+ 2025-12-03 16:19:52 INFO: Finished STEP 4520/50000, loss = 2.526277 (0.037 sec/batch), lr: 0.003000
751
+ 2025-12-03 16:19:53 INFO: Finished STEP 4540/50000, loss = 3.287923 (0.038 sec/batch), lr: 0.003000
752
+ 2025-12-03 16:19:53 INFO: Finished STEP 4560/50000, loss = 3.452159 (0.037 sec/batch), lr: 0.003000
753
+ 2025-12-03 16:19:54 INFO: Finished STEP 4580/50000, loss = 2.489482 (0.037 sec/batch), lr: 0.003000
754
+ 2025-12-03 16:19:55 INFO: Finished STEP 4600/50000, loss = 2.805597 (0.038 sec/batch), lr: 0.003000
755
+ 2025-12-03 16:19:55 INFO: Evaluating on dev set...
756
+ 2025-12-03 16:19:56 INFO: LAS MLAS BLEX
757
+ 2025-12-03 16:19:56 INFO: 43.07 28.94 31.91
758
+ 2025-12-03 16:19:56 INFO: step 4600: train_loss = 3.240289, dev_score = 0.4307
759
+ 2025-12-03 16:19:56 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
760
+ 2025-12-03 16:19:56 INFO: new model checkpoint saved.
761
+ 2025-12-03 16:19:57 INFO: Finished STEP 4620/50000, loss = 1.535582 (0.038 sec/batch), lr: 0.003000
762
+ 2025-12-03 16:19:58 INFO: Finished STEP 4640/50000, loss = 2.746229 (0.037 sec/batch), lr: 0.003000
763
+ 2025-12-03 16:19:58 INFO: Finished STEP 4660/50000, loss = 6.811559 (0.038 sec/batch), lr: 0.003000
764
+ 2025-12-03 16:19:59 INFO: Finished STEP 4680/50000, loss = 2.475586 (0.038 sec/batch), lr: 0.003000
765
+ 2025-12-03 16:20:00 INFO: Finished STEP 4700/50000, loss = 4.254340 (0.037 sec/batch), lr: 0.003000
766
+ 2025-12-03 16:20:00 INFO: Evaluating on dev set...
767
+ 2025-12-03 16:20:00 INFO: LAS MLAS BLEX
768
+ 2025-12-03 16:20:00 INFO: 45.54 29.79 35.32
769
+ 2025-12-03 16:20:00 INFO: step 4700: train_loss = 3.005051, dev_score = 0.4554
770
+ 2025-12-03 16:20:01 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
771
+ 2025-12-03 16:20:01 INFO: new model checkpoint saved.
772
+ 2025-12-03 16:20:02 INFO: Finished STEP 4720/50000, loss = 3.124645 (0.037 sec/batch), lr: 0.003000
773
+ 2025-12-03 16:20:03 INFO: Finished STEP 4740/50000, loss = 2.382721 (0.039 sec/batch), lr: 0.003000
774
+ 2025-12-03 16:20:03 INFO: Finished STEP 4760/50000, loss = 2.495845 (0.038 sec/batch), lr: 0.003000
775
+ 2025-12-03 16:20:04 INFO: Finished STEP 4780/50000, loss = 2.852123 (0.037 sec/batch), lr: 0.003000
776
+ 2025-12-03 16:20:05 INFO: Finished STEP 4800/50000, loss = 3.299480 (0.038 sec/batch), lr: 0.003000
777
+ 2025-12-03 16:20:05 INFO: Evaluating on dev set...
778
+ 2025-12-03 16:20:05 INFO: LAS MLAS BLEX
779
+ 2025-12-03 16:20:05 INFO: 47.03 32.07 36.29
780
+ 2025-12-03 16:20:05 INFO: step 4800: train_loss = 3.002937, dev_score = 0.4703
781
+ 2025-12-03 16:20:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
782
+ 2025-12-03 16:20:06 INFO: new model checkpoint saved.
783
+ 2025-12-03 16:20:07 INFO: Finished STEP 4820/50000, loss = 4.108068 (0.038 sec/batch), lr: 0.003000
784
+ 2025-12-03 16:20:08 INFO: Finished STEP 4840/50000, loss = 2.683535 (0.040 sec/batch), lr: 0.003000
785
+ 2025-12-03 16:20:08 INFO: Finished STEP 4860/50000, loss = 2.515482 (0.038 sec/batch), lr: 0.003000
786
+ 2025-12-03 16:20:09 INFO: Finished STEP 4880/50000, loss = 7.754747 (0.038 sec/batch), lr: 0.003000
787
+ 2025-12-03 16:20:10 INFO: Finished STEP 4900/50000, loss = 2.420208 (0.037 sec/batch), lr: 0.003000
788
+ 2025-12-03 16:20:10 INFO: Evaluating on dev set...
789
+ 2025-12-03 16:20:10 INFO: LAS MLAS BLEX
790
+ 2025-12-03 16:20:10 INFO: 49.26 35.17 40.68
791
+ 2025-12-03 16:20:10 INFO: step 4900: train_loss = 3.131104, dev_score = 0.4926
792
+ 2025-12-03 16:20:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
793
+ 2025-12-03 16:20:11 INFO: new best model saved.
794
+ 2025-12-03 16:20:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
795
+ 2025-12-03 16:20:11 INFO: new model checkpoint saved.
796
+ 2025-12-03 16:20:12 INFO: Finished STEP 4920/50000, loss = 2.357713 (0.041 sec/batch), lr: 0.003000
797
+ 2025-12-03 16:20:13 INFO: Finished STEP 4940/50000, loss = 3.332856 (0.037 sec/batch), lr: 0.003000
798
+ 2025-12-03 16:20:14 INFO: Finished STEP 4960/50000, loss = 3.024401 (0.035 sec/batch), lr: 0.003000
799
+ 2025-12-03 16:20:14 INFO: Finished STEP 4980/50000, loss = 3.875421 (0.037 sec/batch), lr: 0.003000
800
+ 2025-12-03 16:20:15 INFO: Finished STEP 5000/50000, loss = 2.674303 (0.037 sec/batch), lr: 0.003000
801
+ 2025-12-03 16:20:15 INFO: Evaluating on dev set...
802
+ 2025-12-03 16:20:16 INFO: LAS MLAS BLEX
803
+ 2025-12-03 16:20:16 INFO: 47.28 30.51 35.17
804
+ 2025-12-03 16:20:16 INFO: step 5000: train_loss = 3.098171, dev_score = 0.4728
805
+ 2025-12-03 16:20:16 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
806
+ 2025-12-03 16:20:16 INFO: new model checkpoint saved.
807
+ 2025-12-03 16:20:17 INFO: Finished STEP 5020/50000, loss = 2.521623 (0.037 sec/batch), lr: 0.003000
808
+ 2025-12-03 16:20:18 INFO: Finished STEP 5040/50000, loss = 4.132445 (0.039 sec/batch), lr: 0.003000
809
+ 2025-12-03 16:20:19 INFO: Finished STEP 5060/50000, loss = 2.462947 (0.037 sec/batch), lr: 0.003000
810
+ 2025-12-03 16:20:19 INFO: Finished STEP 5080/50000, loss = 3.218666 (0.039 sec/batch), lr: 0.003000
811
+ 2025-12-03 16:20:20 INFO: Finished STEP 5100/50000, loss = 4.503098 (0.038 sec/batch), lr: 0.003000
812
+ 2025-12-03 16:20:20 INFO: Evaluating on dev set...
813
+ 2025-12-03 16:20:21 INFO: LAS MLAS BLEX
814
+ 2025-12-03 16:20:21 INFO: 47.28 30.44 38.05
815
+ 2025-12-03 16:20:21 INFO: step 5100: train_loss = 3.067273, dev_score = 0.4728
816
+ 2025-12-03 16:20:21 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
817
+ 2025-12-03 16:20:21 INFO: new model checkpoint saved.
818
+ 2025-12-03 16:20:22 INFO: Finished STEP 5120/50000, loss = 3.300785 (0.037 sec/batch), lr: 0.003000
819
+ 2025-12-03 16:20:23 INFO: Finished STEP 5140/50000, loss = 2.907721 (0.040 sec/batch), lr: 0.003000
820
+ 2025-12-03 16:20:24 INFO: Finished STEP 5160/50000, loss = 2.800373 (0.037 sec/batch), lr: 0.003000
821
+ 2025-12-03 16:20:24 INFO: Finished STEP 5180/50000, loss = 2.987764 (0.038 sec/batch), lr: 0.003000
822
+ 2025-12-03 16:20:25 INFO: Finished STEP 5200/50000, loss = 2.266387 (0.037 sec/batch), lr: 0.003000
823
+ 2025-12-03 16:20:25 INFO: Evaluating on dev set...
824
+ 2025-12-03 16:20:26 INFO: LAS MLAS BLEX
825
+ 2025-12-03 16:20:26 INFO: 46.53 33.40 37.21
826
+ 2025-12-03 16:20:26 INFO: step 5200: train_loss = 3.078057, dev_score = 0.4653
827
+ 2025-12-03 16:20:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
828
+ 2025-12-03 16:20:26 INFO: new model checkpoint saved.
829
+ 2025-12-03 16:20:27 INFO: Finished STEP 5220/50000, loss = 4.658113 (0.041 sec/batch), lr: 0.003000
830
+ 2025-12-03 16:20:28 INFO: Finished STEP 5240/50000, loss = 2.636468 (0.036 sec/batch), lr: 0.003000
831
+ 2025-12-03 16:20:28 INFO: Finished STEP 5260/50000, loss = 2.274219 (0.036 sec/batch), lr: 0.003000
832
+ 2025-12-03 16:20:29 INFO: Finished STEP 5280/50000, loss = 2.514667 (0.037 sec/batch), lr: 0.003000
833
+ 2025-12-03 16:20:30 INFO: Finished STEP 5300/50000, loss = 1.738609 (0.037 sec/batch), lr: 0.003000
834
+ 2025-12-03 16:20:30 INFO: Evaluating on dev set...
835
+ 2025-12-03 16:20:31 INFO: LAS MLAS BLEX
836
+ 2025-12-03 16:20:31 INFO: 45.30 30.44 35.52
837
+ 2025-12-03 16:20:31 INFO: step 5300: train_loss = 2.970610, dev_score = 0.4530
838
+ 2025-12-03 16:20:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
839
+ 2025-12-03 16:20:31 INFO: new model checkpoint saved.
840
+ 2025-12-03 16:20:32 INFO: Finished STEP 5320/50000, loss = 3.239784 (0.037 sec/batch), lr: 0.003000
841
+ 2025-12-03 16:20:33 INFO: Finished STEP 5340/50000, loss = 3.596699 (0.037 sec/batch), lr: 0.003000
842
+ 2025-12-03 16:20:34 INFO: Finished STEP 5360/50000, loss = 3.680211 (0.040 sec/batch), lr: 0.003000
843
+ 2025-12-03 16:20:34 INFO: Finished STEP 5380/50000, loss = 3.672400 (0.037 sec/batch), lr: 0.003000
844
+ 2025-12-03 16:20:35 INFO: Finished STEP 5400/50000, loss = 1.556455 (0.038 sec/batch), lr: 0.003000
845
+ 2025-12-03 16:20:35 INFO: Evaluating on dev set...
846
+ 2025-12-03 16:20:36 INFO: LAS MLAS BLEX
847
+ 2025-12-03 16:20:36 INFO: 47.52 33.55 39.49
848
+ 2025-12-03 16:20:36 INFO: step 5400: train_loss = 3.024760, dev_score = 0.4752
849
+ 2025-12-03 16:20:36 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
850
+ 2025-12-03 16:20:36 INFO: new model checkpoint saved.
851
+ 2025-12-03 16:20:37 INFO: Finished STEP 5420/50000, loss = 3.705180 (0.036 sec/batch), lr: 0.003000
852
+ 2025-12-03 16:20:38 INFO: Finished STEP 5440/50000, loss = 13.936595 (0.037 sec/batch), lr: 0.003000
853
+ 2025-12-03 16:20:39 INFO: Finished STEP 5460/50000, loss = 2.999550 (0.036 sec/batch), lr: 0.003000
854
+ 2025-12-03 16:20:39 INFO: Finished STEP 5480/50000, loss = 2.945333 (0.038 sec/batch), lr: 0.003000
855
+ 2025-12-03 16:20:40 INFO: Finished STEP 5500/50000, loss = 1.997427 (0.038 sec/batch), lr: 0.003000
856
+ 2025-12-03 16:20:40 INFO: Evaluating on dev set...
857
+ 2025-12-03 16:20:41 INFO: LAS MLAS BLEX
858
+ 2025-12-03 16:20:41 INFO: 45.54 28.94 33.19
859
+ 2025-12-03 16:20:41 INFO: step 5500: train_loss = 3.000114, dev_score = 0.4554
860
+ 2025-12-03 16:20:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
861
+ 2025-12-03 16:20:41 INFO: new model checkpoint saved.
862
+ 2025-12-03 16:20:42 INFO: Finished STEP 5520/50000, loss = 2.699526 (0.038 sec/batch), lr: 0.003000
863
+ 2025-12-03 16:20:43 INFO: Finished STEP 5540/50000, loss = 2.186383 (0.037 sec/batch), lr: 0.003000
864
+ 2025-12-03 16:20:44 INFO: Finished STEP 5560/50000, loss = 3.531899 (0.038 sec/batch), lr: 0.003000
865
+ 2025-12-03 16:20:45 INFO: Finished STEP 5580/50000, loss = 3.039118 (0.042 sec/batch), lr: 0.003000
866
+ 2025-12-03 16:20:45 INFO: Finished STEP 5600/50000, loss = 2.736473 (0.037 sec/batch), lr: 0.003000
867
+ 2025-12-03 16:20:45 INFO: Evaluating on dev set...
868
+ 2025-12-03 16:20:46 INFO: LAS MLAS BLEX
869
+ 2025-12-03 16:20:46 INFO: 52.23 39.58 43.79
870
+ 2025-12-03 16:20:46 INFO: step 5600: train_loss = 3.085800, dev_score = 0.5223
871
+ 2025-12-03 16:20:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
872
+ 2025-12-03 16:20:46 INFO: new best model saved.
873
+ 2025-12-03 16:20:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
874
+ 2025-12-03 16:20:47 INFO: new model checkpoint saved.
875
+ 2025-12-03 16:20:48 INFO: Finished STEP 5620/50000, loss = 1.931772 (0.038 sec/batch), lr: 0.003000
876
+ 2025-12-03 16:20:48 INFO: Finished STEP 5640/50000, loss = 3.113855 (0.038 sec/batch), lr: 0.003000
877
+ 2025-12-03 16:20:49 INFO: Finished STEP 5660/50000, loss = 3.021896 (0.037 sec/batch), lr: 0.003000
878
+ 2025-12-03 16:20:50 INFO: Finished STEP 5680/50000, loss = 2.850614 (0.041 sec/batch), lr: 0.003000
879
+ 2025-12-03 16:20:51 INFO: Finished STEP 5700/50000, loss = 4.736036 (0.039 sec/batch), lr: 0.003000
880
+ 2025-12-03 16:20:51 INFO: Evaluating on dev set...
881
+ 2025-12-03 16:20:51 INFO: LAS MLAS BLEX
882
+ 2025-12-03 16:20:51 INFO: 47.52 32.00 35.37
883
+ 2025-12-03 16:20:51 INFO: step 5700: train_loss = 3.287904, dev_score = 0.4752
884
+ 2025-12-03 16:20:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
885
+ 2025-12-03 16:20:52 INFO: new model checkpoint saved.
886
+ 2025-12-03 16:20:53 INFO: Finished STEP 5720/50000, loss = 3.077649 (0.037 sec/batch), lr: 0.003000
887
+ 2025-12-03 16:20:54 INFO: Finished STEP 5740/50000, loss = 2.439712 (0.039 sec/batch), lr: 0.003000
888
+ 2025-12-03 16:20:54 INFO: Finished STEP 5760/50000, loss = 3.239717 (0.037 sec/batch), lr: 0.003000
889
+ 2025-12-03 16:20:55 INFO: Finished STEP 5780/50000, loss = 4.072330 (0.041 sec/batch), lr: 0.003000
890
+ 2025-12-03 16:20:56 INFO: Finished STEP 5800/50000, loss = 2.996701 (0.039 sec/batch), lr: 0.003000
891
+ 2025-12-03 16:20:56 INFO: Evaluating on dev set...
892
+ 2025-12-03 16:20:57 INFO: LAS MLAS BLEX
893
+ 2025-12-03 16:20:57 INFO: 49.26 35.52 40.17
894
+ 2025-12-03 16:20:57 INFO: step 5800: train_loss = 3.352661, dev_score = 0.4926
895
+ 2025-12-03 16:20:57 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
896
+ 2025-12-03 16:20:57 INFO: new model checkpoint saved.
897
+ 2025-12-03 16:20:58 INFO: Finished STEP 5820/50000, loss = 3.411130 (0.039 sec/batch), lr: 0.003000
898
+ 2025-12-03 16:20:59 INFO: Finished STEP 5840/50000, loss = 3.410057 (0.038 sec/batch), lr: 0.003000
899
+ 2025-12-03 16:21:00 INFO: Finished STEP 5860/50000, loss = 2.305581 (0.038 sec/batch), lr: 0.003000
900
+ 2025-12-03 16:21:00 INFO: Finished STEP 5880/50000, loss = 5.576043 (0.040 sec/batch), lr: 0.003000
901
+ 2025-12-03 16:21:01 INFO: Finished STEP 5900/50000, loss = 4.071429 (0.039 sec/batch), lr: 0.003000
902
+ 2025-12-03 16:21:01 INFO: Evaluating on dev set...
903
+ 2025-12-03 16:21:02 INFO: LAS MLAS BLEX
904
+ 2025-12-03 16:21:02 INFO: 48.76 34.11 37.89
905
+ 2025-12-03 16:21:02 INFO: step 5900: train_loss = 3.114969, dev_score = 0.4876
906
+ 2025-12-03 16:21:02 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
907
+ 2025-12-03 16:21:02 INFO: new model checkpoint saved.
908
+ 2025-12-03 16:21:03 INFO: Finished STEP 5920/50000, loss = 2.589014 (0.038 sec/batch), lr: 0.003000
909
+ 2025-12-03 16:21:04 INFO: Finished STEP 5940/50000, loss = 2.701970 (0.038 sec/batch), lr: 0.003000
910
+ 2025-12-03 16:21:05 INFO: Finished STEP 5960/50000, loss = 2.683145 (0.037 sec/batch), lr: 0.003000
911
+ 2025-12-03 16:21:05 INFO: Finished STEP 5980/50000, loss = 2.532533 (0.038 sec/batch), lr: 0.003000
912
+ 2025-12-03 16:21:06 INFO: Finished STEP 6000/50000, loss = 3.076605 (0.038 sec/batch), lr: 0.003000
913
+ 2025-12-03 16:21:06 INFO: Evaluating on dev set...
914
+ 2025-12-03 16:21:07 INFO: LAS MLAS BLEX
915
+ 2025-12-03 16:21:07 INFO: 52.23 39.75 42.71
916
+ 2025-12-03 16:21:07 INFO: step 6000: train_loss = 3.156836, dev_score = 0.5223
917
+ 2025-12-03 16:21:07 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
918
+ 2025-12-03 16:21:07 INFO: new best model saved.
919
+ 2025-12-03 16:21:08 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
920
+ 2025-12-03 16:21:08 INFO: new model checkpoint saved.
921
+ 2025-12-03 16:21:08 INFO: Finished STEP 6020/50000, loss = 2.644441 (0.037 sec/batch), lr: 0.003000
922
+ 2025-12-03 16:21:09 INFO: Finished STEP 6040/50000, loss = 3.979600 (0.038 sec/batch), lr: 0.003000
923
+ 2025-12-03 16:21:10 INFO: Finished STEP 6060/50000, loss = 3.269711 (0.040 sec/batch), lr: 0.003000
924
+ 2025-12-03 16:21:11 INFO: Finished STEP 6080/50000, loss = 3.899095 (0.037 sec/batch), lr: 0.003000
925
+ 2025-12-03 16:21:12 INFO: Finished STEP 6100/50000, loss = 3.321356 (0.037 sec/batch), lr: 0.003000
926
+ 2025-12-03 16:21:12 INFO: Evaluating on dev set...
927
+ 2025-12-03 16:21:12 INFO: LAS MLAS BLEX
928
+ 2025-12-03 16:21:12 INFO: 51.98 36.86 44.07
929
+ 2025-12-03 16:21:12 INFO: step 6100: train_loss = 3.226123, dev_score = 0.5198
930
+ 2025-12-03 16:21:13 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
931
+ 2025-12-03 16:21:13 INFO: new model checkpoint saved.
932
+ 2025-12-03 16:21:13 INFO: Finished STEP 6120/50000, loss = 2.907549 (0.038 sec/batch), lr: 0.003000
933
+ 2025-12-03 16:21:14 INFO: Finished STEP 6140/50000, loss = 3.659036 (0.038 sec/batch), lr: 0.003000
934
+ 2025-12-03 16:21:15 INFO: Finished STEP 6160/50000, loss = 3.667205 (0.040 sec/batch), lr: 0.003000
935
+ 2025-12-03 16:21:16 INFO: Finished STEP 6180/50000, loss = 3.177562 (0.039 sec/batch), lr: 0.003000
936
+ 2025-12-03 16:21:17 INFO: Finished STEP 6200/50000, loss = 3.213892 (0.040 sec/batch), lr: 0.003000
937
+ 2025-12-03 16:21:17 INFO: Evaluating on dev set...
938
+ 2025-12-03 16:21:17 INFO: LAS MLAS BLEX
939
+ 2025-12-03 16:21:17 INFO: 46.53 32.35 35.29
940
+ 2025-12-03 16:21:17 INFO: step 6200: train_loss = 3.308364, dev_score = 0.4653
941
+ 2025-12-03 16:21:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
942
+ 2025-12-03 16:21:18 INFO: new model checkpoint saved.
943
+ 2025-12-03 16:21:19 INFO: Finished STEP 6220/50000, loss = 2.394730 (0.044 sec/batch), lr: 0.003000
944
+ 2025-12-03 16:21:19 INFO: Finished STEP 6240/50000, loss = 4.634624 (0.039 sec/batch), lr: 0.003000
945
+ 2025-12-03 16:21:20 INFO: Finished STEP 6260/50000, loss = 4.376540 (0.042 sec/batch), lr: 0.003000
946
+ 2025-12-03 16:21:21 INFO: Finished STEP 6280/50000, loss = 3.334856 (0.041 sec/batch), lr: 0.003000
947
+ 2025-12-03 16:21:22 INFO: Finished STEP 6300/50000, loss = 2.713651 (0.038 sec/batch), lr: 0.003000
948
+ 2025-12-03 16:21:22 INFO: Evaluating on dev set...
949
+ 2025-12-03 16:21:22 INFO: LAS MLAS BLEX
950
+ 2025-12-03 16:21:22 INFO: 46.29 33.12 36.48
951
+ 2025-12-03 16:21:22 INFO: step 6300: train_loss = 3.222710, dev_score = 0.4629
952
+ 2025-12-03 16:21:23 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
953
+ 2025-12-03 16:21:23 INFO: new model checkpoint saved.
954
+ 2025-12-03 16:21:24 INFO: Finished STEP 6320/50000, loss = 3.268805 (0.044 sec/batch), lr: 0.003000
955
+ 2025-12-03 16:21:25 INFO: Finished STEP 6340/50000, loss = 2.632533 (0.038 sec/batch), lr: 0.003000
956
+ 2025-12-03 16:21:25 INFO: Finished STEP 6360/50000, loss = 3.789367 (0.043 sec/batch), lr: 0.003000
957
+ 2025-12-03 16:21:26 INFO: Finished STEP 6380/50000, loss = 3.423658 (0.038 sec/batch), lr: 0.003000
958
+ 2025-12-03 16:21:27 INFO: Finished STEP 6400/50000, loss = 3.462103 (0.039 sec/batch), lr: 0.003000
959
+ 2025-12-03 16:21:27 INFO: Evaluating on dev set...
960
+ 2025-12-03 16:21:28 INFO: LAS MLAS BLEX
961
+ 2025-12-03 16:21:28 INFO: 50.99 37.47 41.68
962
+ 2025-12-03 16:21:28 INFO: step 6400: train_loss = 3.299784, dev_score = 0.5099
963
+ 2025-12-03 16:21:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
964
+ 2025-12-03 16:21:28 INFO: new model checkpoint saved.
965
+ 2025-12-03 16:21:29 INFO: Finished STEP 6420/50000, loss = 2.859281 (0.039 sec/batch), lr: 0.003000
966
+ 2025-12-03 16:21:30 INFO: Finished STEP 6440/50000, loss = 2.888271 (0.042 sec/batch), lr: 0.003000
967
+ 2025-12-03 16:21:31 INFO: Finished STEP 6460/50000, loss = 2.590820 (0.040 sec/batch), lr: 0.003000
968
+ 2025-12-03 16:21:32 INFO: Finished STEP 6480/50000, loss = 2.824554 (0.042 sec/batch), lr: 0.003000
969
+ 2025-12-03 16:21:32 INFO: Finished STEP 6500/50000, loss = 4.451169 (0.040 sec/batch), lr: 0.003000
970
+ 2025-12-03 16:21:32 INFO: Evaluating on dev set...
971
+ 2025-12-03 16:21:33 INFO: LAS MLAS BLEX
972
+ 2025-12-03 16:21:33 INFO: 50.00 38.05 40.59
973
+ 2025-12-03 16:21:33 INFO: step 6500: train_loss = 3.046541, dev_score = 0.5000
974
+ 2025-12-03 16:21:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
975
+ 2025-12-03 16:21:34 INFO: new model checkpoint saved.
976
+ 2025-12-03 16:21:34 INFO: Finished STEP 6520/50000, loss = 2.452580 (0.038 sec/batch), lr: 0.003000
977
+ 2025-12-03 16:21:35 INFO: Finished STEP 6540/50000, loss = 3.287420 (0.040 sec/batch), lr: 0.003000
978
+ 2025-12-03 16:21:36 INFO: Finished STEP 6560/50000, loss = 4.250047 (0.042 sec/batch), lr: 0.003000
979
+ 2025-12-03 16:21:37 INFO: Finished STEP 6580/50000, loss = 2.958064 (0.040 sec/batch), lr: 0.003000
980
+ 2025-12-03 16:21:38 INFO: Finished STEP 6600/50000, loss = 2.743094 (0.039 sec/batch), lr: 0.003000
981
+ 2025-12-03 16:21:38 INFO: Evaluating on dev set...
982
+ 2025-12-03 16:21:38 INFO: LAS MLAS BLEX
983
+ 2025-12-03 16:21:38 INFO: 51.24 38.56 41.95
984
+ 2025-12-03 16:21:38 INFO: step 6600: train_loss = 3.295462, dev_score = 0.5124
985
+ 2025-12-03 16:21:39 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
986
+ 2025-12-03 16:21:39 INFO: new model checkpoint saved.
987
+ 2025-12-03 16:21:40 INFO: Finished STEP 6620/50000, loss = 3.130455 (0.038 sec/batch), lr: 0.003000
988
+ 2025-12-03 16:21:40 INFO: Finished STEP 6640/50000, loss = 2.628382 (0.039 sec/batch), lr: 0.003000
989
+ 2025-12-03 16:21:41 INFO: Finished STEP 6660/50000, loss = 3.539381 (0.040 sec/batch), lr: 0.003000
990
+ 2025-12-03 16:21:42 INFO: Finished STEP 6680/50000, loss = 3.186551 (0.039 sec/batch), lr: 0.003000
991
+ 2025-12-03 16:21:43 INFO: Finished STEP 6700/50000, loss = 2.653069 (0.041 sec/batch), lr: 0.003000
992
+ 2025-12-03 16:21:43 INFO: Evaluating on dev set...
993
+ 2025-12-03 16:21:43 INFO: LAS MLAS BLEX
994
+ 2025-12-03 16:21:43 INFO: 50.00 36.36 41.86
995
+ 2025-12-03 16:21:43 INFO: step 6700: train_loss = 3.093319, dev_score = 0.5000
996
+ 2025-12-03 16:21:44 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
997
+ 2025-12-03 16:21:44 INFO: new model checkpoint saved.
998
+ 2025-12-03 16:21:45 INFO: Finished STEP 6720/50000, loss = 3.911385 (0.040 sec/batch), lr: 0.003000
999
+ 2025-12-03 16:21:46 INFO: Finished STEP 6740/50000, loss = 2.806738 (0.043 sec/batch), lr: 0.003000
1000
+ 2025-12-03 16:21:47 INFO: Finished STEP 6760/50000, loss = 3.492468 (0.040 sec/batch), lr: 0.003000
1001
+ 2025-12-03 16:21:47 INFO: Finished STEP 6780/50000, loss = 2.965401 (0.041 sec/batch), lr: 0.003000
1002
+ 2025-12-03 16:21:48 INFO: Finished STEP 6800/50000, loss = 3.246824 (0.039 sec/batch), lr: 0.003000
1003
+ 2025-12-03 16:21:48 INFO: Evaluating on dev set...
1004
+ 2025-12-03 16:21:49 INFO: LAS MLAS BLEX
1005
+ 2025-12-03 16:21:49 INFO: 50.00 36.44 43.64
1006
+ 2025-12-03 16:21:49 INFO: step 6800: train_loss = 3.166321, dev_score = 0.5000
1007
+ 2025-12-03 16:21:49 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
1008
+ 2025-12-03 16:21:49 INFO: new model checkpoint saved.
1009
+ 2025-12-03 16:21:50 INFO: Finished STEP 6820/50000, loss = 4.416415 (0.040 sec/batch), lr: 0.003000
1010
+ 2025-12-03 16:21:51 INFO: Finished STEP 6840/50000, loss = 3.102331 (0.040 sec/batch), lr: 0.003000
1011
+ 2025-12-03 16:21:52 INFO: Finished STEP 6860/50000, loss = 3.115081 (0.041 sec/batch), lr: 0.003000
1012
+ 2025-12-03 16:21:53 INFO: Finished STEP 6880/50000, loss = 3.645265 (0.040 sec/batch), lr: 0.003000
1013
+ 2025-12-03 16:21:53 INFO: Finished STEP 6900/50000, loss = 3.775108 (0.044 sec/batch), lr: 0.003000
1014
+ 2025-12-03 16:21:53 INFO: Evaluating on dev set...
1015
+ 2025-12-03 16:21:54 INFO: LAS MLAS BLEX
1016
+ 2025-12-03 16:21:54 INFO: 50.00 35.10 41.44
1017
+ 2025-12-03 16:21:54 INFO: step 6900: train_loss = 3.394606, dev_score = 0.5000
1018
+ 2025-12-03 16:21:55 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
1019
+ 2025-12-03 16:21:55 INFO: new model checkpoint saved.
1020
+ 2025-12-03 16:21:55 INFO: Finished STEP 6920/50000, loss = 2.923577 (0.039 sec/batch), lr: 0.003000
1021
+ 2025-12-03 16:21:56 INFO: Finished STEP 6940/50000, loss = 2.825840 (0.039 sec/batch), lr: 0.003000
1022
+ 2025-12-03 16:21:57 INFO: Finished STEP 6960/50000, loss = 6.931494 (0.040 sec/batch), lr: 0.003000
1023
+ 2025-12-03 16:21:58 INFO: Finished STEP 6980/50000, loss = 2.818433 (0.040 sec/batch), lr: 0.003000
1024
+ 2025-12-03 16:21:59 INFO: Finished STEP 7000/50000, loss = 2.928462 (0.039 sec/batch), lr: 0.003000
1025
+ 2025-12-03 16:21:59 INFO: Evaluating on dev set...
1026
+ 2025-12-03 16:21:59 INFO: LAS MLAS BLEX
1027
+ 2025-12-03 16:21:59 INFO: 49.26 35.22 41.09
1028
+ 2025-12-03 16:21:59 INFO: step 7000: train_loss = 3.173160, dev_score = 0.4926
1029
+ 2025-12-03 16:21:59 INFO: Training ended with 7000 steps.
1030
+ 2025-12-03 16:21:59 INFO: Best dev F1 = 52.23, at iteration = 5600
1031
+ 2025-12-03 16:22:01 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
1032
+ 2025-12-03 16:22:01 INFO: Running parser in predict mode
1033
+ 2025-12-03 16:22:01 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
1034
+ 2025-12-03 16:22:03 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
1035
+ 2025-12-03 16:22:03 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
1036
+ 2025-12-03 16:22:03 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
1037
+ 2025-12-03 16:22:03 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
1038
+ 2025-12-03 16:22:03 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
1039
+ 2025-12-03 16:22:03 INFO: Loading data with batch size 32...
1040
+ 2025-12-03 16:22:03 DEBUG: 9 batches created.
1041
+ 2025-12-03 16:22:04 INFO: F1 scores for each dependency:
1042
+ Note that unlabeled attachment errors hurt the labeled attachment scores
1043
+ acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
1044
+ acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
1045
+ advcl: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
1046
+ advmod: p 0.3750 r 0.4800 f1 0.4211 (25 actual)
1047
+ amod: p 0.8571 r 0.7742 f1 0.8136 (31 actual)
1048
+ appos: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
1049
+ aux: p 0.7273 r 0.7273 f1 0.7273 (11 actual)
1050
+ case: p 0.9107 r 0.9107 f1 0.9107 (56 actual)
1051
+ cc: p 0.6429 r 0.6923 f1 0.6667 (13 actual)
1052
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
1053
+ conj: p 0.1200 r 0.2500 f1 0.1622 (12 actual)
1054
+ cop: p 0.5000 r 0.3333 f1 0.4000 (3 actual)
1055
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
1056
+ det: p 0.8696 r 0.9091 f1 0.8889 (22 actual)
1057
+ expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
1058
+ iobj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
1059
+ mark: p 0.5000 r 0.3333 f1 0.4000 (12 actual)
1060
+ nmod: p 0.1667 r 0.0667 f1 0.0952 (15 actual)
1061
+ nmod:poss: p 0.9444 r 0.8947 f1 0.9189 (19 actual)
1062
+ nsubj: p 0.3636 r 0.7059 f1 0.4800 (17 actual)
1063
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
1064
+ obj: p 0.3261 r 0.6818 f1 0.4412 (22 actual)
1065
+ obl: p 0.3429 r 0.2927 f1 0.3158 (41 actual)
1066
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
1067
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
1068
+ parataxis: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
1069
+ punct: p 0.3269 r 0.3269 f1 0.3269 (52 actual)
1070
+ root: p 0.5556 r 0.5556 f1 0.5556 (9 actual)
1071
+ xcomp: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
1072
+ 2025-12-03 16:22:04 INFO: LAS MLAS BLEX
1073
+ 2025-12-03 16:22:04 INFO: 52.23 39.75 42.71
1074
+ 2025-12-03 16:22:04 INFO: Parser score:
1075
+ 2025-12-03 16:22:04 INFO: sv_diachronic 52.23
1076
+ 2025-12-03 16:22:04 INFO: Finished running dev set on
1077
+ UD_Swedish-diachronic
1078
+ UAS LAS CLAS MLAS BLEX
1079
+ 64.85 52.23 42.71 39.75 42.71
1080
+ 2025-12-03 16:22:04 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt', '--batch_size', '32', '--dropout', '0.33']
1081
+ 2025-12-03 16:22:04 INFO: Running parser in predict mode
1082
+ 2025-12-03 16:22:04 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
1083
+ 2025-12-03 16:22:06 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/conll17.pt
1084
+ 2025-12-03 16:22:06 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
1085
+ 2025-12-03 16:22:06 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
1086
+ 2025-12-03 16:22:06 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
1087
+ 2025-12-03 16:22:06 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
1088
+ 2025-12-03 16:22:06 INFO: Loading data with batch size 32...
1089
+ 2025-12-03 16:22:06 DEBUG: 93 batches created.
1090
+ 2025-12-03 16:22:11 INFO: F1 scores for each dependency:
1091
+ Note that unlabeled attachment errors hurt the labeled attachment scores
1092
+ acl: p 0.0000 r 0.0000 f1 0.0000 (32 actual)
1093
+ acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
1094
+ acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (75 actual)
1095
+ advcl: p 0.0000 r 0.0000 f1 0.0000 (60 actual)
1096
+ advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
1097
+ advmod: p 0.4437 r 0.5299 f1 0.4830 (268 actual)
1098
+ amod: p 0.7778 r 0.7913 f1 0.7845 (230 actual)
1099
+ appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
1100
+ aux: p 0.7765 r 0.7857 f1 0.7811 (84 actual)
1101
+ aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
1102
+ case: p 0.8284 r 0.8284 f1 0.8284 (373 actual)
1103
+ cc: p 0.5460 r 0.5742 f1 0.5597 (155 actual)
1104
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
1105
+ compound:prt: p 0.0000 r 0.0000 f1 0.0000 (21 actual)
1106
+ conj: p 0.0976 r 0.1772 f1 0.1258 (158 actual)
1107
+ cop: p 0.6552 r 0.4130 f1 0.5067 (46 actual)
1108
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
1109
+ dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
1110
+ det: p 0.8209 r 0.7933 f1 0.8068 (208 actual)
1111
+ discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
1112
+ dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
1113
+ expl: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
1114
+ expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
1115
+ fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
1116
+ flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
1117
+ flat:name: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
1118
+ goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
1119
+ iobj: p 0.0000 r 0.0000 f1 0.0000 (14 actual)
1120
+ mark: p 0.6612 r 0.5229 f1 0.5839 (153 actual)
1121
+ nmod: p 0.1333 r 0.0588 f1 0.0816 (102 actual)
1122
+ nmod:poss: p 0.8621 r 0.8803 f1 0.8711 (142 actual)
1123
+ nsubj: p 0.3650 r 0.5357 f1 0.4342 (280 actual)
1124
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
1125
+ nummod: p 0.0000 r 0.0000 f1 0.0000 (10 actual)
1126
+ obj: p 0.3442 r 0.6339 f1 0.4462 (183 actual)
1127
+ obl: p 0.2196 r 0.2014 f1 0.2101 (278 actual)
1128
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
1129
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
1130
+ parataxis: p 0.0000 r 0.0000 f1 0.0000 (18 actual)
1131
+ punct: p 0.3186 r 0.3224 f1 0.3205 (425 actual)
1132
+ reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
1133
+ root: p 0.4949 r 0.4949 f1 0.4949 (99 actual)
1134
+ vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
1135
+ xcomp: p 0.1702 r 0.2133 f1 0.1893 (75 actual)
1136
+ 2025-12-03 16:22:11 INFO: LAS MLAS BLEX
1137
+ 2025-12-03 16:22:11 INFO: 48.15 37.99 39.98
1138
+ 2025-12-03 16:22:11 INFO: Parser score:
1139
+ 2025-12-03 16:22:11 INFO: sv_diachronic 48.15
1140
+ 2025-12-03 16:22:11 INFO: Finished running test set on
1141
+ UD_Swedish-diachronic
1142
+ UAS LAS CLAS MLAS BLEX
1143
+ 61.93 48.15 39.98 37.99 39.98
1144
+ DONE.
1145
+ Full log saved to: logs/log_conll17.pt_sv_diachron_nn_20251203_160846.txt
1146
+ Symlink updated: logs/latest.txt → log_conll17.pt_sv_diachron_nn_20251203_160846.txt
logs/log_diachronic.pt_sv_diachron_de_lit_20251203_150832.txt ADDED
@@ -0,0 +1,188 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ === LOGFILE: logs/log_diachronic.pt_sv_diachron_de_lit_20251203_150832.txt ===
2
+ Language codes: sv diachron de_lit
3
+ Using pretrained model: diachronic.pt
4
+
5
+ Running: python prepare-train-val-test.py sv diachron de_lit
6
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
7
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
8
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
9
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
10
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
11
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
12
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
13
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
14
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
15
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
16
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
17
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-de_lit/de_lit-ud-test.conllu
18
+ Including DigPhil MACHINE in TRAIN (minus gold)…
19
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
20
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
21
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
22
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
23
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
24
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
25
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
26
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
27
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
28
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
29
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
30
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
31
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
32
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
33
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
34
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
35
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
36
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
37
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
38
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
39
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
40
+ Cleaning TRAIN...
41
+ [REMOVED] sent_id=6 ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
42
+ [REMOVED] sent_id=7_8 ERRORS=['Multiple roots found: [5, 10]']
43
+ [REMOVED] sent_id=30_31 ERRORS=['Multiple roots found: [3, 18]']
44
+ [REMOVED] sent_id=35 ERRORS=['Line 36: Invalid token ID or head']
45
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [1, 5]']
46
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [7, 20]']
47
+ [REMOVED] sent_id=8_9 ERRORS=['Multiple roots found: [24, 57]']
48
+ [REMOVED] sent_id=12_13 ERRORS=['Multiple roots found: [11, 16]']
49
+ [REMOVED] sent_id=124_split2 ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
50
+ [REMOVED] sent_id=396 ERRORS=['Token 2: Missing form']
51
+ [REMOVED] sent_id=416 ERRORS=['Token 2: Missing form']
52
+ [REMOVED] sent_id=589 ERRORS=['Token 2: Missing form']
53
+ [REMOVED] sent_id=909 ERRORS=['Token 2: Missing form']
54
+ [REMOVED] sent_id=912 ERRORS=['Token 2: Missing form']
55
+ [REMOVED] sent_id=3_split1 ERRORS=['Multiple roots found: [4, 15, 17]']
56
+ [REMOVED] sent_id=3_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
57
+ [REMOVED] sent_id=3_4 ERRORS=['Multiple roots found: [1, 5]']
58
+ [REMOVED] sent_id=5_6 ERRORS=['Multiple roots found: [3, 24]']
59
+ [REMOVED] sent_id=11_12_13 ERRORS=['Multiple roots found: [5, 17, 25]']
60
+ [REMOVED] sent_id=119 ERRORS=['Token 2: Missing form']
61
+ [REMOVED] sent_id=179 ERRORS=['Token 2: Missing form']
62
+ [REMOVED] sent_id=188 ERRORS=['Token 2: Missing form']
63
+ [REMOVED] sent_id=223 ERRORS=['Token 2: Missing form']
64
+ [REMOVED] sent_id=268 ERRORS=['Token 2: Missing form']
65
+ [REMOVED] sent_id=325 ERRORS=['Token 2: Missing form']
66
+ [REMOVED] sent_id=388 ERRORS=['Token 2: Missing form']
67
+ [REMOVED] sent_id=399 ERRORS=['Token 2: Missing form']
68
+ [REMOVED] sent_id=475 ERRORS=['Token 2: Missing form']
69
+ [REMOVED] sent_id=505 ERRORS=['Token 2: Missing form']
70
+ [REMOVED] sent_id=520 ERRORS=['Token 2: Missing form']
71
+ [REMOVED] sent_id=562 ERRORS=['Token 2: Missing form']
72
+ [REMOVED] sent_id=669 ERRORS=['Token 2: Missing form']
73
+ [REMOVED] sent_id=711 ERRORS=['Token 2: Missing form']
74
+ [REMOVED] sent_id=731 ERRORS=['Token 2: Missing form']
75
+ [REMOVED] sent_id=867 ERRORS=['Token 2: Missing form']
76
+ [REMOVED] sent_id=884 ERRORS=['Token 2: Missing form']
77
+ [REMOVED] sent_id=923 ERRORS=['Token 2: Missing form']
78
+ [REMOVED] sent_id=939 ERRORS=['Token 2: Missing form']
79
+ [REMOVED] sent_id=1086 ERRORS=['Token 2: Missing form']
80
+ [REMOVED] sent_id=1179 ERRORS=['Token 2: Missing form']
81
+ [REMOVED] sent_id=1251 ERRORS=['Token 2: Missing form']
82
+ [REMOVED] sent_id=1345 ERRORS=['Token 2: Missing form']
83
+ [REMOVED] sent_id=1459 ERRORS=['Token 2: Missing form']
84
+ [REMOVED] sent_id=1656 ERRORS=['Token 2: Missing form']
85
+ [REMOVED] sent_id=1669 ERRORS=['Token 2: Missing form']
86
+ [REMOVED] sent_id=87_88 ERRORS=['Multiple roots found: [3, 6]']
87
+ [REMOVED] sent_id=65_split2_66_split2 ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
88
+ [REMOVED] sent_id=25 ERRORS=['Token 2: Missing form']
89
+ [REMOVED] sent_id=136 ERRORS=['Token 2: Missing form']
90
+ [REMOVED] sent_id=208 ERRORS=['Token 2: Missing form']
91
+ [REMOVED] sent_id=230 ERRORS=['Token 2: Missing form']
92
+ [REMOVED] sent_id=245 ERRORS=['Token 2: Missing form']
93
+ [REMOVED] sent_id=276 ERRORS=['Token 2: Missing form']
94
+ [REMOVED] sent_id=320 ERRORS=['Token 2: Missing form']
95
+ [REMOVED] sent_id=366 ERRORS=['Token 2: Missing form']
96
+ [REMOVED] sent_id=519 ERRORS=['Token 2: Missing form']
97
+ [REMOVED] sent_id=569 ERRORS=['Token 2: Missing form']
98
+ [REMOVED] sent_id=50_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
99
+ [REMOVED] sent_id=53_54 ERRORS=['Multiple roots found: [27, 91]']
100
+ [REMOVED] sent_id=55_56_57 ERRORS=['Multiple roots found: [2, 4, 13]']
101
+ [REMOVED] sent_id=17_split1 ERRORS=['Multiple roots found: [2, 14, 17]']
102
+ [REMOVED] sent_id=17_split2 ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
103
+ [REMOVED] sent_id=19_split1 ERRORS=['Multiple roots found: [3, 31]']
104
+ Cleaning DEV...
105
+ [REMOVED] sent_id=33 ERRORS=['Token 15: Missing deprel']
106
+ Cleaning TEST...
107
+ Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (68172 valid sentences)
108
+ Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
109
+ Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
110
+ Done.
111
+ Sourcing scripts/config_alvis.sh
112
+ Running stanza dataset preparation…
113
+ 2025-12-03 15:08:42 INFO: Datasets program called with:
114
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
115
+ 2025-12-03 15:08:42 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
116
+
117
+ 2025-12-03 15:08:42 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
118
+ 2025-12-03 15:08:42 DEBUG: Processing parameter "processors"...
119
+ 2025-12-03 15:08:42 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
120
+ 2025-12-03 15:08:42 INFO: Downloading these customized packages for language: sv (Swedish)...
121
+ =======================
122
+ | Processor | Package |
123
+ -----------------------
124
+ =======================
125
+
126
+ 2025-12-03 15:08:42 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
127
+ 2025-12-03 15:08:42 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
128
+ 2025-12-03 15:08:42 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
129
+ 2025-12-03 15:08:42 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
130
+ Augmented 188 quotes: Counter({'«»': 23, '»«': 22, '″″': 22, '""': 20, '““': 20, '„”': 19, '「」': 18, '《》': 16, '„“': 15, '””': 13})
131
+ 2025-12-03 15:08:46 INFO: Running tagger to retag /local/tmp.5440223/tmp8xxzfsmg/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
132
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmp8xxzfsmg/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
133
+ 2025-12-03 15:08:46 INFO: Running tagger in predict mode
134
+ 2025-12-03 15:08:46 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
135
+ 2025-12-03 15:08:48 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
136
+ 2025-12-03 15:08:48 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
137
+ 2025-12-03 15:08:48 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
138
+ 2025-12-03 15:08:48 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
139
+ 2025-12-03 15:08:49 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
140
+ 2025-12-03 15:08:52 INFO: Loading data with batch size 250...
141
+ 2025-12-03 15:09:33 INFO: Start evaluation...
142
+ 2025-12-03 15:13:12 INFO: UPOS XPOS UFeats AllTags
143
+ 2025-12-03 15:13:12 INFO: 92.42 77.41 88.14 75.59
144
+ 2025-12-03 15:13:12 INFO: POS Tagger score: sv_diachronic 75.59
145
+ 2025-12-03 15:13:13 INFO: Running tagger to retag /local/tmp.5440223/tmp8xxzfsmg/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
146
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmp8xxzfsmg/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
147
+ 2025-12-03 15:13:13 INFO: Running tagger in predict mode
148
+ 2025-12-03 15:13:13 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
149
+ 2025-12-03 15:13:15 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
150
+ 2025-12-03 15:13:15 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
151
+ 2025-12-03 15:13:15 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
152
+ 2025-12-03 15:13:15 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
153
+ 2025-12-03 15:13:16 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
154
+ 2025-12-03 15:13:16 INFO: Loading data with batch size 250...
155
+ 2025-12-03 15:13:16 INFO: Start evaluation...
156
+ 2025-12-03 15:13:16 INFO: UPOS XPOS UFeats AllTags
157
+ 2025-12-03 15:13:16 INFO: 88.61 79.46 83.42 74.75
158
+ 2025-12-03 15:13:16 INFO: POS Tagger score: sv_diachronic 74.75
159
+ 2025-12-03 15:13:16 INFO: Running tagger to retag /local/tmp.5440223/tmp8xxzfsmg/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
160
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmp8xxzfsmg/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
161
+ 2025-12-03 15:13:16 INFO: Running tagger in predict mode
162
+ 2025-12-03 15:13:16 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
163
+ 2025-12-03 15:13:19 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
164
+ 2025-12-03 15:13:19 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
165
+ 2025-12-03 15:13:19 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
166
+ 2025-12-03 15:13:19 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
167
+ 2025-12-03 15:13:19 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
168
+ 2025-12-03 15:13:19 INFO: Loading data with batch size 250...
169
+ 2025-12-03 15:13:19 INFO: Start evaluation...
170
+ 2025-12-03 15:13:20 INFO: UPOS XPOS UFeats AllTags
171
+ 2025-12-03 15:13:20 INFO: 89.61 86.64 86.69 81.18
172
+ 2025-12-03 15:13:20 INFO: POS Tagger score: sv_diachronic 81.18
173
+ Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
174
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5440223/tmp8xxzfsmg/sv_diachronic.train.gold.conllu
175
+ Swapped 'w1, w2' for 'w1 ,w2' 141 times
176
+ Added 573 new sentences with asdf, zzzz -> asdf,zzzz
177
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5440223/tmp8xxzfsmg/sv_diachronic.dev.gold.conllu
178
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5440223/tmp8xxzfsmg/sv_diachronic.test.gold.conllu
179
+ Running stanza dependency parser training…
180
+ 2025-12-03 15:13:34 INFO: Training program called with:
181
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt --batch_size 32 --dropout 0.33
182
+ 2025-12-03 15:13:34 DEBUG: UD_Swedish-diachronic: sv_diachronic
183
+ 2025-12-03 15:13:34 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
184
+ 2025-12-03 15:13:34 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
185
+ 2025-12-03 15:13:34 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt exists, skipping!
186
+ DONE.
187
+ Full log saved to: logs/log_diachronic.pt_sv_diachron_de_lit_20251203_150832.txt
188
+ Symlink updated: logs/latest.txt → log_diachronic.pt_sv_diachron_de_lit_20251203_150832.txt
logs/log_diachronic.pt_sv_diachron_de_lit_20251203_151554.txt ADDED
@@ -0,0 +1,650 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ === LOGFILE: logs/log_diachronic.pt_sv_diachron_de_lit_20251203_151554.txt ===
2
+ Language codes: sv diachron de_lit
3
+ Using pretrained model: diachronic.pt
4
+
5
+ Running: python prepare-train-val-test.py sv diachron de_lit
6
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
7
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
8
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
9
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
10
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
11
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
12
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
13
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
14
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
15
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
16
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
17
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-de_lit/de_lit-ud-test.conllu
18
+ Including DigPhil MACHINE in TRAIN (minus gold)…
19
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
20
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
21
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
22
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
23
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
24
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
25
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
26
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
27
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
28
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
29
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
30
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
31
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
32
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
33
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
34
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
35
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
36
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
37
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
38
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
39
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
40
+ Cleaning TRAIN...
41
+ [REMOVED] sent_id=6 ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
42
+ [REMOVED] sent_id=7_8 ERRORS=['Multiple roots found: [5, 10]']
43
+ [REMOVED] sent_id=30_31 ERRORS=['Multiple roots found: [3, 18]']
44
+ [REMOVED] sent_id=35 ERRORS=['Line 36: Invalid token ID or head']
45
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [1, 5]']
46
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [7, 20]']
47
+ [REMOVED] sent_id=8_9 ERRORS=['Multiple roots found: [24, 57]']
48
+ [REMOVED] sent_id=12_13 ERRORS=['Multiple roots found: [11, 16]']
49
+ [REMOVED] sent_id=124_split2 ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
50
+ [REMOVED] sent_id=396 ERRORS=['Token 2: Missing form']
51
+ [REMOVED] sent_id=416 ERRORS=['Token 2: Missing form']
52
+ [REMOVED] sent_id=589 ERRORS=['Token 2: Missing form']
53
+ [REMOVED] sent_id=909 ERRORS=['Token 2: Missing form']
54
+ [REMOVED] sent_id=912 ERRORS=['Token 2: Missing form']
55
+ [REMOVED] sent_id=3_split1 ERRORS=['Multiple roots found: [4, 15, 17]']
56
+ [REMOVED] sent_id=3_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
57
+ [REMOVED] sent_id=3_4 ERRORS=['Multiple roots found: [1, 5]']
58
+ [REMOVED] sent_id=5_6 ERRORS=['Multiple roots found: [3, 24]']
59
+ [REMOVED] sent_id=11_12_13 ERRORS=['Multiple roots found: [5, 17, 25]']
60
+ [REMOVED] sent_id=119 ERRORS=['Token 2: Missing form']
61
+ [REMOVED] sent_id=179 ERRORS=['Token 2: Missing form']
62
+ [REMOVED] sent_id=188 ERRORS=['Token 2: Missing form']
63
+ [REMOVED] sent_id=223 ERRORS=['Token 2: Missing form']
64
+ [REMOVED] sent_id=268 ERRORS=['Token 2: Missing form']
65
+ [REMOVED] sent_id=325 ERRORS=['Token 2: Missing form']
66
+ [REMOVED] sent_id=388 ERRORS=['Token 2: Missing form']
67
+ [REMOVED] sent_id=399 ERRORS=['Token 2: Missing form']
68
+ [REMOVED] sent_id=475 ERRORS=['Token 2: Missing form']
69
+ [REMOVED] sent_id=505 ERRORS=['Token 2: Missing form']
70
+ [REMOVED] sent_id=520 ERRORS=['Token 2: Missing form']
71
+ [REMOVED] sent_id=562 ERRORS=['Token 2: Missing form']
72
+ [REMOVED] sent_id=669 ERRORS=['Token 2: Missing form']
73
+ [REMOVED] sent_id=711 ERRORS=['Token 2: Missing form']
74
+ [REMOVED] sent_id=731 ERRORS=['Token 2: Missing form']
75
+ [REMOVED] sent_id=867 ERRORS=['Token 2: Missing form']
76
+ [REMOVED] sent_id=884 ERRORS=['Token 2: Missing form']
77
+ [REMOVED] sent_id=923 ERRORS=['Token 2: Missing form']
78
+ [REMOVED] sent_id=939 ERRORS=['Token 2: Missing form']
79
+ [REMOVED] sent_id=1086 ERRORS=['Token 2: Missing form']
80
+ [REMOVED] sent_id=1179 ERRORS=['Token 2: Missing form']
81
+ [REMOVED] sent_id=1251 ERRORS=['Token 2: Missing form']
82
+ [REMOVED] sent_id=1345 ERRORS=['Token 2: Missing form']
83
+ [REMOVED] sent_id=1459 ERRORS=['Token 2: Missing form']
84
+ [REMOVED] sent_id=1656 ERRORS=['Token 2: Missing form']
85
+ [REMOVED] sent_id=1669 ERRORS=['Token 2: Missing form']
86
+ [REMOVED] sent_id=87_88 ERRORS=['Multiple roots found: [3, 6]']
87
+ [REMOVED] sent_id=65_split2_66_split2 ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
88
+ [REMOVED] sent_id=25 ERRORS=['Token 2: Missing form']
89
+ [REMOVED] sent_id=136 ERRORS=['Token 2: Missing form']
90
+ [REMOVED] sent_id=208 ERRORS=['Token 2: Missing form']
91
+ [REMOVED] sent_id=230 ERRORS=['Token 2: Missing form']
92
+ [REMOVED] sent_id=245 ERRORS=['Token 2: Missing form']
93
+ [REMOVED] sent_id=276 ERRORS=['Token 2: Missing form']
94
+ [REMOVED] sent_id=320 ERRORS=['Token 2: Missing form']
95
+ [REMOVED] sent_id=366 ERRORS=['Token 2: Missing form']
96
+ [REMOVED] sent_id=519 ERRORS=['Token 2: Missing form']
97
+ [REMOVED] sent_id=569 ERRORS=['Token 2: Missing form']
98
+ [REMOVED] sent_id=50_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
99
+ [REMOVED] sent_id=53_54 ERRORS=['Multiple roots found: [27, 91]']
100
+ [REMOVED] sent_id=55_56_57 ERRORS=['Multiple roots found: [2, 4, 13]']
101
+ [REMOVED] sent_id=17_split1 ERRORS=['Multiple roots found: [2, 14, 17]']
102
+ [REMOVED] sent_id=17_split2 ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
103
+ [REMOVED] sent_id=19_split1 ERRORS=['Multiple roots found: [3, 31]']
104
+ Cleaning DEV...
105
+ [REMOVED] sent_id=33 ERRORS=['Token 15: Missing deprel']
106
+ Cleaning TEST...
107
+ Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (68172 valid sentences)
108
+ Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
109
+ Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
110
+ Done.
111
+ Sourcing scripts/config_alvis.sh
112
+ Running stanza dataset preparation…
113
+ 2025-12-03 15:16:04 INFO: Datasets program called with:
114
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
115
+ 2025-12-03 15:16:04 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
116
+
117
+ 2025-12-03 15:16:04 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
118
+ 2025-12-03 15:16:04 DEBUG: Processing parameter "processors"...
119
+ 2025-12-03 15:16:04 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
120
+ 2025-12-03 15:16:04 INFO: Downloading these customized packages for language: sv (Swedish)...
121
+ =======================
122
+ | Processor | Package |
123
+ -----------------------
124
+ =======================
125
+
126
+ 2025-12-03 15:16:04 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
127
+ 2025-12-03 15:16:04 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
128
+ 2025-12-03 15:16:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
129
+ 2025-12-03 15:16:04 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
130
+ Augmented 188 quotes: Counter({'«»': 23, '»«': 22, '″″': 22, '""': 20, '““': 20, '„”': 19, '「」': 18, '《》': 16, '„“': 15, '””': 13})
131
+ 2025-12-03 15:16:07 INFO: Running tagger to retag /local/tmp.5440223/tmpbokazz5p/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
132
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmpbokazz5p/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
133
+ 2025-12-03 15:16:07 INFO: Running tagger in predict mode
134
+ 2025-12-03 15:16:07 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
135
+ 2025-12-03 15:16:10 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
136
+ 2025-12-03 15:16:10 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
137
+ 2025-12-03 15:16:10 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
138
+ 2025-12-03 15:16:10 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
139
+ 2025-12-03 15:16:10 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
140
+ 2025-12-03 15:16:13 INFO: Loading data with batch size 250...
141
+ 2025-12-03 15:16:54 INFO: Start evaluation...
142
+ 2025-12-03 15:20:34 INFO: UPOS XPOS UFeats AllTags
143
+ 2025-12-03 15:20:34 INFO: 92.42 77.41 88.14 75.59
144
+ 2025-12-03 15:20:34 INFO: POS Tagger score: sv_diachronic 75.59
145
+ 2025-12-03 15:20:35 INFO: Running tagger to retag /local/tmp.5440223/tmpbokazz5p/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
146
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmpbokazz5p/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
147
+ 2025-12-03 15:20:35 INFO: Running tagger in predict mode
148
+ 2025-12-03 15:20:35 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
149
+ 2025-12-03 15:20:37 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
150
+ 2025-12-03 15:20:37 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
151
+ 2025-12-03 15:20:37 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
152
+ 2025-12-03 15:20:37 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
153
+ 2025-12-03 15:20:38 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
154
+ 2025-12-03 15:20:38 INFO: Loading data with batch size 250...
155
+ 2025-12-03 15:20:38 INFO: Start evaluation...
156
+ 2025-12-03 15:20:38 INFO: UPOS XPOS UFeats AllTags
157
+ 2025-12-03 15:20:38 INFO: 88.61 79.46 83.42 74.75
158
+ 2025-12-03 15:20:38 INFO: POS Tagger score: sv_diachronic 74.75
159
+ 2025-12-03 15:20:38 INFO: Running tagger to retag /local/tmp.5440223/tmpbokazz5p/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
160
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5440223/tmpbokazz5p/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
161
+ 2025-12-03 15:20:38 INFO: Running tagger in predict mode
162
+ 2025-12-03 15:20:38 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
163
+ 2025-12-03 15:20:41 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
164
+ 2025-12-03 15:20:41 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
165
+ 2025-12-03 15:20:41 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
166
+ 2025-12-03 15:20:41 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
167
+ 2025-12-03 15:20:41 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
168
+ 2025-12-03 15:20:41 INFO: Loading data with batch size 250...
169
+ 2025-12-03 15:20:41 INFO: Start evaluation...
170
+ 2025-12-03 15:20:42 INFO: UPOS XPOS UFeats AllTags
171
+ 2025-12-03 15:20:42 INFO: 89.61 86.64 86.69 81.18
172
+ 2025-12-03 15:20:42 INFO: POS Tagger score: sv_diachronic 81.18
173
+ Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
174
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5440223/tmpbokazz5p/sv_diachronic.train.gold.conllu
175
+ Swapped 'w1, w2' for 'w1 ,w2' 141 times
176
+ Added 573 new sentences with asdf, zzzz -> asdf,zzzz
177
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5440223/tmpbokazz5p/sv_diachronic.dev.gold.conllu
178
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5440223/tmpbokazz5p/sv_diachronic.test.gold.conllu
179
+ Running stanza dependency parser training…
180
+ 2025-12-03 15:20:57 INFO: Training program called with:
181
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt --batch_size 32 --dropout 0.33
182
+ 2025-12-03 15:20:57 DEBUG: UD_Swedish-diachronic: sv_diachronic
183
+ 2025-12-03 15:20:57 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
184
+ 2025-12-03 15:20:57 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
185
+ 2025-12-03 15:20:57 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
186
+ 2025-12-03 15:20:57 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
187
+ 2025-12-03 15:20:57 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
188
+ 2025-12-03 15:20:57 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--batch_size', '32', '--dropout', '0.33']
189
+ 2025-12-03 15:20:57 INFO: Running parser in train mode
190
+ 2025-12-03 15:20:57 INFO: Using pretrained contextualized char embedding
191
+ 2025-12-03 15:20:57 INFO: Loading data with batch size 32...
192
+ 2025-12-03 15:21:05 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 68745
193
+ 2025-12-03 15:21:05 INFO: Original data size: 68745
194
+ 2025-12-03 15:21:05 INFO: Augmented data size: 69070
195
+ 2025-12-03 15:21:24 WARNING: sv_diachronic is not a known dataset. Examining the data to choose which xpos vocab to use
196
+ 2025-12-03 15:21:24 INFO: Original length = 69070
197
+ 2025-12-03 15:21:25 INFO: Filtered length = 69070
198
+ 2025-12-03 15:21:40 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
199
+ 2025-12-03 15:21:49 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
200
+ 2025-12-03 15:22:04 DEBUG: 40474 batches created.
201
+ 2025-12-03 15:22:04 DEBUG: 9 batches created.
202
+ 2025-12-03 15:22:04 INFO: Training parser...
203
+ 2025-12-03 15:22:04 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
204
+ 2025-12-03 15:22:04 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
205
+ 2025-12-03 15:22:04 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
206
+ 2025-12-03 15:22:04 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
207
+ 2025-12-03 15:22:13 INFO: Finished STEP 20/50000, loss = 6.823766 (0.201 sec/batch), lr: 0.003000
208
+ 2025-12-03 15:22:17 INFO: Finished STEP 40/50000, loss = 5.197714 (0.167 sec/batch), lr: 0.003000
209
+ 2025-12-03 15:22:20 INFO: Finished STEP 60/50000, loss = 4.833910 (0.151 sec/batch), lr: 0.003000
210
+ 2025-12-03 15:22:23 INFO: Finished STEP 80/50000, loss = 4.588304 (0.144 sec/batch), lr: 0.003000
211
+ 2025-12-03 15:22:26 INFO: Finished STEP 100/50000, loss = 4.913873 (0.136 sec/batch), lr: 0.003000
212
+ 2025-12-03 15:22:26 INFO: Evaluating on dev set...
213
+ 2025-12-03 15:22:26 INFO: LAS MLAS BLEX
214
+ 2025-12-03 15:22:26 INFO: 38.37 22.69 25.63
215
+ 2025-12-03 15:22:27 INFO: step 100: train_loss = 8.803269, dev_score = 0.3837
216
+ 2025-12-03 15:22:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
217
+ 2025-12-03 15:22:27 INFO: new best model saved.
218
+ 2025-12-03 15:22:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
219
+ 2025-12-03 15:22:27 INFO: new model checkpoint saved.
220
+ 2025-12-03 15:22:30 INFO: Finished STEP 120/50000, loss = 4.934840 (0.121 sec/batch), lr: 0.003000
221
+ 2025-12-03 15:22:33 INFO: Finished STEP 140/50000, loss = 4.262431 (0.136 sec/batch), lr: 0.003000
222
+ 2025-12-03 15:22:35 INFO: Finished STEP 160/50000, loss = 4.450234 (0.137 sec/batch), lr: 0.003000
223
+ 2025-12-03 15:22:38 INFO: Finished STEP 180/50000, loss = 5.214414 (0.110 sec/batch), lr: 0.003000
224
+ 2025-12-03 15:22:40 INFO: Finished STEP 200/50000, loss = 4.563318 (0.125 sec/batch), lr: 0.003000
225
+ 2025-12-03 15:22:40 INFO: Evaluating on dev set...
226
+ 2025-12-03 15:22:41 INFO: LAS MLAS BLEX
227
+ 2025-12-03 15:22:41 INFO: 47.77 34.31 38.49
228
+ 2025-12-03 15:22:41 INFO: step 200: train_loss = 4.948097, dev_score = 0.4777
229
+ 2025-12-03 15:22:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
230
+ 2025-12-03 15:22:41 INFO: new best model saved.
231
+ 2025-12-03 15:22:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
232
+ 2025-12-03 15:22:42 INFO: new model checkpoint saved.
233
+ 2025-12-03 15:22:44 INFO: Finished STEP 220/50000, loss = 4.435537 (0.124 sec/batch), lr: 0.003000
234
+ 2025-12-03 15:22:47 INFO: Finished STEP 240/50000, loss = 4.125856 (0.105 sec/batch), lr: 0.003000
235
+ 2025-12-03 15:22:49 INFO: Finished STEP 260/50000, loss = 4.996243 (0.118 sec/batch), lr: 0.003000
236
+ 2025-12-03 15:22:52 INFO: Finished STEP 280/50000, loss = 3.675077 (0.114 sec/batch), lr: 0.003000
237
+ 2025-12-03 15:22:54 INFO: Finished STEP 300/50000, loss = 4.183073 (0.111 sec/batch), lr: 0.003000
238
+ 2025-12-03 15:22:54 INFO: Evaluating on dev set...
239
+ 2025-12-03 15:22:54 INFO: LAS MLAS BLEX
240
+ 2025-12-03 15:22:54 INFO: 50.74 39.50 42.02
241
+ 2025-12-03 15:22:54 INFO: step 300: train_loss = 4.627439, dev_score = 0.5074
242
+ 2025-12-03 15:22:55 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
243
+ 2025-12-03 15:22:55 INFO: new best model saved.
244
+ 2025-12-03 15:22:55 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
245
+ 2025-12-03 15:22:55 INFO: new model checkpoint saved.
246
+ 2025-12-03 15:22:57 INFO: Finished STEP 320/50000, loss = 4.655065 (0.117 sec/batch), lr: 0.003000
247
+ 2025-12-03 15:23:00 INFO: Finished STEP 340/50000, loss = 3.504390 (0.103 sec/batch), lr: 0.003000
248
+ 2025-12-03 15:23:02 INFO: Finished STEP 360/50000, loss = 3.996336 (0.116 sec/batch), lr: 0.003000
249
+ 2025-12-03 15:23:04 INFO: Finished STEP 380/50000, loss = 4.730554 (0.100 sec/batch), lr: 0.003000
250
+ 2025-12-03 15:23:06 INFO: Finished STEP 400/50000, loss = 4.109470 (0.108 sec/batch), lr: 0.003000
251
+ 2025-12-03 15:23:06 INFO: Evaluating on dev set...
252
+ 2025-12-03 15:23:07 INFO: LAS MLAS BLEX
253
+ 2025-12-03 15:23:07 INFO: 50.00 37.32 44.86
254
+ 2025-12-03 15:23:07 INFO: step 400: train_loss = 4.481576, dev_score = 0.5000
255
+ 2025-12-03 15:23:07 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
256
+ 2025-12-03 15:23:07 INFO: new model checkpoint saved.
257
+ 2025-12-03 15:23:10 INFO: Finished STEP 420/50000, loss = 3.610042 (0.114 sec/batch), lr: 0.003000
258
+ 2025-12-03 15:23:12 INFO: Finished STEP 440/50000, loss = 4.877528 (0.099 sec/batch), lr: 0.003000
259
+ 2025-12-03 15:23:14 INFO: Finished STEP 460/50000, loss = 5.707459 (0.112 sec/batch), lr: 0.003000
260
+ 2025-12-03 15:23:16 INFO: Finished STEP 480/50000, loss = 3.516663 (0.093 sec/batch), lr: 0.003000
261
+ 2025-12-03 15:23:18 INFO: Finished STEP 500/50000, loss = 4.809783 (0.094 sec/batch), lr: 0.003000
262
+ 2025-12-03 15:23:18 INFO: Evaluating on dev set...
263
+ 2025-12-03 15:23:18 INFO: LAS MLAS BLEX
264
+ 2025-12-03 15:23:18 INFO: 55.20 44.86 49.06
265
+ 2025-12-03 15:23:18 INFO: step 500: train_loss = 4.382052, dev_score = 0.5520
266
+ 2025-12-03 15:23:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
267
+ 2025-12-03 15:23:19 INFO: new best model saved.
268
+ 2025-12-03 15:23:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
269
+ 2025-12-03 15:23:19 INFO: new model checkpoint saved.
270
+ 2025-12-03 15:23:21 INFO: Finished STEP 520/50000, loss = 3.225300 (0.095 sec/batch), lr: 0.003000
271
+ 2025-12-03 15:23:23 INFO: Finished STEP 540/50000, loss = 4.299254 (0.104 sec/batch), lr: 0.003000
272
+ 2025-12-03 15:23:25 INFO: Finished STEP 560/50000, loss = 4.103633 (0.099 sec/batch), lr: 0.003000
273
+ 2025-12-03 15:23:27 INFO: Finished STEP 580/50000, loss = 2.937032 (0.097 sec/batch), lr: 0.003000
274
+ 2025-12-03 15:23:29 INFO: Finished STEP 600/50000, loss = 4.775133 (0.085 sec/batch), lr: 0.003000
275
+ 2025-12-03 15:23:29 INFO: Evaluating on dev set...
276
+ 2025-12-03 15:23:30 INFO: LAS MLAS BLEX
277
+ 2025-12-03 15:23:30 INFO: 55.69 46.41 49.37
278
+ 2025-12-03 15:23:30 INFO: step 600: train_loss = 4.127119, dev_score = 0.5569
279
+ 2025-12-03 15:23:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
280
+ 2025-12-03 15:23:30 INFO: new best model saved.
281
+ 2025-12-03 15:23:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
282
+ 2025-12-03 15:23:31 INFO: new model checkpoint saved.
283
+ 2025-12-03 15:23:33 INFO: Finished STEP 620/50000, loss = 4.320338 (0.098 sec/batch), lr: 0.003000
284
+ 2025-12-03 15:23:35 INFO: Finished STEP 640/50000, loss = 4.328771 (0.093 sec/batch), lr: 0.003000
285
+ 2025-12-03 15:23:37 INFO: Finished STEP 660/50000, loss = 4.174691 (0.099 sec/batch), lr: 0.003000
286
+ 2025-12-03 15:23:39 INFO: Finished STEP 680/50000, loss = 4.365898 (0.095 sec/batch), lr: 0.003000
287
+ 2025-12-03 15:23:41 INFO: Finished STEP 700/50000, loss = 4.004156 (0.101 sec/batch), lr: 0.003000
288
+ 2025-12-03 15:23:41 INFO: Evaluating on dev set...
289
+ 2025-12-03 15:23:41 INFO: LAS MLAS BLEX
290
+ 2025-12-03 15:23:41 INFO: 58.91 49.90 53.67
291
+ 2025-12-03 15:23:41 INFO: step 700: train_loss = 4.123694, dev_score = 0.5891
292
+ 2025-12-03 15:23:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
293
+ 2025-12-03 15:23:41 INFO: new best model saved.
294
+ 2025-12-03 15:23:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
295
+ 2025-12-03 15:23:42 INFO: new model checkpoint saved.
296
+ 2025-12-03 15:23:44 INFO: Finished STEP 720/50000, loss = 3.509187 (0.094 sec/batch), lr: 0.003000
297
+ 2025-12-03 15:23:46 INFO: Finished STEP 740/50000, loss = 3.501531 (0.087 sec/batch), lr: 0.003000
298
+ 2025-12-03 15:23:48 INFO: Finished STEP 760/50000, loss = 7.646075 (0.104 sec/batch), lr: 0.003000
299
+ 2025-12-03 15:23:50 INFO: Finished STEP 780/50000, loss = 4.090045 (0.093 sec/batch), lr: 0.003000
300
+ 2025-12-03 15:23:52 INFO: Finished STEP 800/50000, loss = 4.294971 (0.089 sec/batch), lr: 0.003000
301
+ 2025-12-03 15:23:52 INFO: Evaluating on dev set...
302
+ 2025-12-03 15:23:52 INFO: LAS MLAS BLEX
303
+ 2025-12-03 15:23:52 INFO: 54.95 45.47 49.68
304
+ 2025-12-03 15:23:52 INFO: step 800: train_loss = 3.999161, dev_score = 0.5495
305
+ 2025-12-03 15:23:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
306
+ 2025-12-03 15:23:53 INFO: new model checkpoint saved.
307
+ 2025-12-03 15:23:54 INFO: Finished STEP 820/50000, loss = 5.712487 (0.095 sec/batch), lr: 0.003000
308
+ 2025-12-03 15:23:56 INFO: Finished STEP 840/50000, loss = 2.782573 (0.096 sec/batch), lr: 0.003000
309
+ 2025-12-03 15:23:58 INFO: Finished STEP 860/50000, loss = 7.300400 (0.094 sec/batch), lr: 0.003000
310
+ 2025-12-03 15:24:00 INFO: Finished STEP 880/50000, loss = 4.767128 (0.090 sec/batch), lr: 0.003000
311
+ 2025-12-03 15:24:02 INFO: Finished STEP 900/50000, loss = 3.495617 (0.098 sec/batch), lr: 0.003000
312
+ 2025-12-03 15:24:02 INFO: Evaluating on dev set...
313
+ 2025-12-03 15:24:02 INFO: LAS MLAS BLEX
314
+ 2025-12-03 15:24:02 INFO: 59.16 48.41 52.65
315
+ 2025-12-03 15:24:02 INFO: step 900: train_loss = 4.125686, dev_score = 0.5916
316
+ 2025-12-03 15:24:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
317
+ 2025-12-03 15:24:03 INFO: new best model saved.
318
+ 2025-12-03 15:24:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
319
+ 2025-12-03 15:24:03 INFO: new model checkpoint saved.
320
+ 2025-12-03 15:24:05 INFO: Finished STEP 920/50000, loss = 4.070043 (0.092 sec/batch), lr: 0.003000
321
+ 2025-12-03 15:24:07 INFO: Finished STEP 940/50000, loss = 3.088636 (0.089 sec/batch), lr: 0.003000
322
+ 2025-12-03 15:24:09 INFO: Finished STEP 960/50000, loss = 3.278060 (0.089 sec/batch), lr: 0.003000
323
+ 2025-12-03 15:24:11 INFO: Finished STEP 980/50000, loss = 4.636242 (0.086 sec/batch), lr: 0.003000
324
+ 2025-12-03 15:24:12 INFO: Finished STEP 1000/50000, loss = 4.037950 (0.087 sec/batch), lr: 0.003000
325
+ 2025-12-03 15:24:12 INFO: Evaluating on dev set...
326
+ 2025-12-03 15:24:13 INFO: LAS MLAS BLEX
327
+ 2025-12-03 15:24:13 INFO: 57.92 47.16 50.95
328
+ 2025-12-03 15:24:13 INFO: step 1000: train_loss = 4.166898, dev_score = 0.5792
329
+ 2025-12-03 15:24:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
330
+ 2025-12-03 15:24:14 INFO: new model checkpoint saved.
331
+ 2025-12-03 15:24:15 INFO: Finished STEP 1020/50000, loss = 3.502761 (0.086 sec/batch), lr: 0.003000
332
+ 2025-12-03 15:24:17 INFO: Finished STEP 1040/50000, loss = 3.102474 (0.086 sec/batch), lr: 0.003000
333
+ 2025-12-03 15:24:19 INFO: Finished STEP 1060/50000, loss = 4.179004 (0.083 sec/batch), lr: 0.003000
334
+ 2025-12-03 15:24:21 INFO: Finished STEP 1080/50000, loss = 2.710993 (0.084 sec/batch), lr: 0.003000
335
+ 2025-12-03 15:24:22 INFO: Finished STEP 1100/50000, loss = 3.835454 (0.085 sec/batch), lr: 0.003000
336
+ 2025-12-03 15:24:22 INFO: Evaluating on dev set...
337
+ 2025-12-03 15:24:23 INFO: LAS MLAS BLEX
338
+ 2025-12-03 15:24:23 INFO: 56.44 48.42 51.79
339
+ 2025-12-03 15:24:23 INFO: step 1100: train_loss = 3.991774, dev_score = 0.5644
340
+ 2025-12-03 15:24:23 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
341
+ 2025-12-03 15:24:23 INFO: new model checkpoint saved.
342
+ 2025-12-03 15:24:25 INFO: Finished STEP 1120/50000, loss = 3.077247 (0.078 sec/batch), lr: 0.003000
343
+ 2025-12-03 15:24:27 INFO: Finished STEP 1140/50000, loss = 3.339320 (0.083 sec/batch), lr: 0.003000
344
+ 2025-12-03 15:24:29 INFO: Finished STEP 1160/50000, loss = 3.540397 (0.097 sec/batch), lr: 0.003000
345
+ 2025-12-03 15:24:30 INFO: Finished STEP 1180/50000, loss = 3.843534 (0.087 sec/batch), lr: 0.003000
346
+ 2025-12-03 15:24:32 INFO: Finished STEP 1200/50000, loss = 4.325091 (0.087 sec/batch), lr: 0.003000
347
+ 2025-12-03 15:24:32 INFO: Evaluating on dev set...
348
+ 2025-12-03 15:24:33 INFO: LAS MLAS BLEX
349
+ 2025-12-03 15:24:33 INFO: 59.65 49.37 53.59
350
+ 2025-12-03 15:24:33 INFO: step 1200: train_loss = 3.908015, dev_score = 0.5965
351
+ 2025-12-03 15:24:33 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
352
+ 2025-12-03 15:24:33 INFO: new best model saved.
353
+ 2025-12-03 15:24:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
354
+ 2025-12-03 15:24:34 INFO: new model checkpoint saved.
355
+ 2025-12-03 15:24:35 INFO: Finished STEP 1220/50000, loss = 4.003079 (0.085 sec/batch), lr: 0.003000
356
+ 2025-12-03 15:24:37 INFO: Finished STEP 1240/50000, loss = 3.405023 (0.079 sec/batch), lr: 0.003000
357
+ 2025-12-03 15:24:39 INFO: Finished STEP 1260/50000, loss = 3.795089 (0.082 sec/batch), lr: 0.003000
358
+ 2025-12-03 15:24:40 INFO: Finished STEP 1280/50000, loss = 3.564522 (0.087 sec/batch), lr: 0.003000
359
+ 2025-12-03 15:24:42 INFO: Finished STEP 1300/50000, loss = 6.505885 (0.084 sec/batch), lr: 0.003000
360
+ 2025-12-03 15:24:42 INFO: Evaluating on dev set...
361
+ 2025-12-03 15:24:43 INFO: LAS MLAS BLEX
362
+ 2025-12-03 15:24:43 INFO: 57.67 48.10 52.32
363
+ 2025-12-03 15:24:43 INFO: step 1300: train_loss = 3.923657, dev_score = 0.5767
364
+ 2025-12-03 15:24:43 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
365
+ 2025-12-03 15:24:43 INFO: new model checkpoint saved.
366
+ 2025-12-03 15:24:45 INFO: Finished STEP 1320/50000, loss = 3.060607 (0.088 sec/batch), lr: 0.003000
367
+ 2025-12-03 15:24:47 INFO: Finished STEP 1340/50000, loss = 3.442186 (0.077 sec/batch), lr: 0.003000
368
+ 2025-12-03 15:24:48 INFO: Finished STEP 1360/50000, loss = 4.950190 (0.086 sec/batch), lr: 0.003000
369
+ 2025-12-03 15:24:50 INFO: Finished STEP 1380/50000, loss = 6.227818 (0.083 sec/batch), lr: 0.003000
370
+ 2025-12-03 15:24:52 INFO: Finished STEP 1400/50000, loss = 3.698903 (0.086 sec/batch), lr: 0.003000
371
+ 2025-12-03 15:24:52 INFO: Evaluating on dev set...
372
+ 2025-12-03 15:24:52 INFO: LAS MLAS BLEX
373
+ 2025-12-03 15:24:52 INFO: 59.65 50.11 52.63
374
+ 2025-12-03 15:24:52 INFO: step 1400: train_loss = 3.958746, dev_score = 0.5965
375
+ 2025-12-03 15:24:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
376
+ 2025-12-03 15:24:53 INFO: new best model saved.
377
+ 2025-12-03 15:24:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
378
+ 2025-12-03 15:24:53 INFO: new model checkpoint saved.
379
+ 2025-12-03 15:24:55 INFO: Finished STEP 1420/50000, loss = 3.003988 (0.081 sec/batch), lr: 0.003000
380
+ 2025-12-03 15:24:56 INFO: Finished STEP 1440/50000, loss = 3.274803 (0.082 sec/batch), lr: 0.003000
381
+ 2025-12-03 15:24:58 INFO: Finished STEP 1460/50000, loss = 3.585543 (0.086 sec/batch), lr: 0.003000
382
+ 2025-12-03 15:25:00 INFO: Finished STEP 1480/50000, loss = 3.609994 (0.075 sec/batch), lr: 0.003000
383
+ 2025-12-03 15:25:01 INFO: Finished STEP 1500/50000, loss = 3.533881 (0.080 sec/batch), lr: 0.003000
384
+ 2025-12-03 15:25:01 INFO: Evaluating on dev set...
385
+ 2025-12-03 15:25:02 INFO: LAS MLAS BLEX
386
+ 2025-12-03 15:25:02 INFO: 57.67 47.68 51.48
387
+ 2025-12-03 15:25:02 INFO: step 1500: train_loss = 3.899388, dev_score = 0.5767
388
+ 2025-12-03 15:25:02 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
389
+ 2025-12-03 15:25:02 INFO: new model checkpoint saved.
390
+ 2025-12-03 15:25:04 INFO: Finished STEP 1520/50000, loss = 3.373524 (0.078 sec/batch), lr: 0.003000
391
+ 2025-12-03 15:25:06 INFO: Finished STEP 1540/50000, loss = 4.033227 (0.086 sec/batch), lr: 0.003000
392
+ 2025-12-03 15:25:07 INFO: Finished STEP 1560/50000, loss = 4.103477 (0.078 sec/batch), lr: 0.003000
393
+ 2025-12-03 15:25:09 INFO: Finished STEP 1580/50000, loss = 2.717227 (0.075 sec/batch), lr: 0.003000
394
+ 2025-12-03 15:25:11 INFO: Finished STEP 1600/50000, loss = 4.718042 (0.085 sec/batch), lr: 0.003000
395
+ 2025-12-03 15:25:11 INFO: Evaluating on dev set...
396
+ 2025-12-03 15:25:11 INFO: LAS MLAS BLEX
397
+ 2025-12-03 15:25:11 INFO: 57.92 49.89 53.70
398
+ 2025-12-03 15:25:11 INFO: step 1600: train_loss = 4.067216, dev_score = 0.5792
399
+ 2025-12-03 15:25:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
400
+ 2025-12-03 15:25:12 INFO: new model checkpoint saved.
401
+ 2025-12-03 15:25:13 INFO: Finished STEP 1620/50000, loss = 4.821809 (0.072 sec/batch), lr: 0.003000
402
+ 2025-12-03 15:25:15 INFO: Finished STEP 1640/50000, loss = 3.207412 (0.076 sec/batch), lr: 0.003000
403
+ 2025-12-03 15:25:16 INFO: Finished STEP 1660/50000, loss = 3.344615 (0.086 sec/batch), lr: 0.003000
404
+ 2025-12-03 15:25:18 INFO: Finished STEP 1680/50000, loss = 2.665408 (0.080 sec/batch), lr: 0.003000
405
+ 2025-12-03 15:25:20 INFO: Finished STEP 1700/50000, loss = 4.345899 (0.081 sec/batch), lr: 0.003000
406
+ 2025-12-03 15:25:20 INFO: Evaluating on dev set...
407
+ 2025-12-03 15:25:20 INFO: LAS MLAS BLEX
408
+ 2025-12-03 15:25:20 INFO: 56.68 50.31 52.41
409
+ 2025-12-03 15:25:20 INFO: step 1700: train_loss = 4.037608, dev_score = 0.5668
410
+ 2025-12-03 15:25:21 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
411
+ 2025-12-03 15:25:21 INFO: new model checkpoint saved.
412
+ 2025-12-03 15:25:22 INFO: Finished STEP 1720/50000, loss = 2.968188 (0.081 sec/batch), lr: 0.003000
413
+ 2025-12-03 15:25:24 INFO: Finished STEP 1740/50000, loss = 3.712258 (0.075 sec/batch), lr: 0.003000
414
+ 2025-12-03 15:25:26 INFO: Finished STEP 1760/50000, loss = 5.289080 (0.081 sec/batch), lr: 0.003000
415
+ 2025-12-03 15:25:27 INFO: Finished STEP 1780/50000, loss = 4.869098 (0.079 sec/batch), lr: 0.003000
416
+ 2025-12-03 15:25:29 INFO: Finished STEP 1800/50000, loss = 2.825555 (0.080 sec/batch), lr: 0.003000
417
+ 2025-12-03 15:25:29 INFO: Evaluating on dev set...
418
+ 2025-12-03 15:25:29 INFO: LAS MLAS BLEX
419
+ 2025-12-03 15:25:29 INFO: 60.40 49.68 54.32
420
+ 2025-12-03 15:25:29 INFO: step 1800: train_loss = 4.019084, dev_score = 0.6040
421
+ 2025-12-03 15:25:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
422
+ 2025-12-03 15:25:30 INFO: new best model saved.
423
+ 2025-12-03 15:25:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
424
+ 2025-12-03 15:25:30 INFO: new model checkpoint saved.
425
+ 2025-12-03 15:25:32 INFO: Finished STEP 1820/50000, loss = 3.218971 (0.074 sec/batch), lr: 0.003000
426
+ 2025-12-03 15:25:33 INFO: Finished STEP 1840/50000, loss = 4.086784 (0.080 sec/batch), lr: 0.003000
427
+ 2025-12-03 15:25:35 INFO: Finished STEP 1860/50000, loss = 3.021889 (0.076 sec/batch), lr: 0.003000
428
+ 2025-12-03 15:25:36 INFO: Finished STEP 1880/50000, loss = 5.154871 (0.082 sec/batch), lr: 0.003000
429
+ 2025-12-03 15:25:38 INFO: Finished STEP 1900/50000, loss = 3.693562 (0.068 sec/batch), lr: 0.003000
430
+ 2025-12-03 15:25:38 INFO: Evaluating on dev set...
431
+ 2025-12-03 15:25:38 INFO: LAS MLAS BLEX
432
+ 2025-12-03 15:25:38 INFO: 59.65 49.37 53.59
433
+ 2025-12-03 15:25:38 INFO: step 1900: train_loss = 4.130100, dev_score = 0.5965
434
+ 2025-12-03 15:25:39 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
435
+ 2025-12-03 15:25:39 INFO: new model checkpoint saved.
436
+ 2025-12-03 15:25:41 INFO: Finished STEP 1920/50000, loss = 4.085068 (0.077 sec/batch), lr: 0.003000
437
+ 2025-12-03 15:25:42 INFO: Finished STEP 1940/50000, loss = 3.556464 (0.078 sec/batch), lr: 0.003000
438
+ 2025-12-03 15:25:44 INFO: Finished STEP 1960/50000, loss = 3.946025 (0.077 sec/batch), lr: 0.003000
439
+ 2025-12-03 15:25:45 INFO: Finished STEP 1980/50000, loss = 2.063778 (0.077 sec/batch), lr: 0.003000
440
+ 2025-12-03 15:25:47 INFO: Finished STEP 2000/50000, loss = 2.791371 (0.079 sec/batch), lr: 0.003000
441
+ 2025-12-03 15:25:47 INFO: Evaluating on dev set...
442
+ 2025-12-03 15:25:47 INFO: LAS MLAS BLEX
443
+ 2025-12-03 15:25:47 INFO: 57.43 45.40 48.39
444
+ 2025-12-03 15:25:47 INFO: step 2000: train_loss = 4.043870, dev_score = 0.5743
445
+ 2025-12-03 15:25:48 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
446
+ 2025-12-03 15:25:48 INFO: new model checkpoint saved.
447
+ 2025-12-03 15:25:50 INFO: Finished STEP 2020/50000, loss = 3.088628 (0.076 sec/batch), lr: 0.003000
448
+ 2025-12-03 15:25:51 INFO: Finished STEP 2040/50000, loss = 2.854034 (0.071 sec/batch), lr: 0.003000
449
+ 2025-12-03 15:25:53 INFO: Finished STEP 2060/50000, loss = 5.301508 (0.079 sec/batch), lr: 0.003000
450
+ 2025-12-03 15:25:54 INFO: Finished STEP 2080/50000, loss = 4.484879 (0.071 sec/batch), lr: 0.003000
451
+ 2025-12-03 15:25:56 INFO: Finished STEP 2100/50000, loss = 3.586477 (0.085 sec/batch), lr: 0.003000
452
+ 2025-12-03 15:25:56 INFO: Evaluating on dev set...
453
+ 2025-12-03 15:25:56 INFO: LAS MLAS BLEX
454
+ 2025-12-03 15:25:56 INFO: 58.66 48.83 51.80
455
+ 2025-12-03 15:25:56 INFO: step 2100: train_loss = 4.001230, dev_score = 0.5866
456
+ 2025-12-03 15:25:57 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
457
+ 2025-12-03 15:25:57 INFO: new model checkpoint saved.
458
+ 2025-12-03 15:25:58 INFO: Finished STEP 2120/50000, loss = 3.297551 (0.073 sec/batch), lr: 0.003000
459
+ 2025-12-03 15:26:00 INFO: Finished STEP 2140/50000, loss = 3.914132 (0.067 sec/batch), lr: 0.003000
460
+ 2025-12-03 15:26:01 INFO: Finished STEP 2160/50000, loss = 3.974120 (0.074 sec/batch), lr: 0.003000
461
+ 2025-12-03 15:26:03 INFO: Finished STEP 2180/50000, loss = 3.764951 (0.069 sec/batch), lr: 0.003000
462
+ 2025-12-03 15:26:04 INFO: Finished STEP 2200/50000, loss = 5.886549 (0.082 sec/batch), lr: 0.003000
463
+ 2025-12-03 15:26:04 INFO: Evaluating on dev set...
464
+ 2025-12-03 15:26:05 INFO: LAS MLAS BLEX
465
+ 2025-12-03 15:26:05 INFO: 59.16 50.63 54.43
466
+ 2025-12-03 15:26:05 INFO: step 2200: train_loss = 4.143704, dev_score = 0.5916
467
+ 2025-12-03 15:26:06 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
468
+ 2025-12-03 15:26:06 INFO: new model checkpoint saved.
469
+ 2025-12-03 15:26:07 INFO: Finished STEP 2220/50000, loss = 4.426775 (0.067 sec/batch), lr: 0.003000
470
+ 2025-12-03 15:26:09 INFO: Finished STEP 2240/50000, loss = 5.679852 (0.086 sec/batch), lr: 0.003000
471
+ 2025-12-03 15:26:10 INFO: Finished STEP 2260/50000, loss = 2.991968 (0.072 sec/batch), lr: 0.003000
472
+ 2025-12-03 15:26:11 INFO: Finished STEP 2280/50000, loss = 3.683204 (0.077 sec/batch), lr: 0.003000
473
+ 2025-12-03 15:26:13 INFO: Finished STEP 2300/50000, loss = 5.155180 (0.077 sec/batch), lr: 0.003000
474
+ 2025-12-03 15:26:13 INFO: Evaluating on dev set...
475
+ 2025-12-03 15:26:14 INFO: LAS MLAS BLEX
476
+ 2025-12-03 15:26:14 INFO: 58.66 47.48 50.84
477
+ 2025-12-03 15:26:14 INFO: step 2300: train_loss = 3.961017, dev_score = 0.5866
478
+ 2025-12-03 15:26:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
479
+ 2025-12-03 15:26:14 INFO: new model checkpoint saved.
480
+ 2025-12-03 15:26:16 INFO: Finished STEP 2320/50000, loss = 4.828197 (0.071 sec/batch), lr: 0.003000
481
+ 2025-12-03 15:26:17 INFO: Finished STEP 2340/50000, loss = 3.383664 (0.076 sec/batch), lr: 0.003000
482
+ 2025-12-03 15:26:19 INFO: Finished STEP 2360/50000, loss = 3.122447 (0.073 sec/batch), lr: 0.003000
483
+ 2025-12-03 15:26:20 INFO: Finished STEP 2380/50000, loss = 2.817877 (0.069 sec/batch), lr: 0.003000
484
+ 2025-12-03 15:26:22 INFO: Finished STEP 2400/50000, loss = 4.030520 (0.071 sec/batch), lr: 0.003000
485
+ 2025-12-03 15:26:22 INFO: Evaluating on dev set...
486
+ 2025-12-03 15:26:22 INFO: LAS MLAS BLEX
487
+ 2025-12-03 15:26:22 INFO: 59.90 49.26 53.05
488
+ 2025-12-03 15:26:22 INFO: step 2400: train_loss = 4.231360, dev_score = 0.5990
489
+ 2025-12-03 15:26:23 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
490
+ 2025-12-03 15:26:23 INFO: new model checkpoint saved.
491
+ 2025-12-03 15:26:24 INFO: Finished STEP 2420/50000, loss = 3.418092 (0.072 sec/batch), lr: 0.003000
492
+ 2025-12-03 15:26:26 INFO: Finished STEP 2440/50000, loss = 3.182749 (0.077 sec/batch), lr: 0.003000
493
+ 2025-12-03 15:26:27 INFO: Finished STEP 2460/50000, loss = 4.582247 (0.077 sec/batch), lr: 0.003000
494
+ 2025-12-03 15:26:29 INFO: Finished STEP 2480/50000, loss = 2.969922 (0.078 sec/batch), lr: 0.003000
495
+ 2025-12-03 15:26:30 INFO: Finished STEP 2500/50000, loss = 3.402813 (0.072 sec/batch), lr: 0.003000
496
+ 2025-12-03 15:26:30 INFO: Evaluating on dev set...
497
+ 2025-12-03 15:26:31 INFO: LAS MLAS BLEX
498
+ 2025-12-03 15:26:31 INFO: 56.44 46.71 50.96
499
+ 2025-12-03 15:26:31 INFO: step 2500: train_loss = 3.973895, dev_score = 0.5644
500
+ 2025-12-03 15:26:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
501
+ 2025-12-03 15:26:31 INFO: new model checkpoint saved.
502
+ 2025-12-03 15:26:33 INFO: Finished STEP 2520/50000, loss = 2.656119 (0.074 sec/batch), lr: 0.003000
503
+ 2025-12-03 15:26:34 INFO: Finished STEP 2540/50000, loss = 2.363447 (0.069 sec/batch), lr: 0.003000
504
+ 2025-12-03 15:26:36 INFO: Finished STEP 2560/50000, loss = 3.910929 (0.077 sec/batch), lr: 0.003000
505
+ 2025-12-03 15:26:37 INFO: Finished STEP 2580/50000, loss = 2.956022 (0.070 sec/batch), lr: 0.003000
506
+ 2025-12-03 15:26:39 INFO: Finished STEP 2600/50000, loss = 3.604472 (0.075 sec/batch), lr: 0.003000
507
+ 2025-12-03 15:26:39 INFO: Evaluating on dev set...
508
+ 2025-12-03 15:26:39 INFO: LAS MLAS BLEX
509
+ 2025-12-03 15:26:39 INFO: 57.67 50.32 53.28
510
+ 2025-12-03 15:26:39 INFO: step 2600: train_loss = 4.207645, dev_score = 0.5767
511
+ 2025-12-03 15:26:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
512
+ 2025-12-03 15:26:40 INFO: new model checkpoint saved.
513
+ 2025-12-03 15:26:41 INFO: Finished STEP 2620/50000, loss = 4.077714 (0.081 sec/batch), lr: 0.003000
514
+ 2025-12-03 15:26:43 INFO: Finished STEP 2640/50000, loss = 2.716412 (0.073 sec/batch), lr: 0.003000
515
+ 2025-12-03 15:26:44 INFO: Finished STEP 2660/50000, loss = 4.703525 (0.068 sec/batch), lr: 0.003000
516
+ 2025-12-03 15:26:45 INFO: Finished STEP 2680/50000, loss = 2.921926 (0.080 sec/batch), lr: 0.003000
517
+ 2025-12-03 15:26:47 INFO: Finished STEP 2700/50000, loss = 2.392154 (0.074 sec/batch), lr: 0.003000
518
+ 2025-12-03 15:26:47 INFO: Evaluating on dev set...
519
+ 2025-12-03 15:26:47 INFO: LAS MLAS BLEX
520
+ 2025-12-03 15:26:47 INFO: 57.43 49.68 52.63
521
+ 2025-12-03 15:26:47 INFO: step 2700: train_loss = 4.078543, dev_score = 0.5743
522
+ 2025-12-03 15:26:48 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
523
+ 2025-12-03 15:26:48 INFO: new model checkpoint saved.
524
+ 2025-12-03 15:26:49 INFO: Finished STEP 2720/50000, loss = 3.484982 (0.072 sec/batch), lr: 0.003000
525
+ 2025-12-03 15:26:51 INFO: Finished STEP 2740/50000, loss = 5.759233 (0.076 sec/batch), lr: 0.003000
526
+ 2025-12-03 15:26:52 INFO: Finished STEP 2760/50000, loss = 3.679373 (0.074 sec/batch), lr: 0.003000
527
+ 2025-12-03 15:26:54 INFO: Finished STEP 2780/50000, loss = 2.794205 (0.075 sec/batch), lr: 0.003000
528
+ 2025-12-03 15:26:55 INFO: Finished STEP 2800/50000, loss = 5.015698 (0.069 sec/batch), lr: 0.003000
529
+ 2025-12-03 15:26:55 INFO: Evaluating on dev set...
530
+ 2025-12-03 15:26:56 INFO: LAS MLAS BLEX
531
+ 2025-12-03 15:26:56 INFO: 59.16 50.00 53.81
532
+ 2025-12-03 15:26:56 INFO: step 2800: train_loss = 4.112154, dev_score = 0.5916
533
+ 2025-12-03 15:26:56 INFO: Training ended with 2800 steps.
534
+ 2025-12-03 15:26:56 INFO: Best dev F1 = 60.40, at iteration = 1800
535
+ 2025-12-03 15:26:57 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--batch_size', '32', '--dropout', '0.33']
536
+ 2025-12-03 15:26:57 INFO: Running parser in predict mode
537
+ 2025-12-03 15:26:57 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
538
+ 2025-12-03 15:27:00 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
539
+ 2025-12-03 15:27:00 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
540
+ 2025-12-03 15:27:00 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
541
+ 2025-12-03 15:27:00 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
542
+ 2025-12-03 15:27:00 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
543
+ 2025-12-03 15:27:00 INFO: Loading data with batch size 32...
544
+ 2025-12-03 15:27:00 DEBUG: 9 batches created.
545
+ 2025-12-03 15:27:01 INFO: F1 scores for each dependency:
546
+ Note that unlabeled attachment errors hurt the labeled attachment scores
547
+ acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
548
+ acl:relcl: p 0.3333 r 0.2857 f1 0.3077 (7 actual)
549
+ advcl: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
550
+ advmod: p 0.4000 r 0.4800 f1 0.4364 (25 actual)
551
+ amod: p 0.7500 r 0.7742 f1 0.7619 (31 actual)
552
+ appos: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
553
+ aux: p 0.6667 r 0.7273 f1 0.6957 (11 actual)
554
+ case: p 0.9608 r 0.8750 f1 0.9159 (56 actual)
555
+ cc: p 0.6923 r 0.6923 f1 0.6923 (13 actual)
556
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
557
+ conj: p 0.2333 r 0.5833 f1 0.3333 (12 actual)
558
+ cop: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
559
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
560
+ det: p 0.8182 r 0.8182 f1 0.8182 (22 actual)
561
+ expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
562
+ iobj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
563
+ mark: p 0.5714 r 0.6667 f1 0.6154 (12 actual)
564
+ nmod: p 0.2632 r 0.3333 f1 0.2941 (15 actual)
565
+ nmod:poss: p 1.0000 r 0.8947 f1 0.9444 (19 actual)
566
+ nsubj: p 0.4583 r 0.6471 f1 0.5366 (17 actual)
567
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
568
+ obj: p 0.6667 r 0.8182 f1 0.7347 (22 actual)
569
+ obl: p 0.5682 r 0.6098 f1 0.5882 (41 actual)
570
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
571
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
572
+ parataxis: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
573
+ punct: p 0.4423 r 0.4423 f1 0.4423 (52 actual)
574
+ root: p 0.4444 r 0.4444 f1 0.4444 (9 actual)
575
+ xcomp: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
576
+ 2025-12-03 15:27:01 INFO: LAS MLAS BLEX
577
+ 2025-12-03 15:27:01 INFO: 60.40 49.68 54.32
578
+ 2025-12-03 15:27:01 INFO: Parser score:
579
+ 2025-12-03 15:27:01 INFO: sv_diachronic 60.40
580
+ 2025-12-03 15:27:01 INFO: Finished running dev set on
581
+ UD_Swedish-diachronic
582
+ UAS LAS CLAS MLAS BLEX
583
+ 68.07 60.40 54.32 49.68 54.32
584
+ 2025-12-03 15:27:01 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--batch_size', '32', '--dropout', '0.33']
585
+ 2025-12-03 15:27:01 INFO: Running parser in predict mode
586
+ 2025-12-03 15:27:01 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
587
+ 2025-12-03 15:27:04 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
588
+ 2025-12-03 15:27:04 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
589
+ 2025-12-03 15:27:04 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
590
+ 2025-12-03 15:27:04 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
591
+ 2025-12-03 15:27:04 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
592
+ 2025-12-03 15:27:04 INFO: Loading data with batch size 32...
593
+ 2025-12-03 15:27:04 DEBUG: 93 batches created.
594
+ 2025-12-03 15:27:08 INFO: F1 scores for each dependency:
595
+ Note that unlabeled attachment errors hurt the labeled attachment scores
596
+ acl: p 0.0000 r 0.0000 f1 0.0000 (32 actual)
597
+ acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
598
+ acl:relcl: p 0.2333 r 0.0933 f1 0.1333 (75 actual)
599
+ advcl: p 0.0217 r 0.0167 f1 0.0189 (60 actual)
600
+ advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
601
+ advmod: p 0.5836 r 0.5858 f1 0.5847 (268 actual)
602
+ amod: p 0.7099 r 0.8087 f1 0.7561 (230 actual)
603
+ appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
604
+ aux: p 0.7640 r 0.8095 f1 0.7861 (84 actual)
605
+ aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
606
+ case: p 0.8476 r 0.8204 f1 0.8338 (373 actual)
607
+ cc: p 0.6200 r 0.6000 f1 0.6098 (155 actual)
608
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
609
+ compound:prt: p 0.8333 r 0.7143 f1 0.7692 (21 actual)
610
+ conj: p 0.1505 r 0.2848 f1 0.1969 (158 actual)
611
+ cop: p 0.9130 r 0.4565 f1 0.6087 (46 actual)
612
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
613
+ dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
614
+ det: p 0.8075 r 0.8269 f1 0.8171 (208 actual)
615
+ discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
616
+ dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
617
+ expl: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
618
+ expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
619
+ fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
620
+ flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
621
+ flat:name: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
622
+ goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
623
+ iobj: p 0.0000 r 0.0000 f1 0.0000 (14 actual)
624
+ mark: p 0.6750 r 0.7059 f1 0.6901 (153 actual)
625
+ nmod: p 0.3509 r 0.3922 f1 0.3704 (102 actual)
626
+ nmod:poss: p 0.8141 r 0.8944 f1 0.8523 (142 actual)
627
+ nsubj: p 0.5340 r 0.6179 f1 0.5728 (280 actual)
628
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
629
+ nummod: p 0.0000 r 0.0000 f1 0.0000 (10 actual)
630
+ obj: p 0.5607 r 0.7322 f1 0.6351 (183 actual)
631
+ obl: p 0.5032 r 0.5576 f1 0.5290 (278 actual)
632
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
633
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
634
+ parataxis: p 0.0714 r 0.0556 f1 0.0625 (18 actual)
635
+ punct: p 0.4131 r 0.4141 f1 0.4136 (425 actual)
636
+ reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
637
+ root: p 0.5354 r 0.5354 f1 0.5354 (99 actual)
638
+ vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
639
+ xcomp: p 0.4444 r 0.1600 f1 0.2353 (75 actual)
640
+ 2025-12-03 15:27:08 INFO: LAS MLAS BLEX
641
+ 2025-12-03 15:27:08 INFO: 57.05 48.00 51.23
642
+ 2025-12-03 15:27:08 INFO: Parser score:
643
+ 2025-12-03 15:27:08 INFO: sv_diachronic 57.05
644
+ 2025-12-03 15:27:09 INFO: Finished running test set on
645
+ UD_Swedish-diachronic
646
+ UAS LAS CLAS MLAS BLEX
647
+ 66.39 57.05 51.23 48.00 51.23
648
+ DONE.
649
+ Full log saved to: logs/log_diachronic.pt_sv_diachron_de_lit_20251203_151554.txt
650
+ Symlink updated: logs/latest.txt → log_diachronic.pt_sv_diachron_de_lit_20251203_151554.txt
logs/log_diachronic.pt_sv_diachron_is_20251203_214751.txt ADDED
@@ -0,0 +1,757 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ === LOGFILE: logs/log_diachronic.pt_sv_diachron_is_20251203_214751.txt ===
2
+ Language codes: sv diachron is
3
+ Using pretrained model: diachronic.pt
4
+
5
+ Running: python prepare-train-val-test.py sv diachron is
6
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-dev.conllu
7
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test.conllu
8
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_pud-ud-test.conllu
9
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-test.conllu
10
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_swell-ud-test-trg.conllu
11
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-dev.conllu
12
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/ucxn_ud_swedish-talbanken.conllu
13
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_talbanken-ud-train.conllu
14
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_old-ud-test.conllu
15
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-train.conllu
16
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-sv/sv_lines-ud-test.conllu
17
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_gc-ud-dev.conllu
18
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_modern-ud-train.conllu
19
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_pud-ud-test.conllu
20
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_modern-ud-test.conllu
21
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_gc-ud-train.conllu
22
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_gc-ud-test.conllu
23
+ Reading: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud-treebanks-is/is_modern-ud-dev.conllu
24
+ Including DigPhil MACHINE in TRAIN (minus gold)…
25
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec330-GyllenborgC_SwenskaSpratthoken.conllu
26
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec254-CederborghF_BerattelseOmJohnHall.conllu
27
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec277-EnbomPU_MedborgeligtSkalde.conllu
28
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec268-DulciU_VitterhetsNojen3.conllu
29
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1063-spf220.conllu
30
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec397-AngeredStrandbergH_UnderSodernsSol.conllu
31
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec324-GranbergPA_Enslighetsalskaren.conllu
32
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec252-BremerF_Teckningar1.conllu
33
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec988-spf145.conllu
34
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec987-spf144.conllu
35
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec631-HasselskogN_HallaHallaGronkoping.conllu
36
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-letter141673-Stalhammar.conllu
37
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1033-spf190.conllu
38
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec25-Runius.conllu
39
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec486-SchwartzMS_BellmansSkor.conllu
40
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec452-NyblomH_FantasierFyra.conllu
41
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec613-EngstromA_StrindbergOchJag.conllu
42
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec208-Anonym_DetGrasligaMordet.conllu
43
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec639-HeidenstamV_Proletarfilosofiens.conllu
44
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec1102-spf259.conllu
45
+ Reading GOLD: /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/alanev_raw_files/diachron-validated/svediakorp-sec991-spf148.conllu
46
+ Cleaning TRAIN...
47
+ [REMOVED] sent_id=6 ERRORS=['Line 24: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 29: Invalid token ID or head', 'Token 30 has invalid head 24']
48
+ [REMOVED] sent_id=7_8 ERRORS=['Multiple roots found: [5, 10]']
49
+ [REMOVED] sent_id=30_31 ERRORS=['Multiple roots found: [3, 18]']
50
+ [REMOVED] sent_id=35 ERRORS=['Line 36: Invalid token ID or head']
51
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [1, 5]']
52
+ [REMOVED] sent_id=2_3 ERRORS=['Multiple roots found: [7, 20]']
53
+ [REMOVED] sent_id=8_9 ERRORS=['Multiple roots found: [24, 57]']
54
+ [REMOVED] sent_id=12_13 ERRORS=['Multiple roots found: [11, 16]']
55
+ [REMOVED] sent_id=124_split2 ERRORS=['Line 4: Invalid token ID or head', 'No root found', 'Token 1 has invalid head 4', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 6 has invalid head 4', 'Token 11 has invalid head 4', 'Token 15 has invalid head 4']
56
+ [REMOVED] sent_id=396 ERRORS=['Token 2: Missing form']
57
+ [REMOVED] sent_id=416 ERRORS=['Token 2: Missing form']
58
+ [REMOVED] sent_id=589 ERRORS=['Token 2: Missing form']
59
+ [REMOVED] sent_id=909 ERRORS=['Token 2: Missing form']
60
+ [REMOVED] sent_id=912 ERRORS=['Token 2: Missing form']
61
+ [REMOVED] sent_id=3_split1 ERRORS=['Multiple roots found: [4, 15, 17]']
62
+ [REMOVED] sent_id=3_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 8: Invalid token ID or head', 'Line 15: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1', 'Token 3 has invalid head 8', 'Token 4 has invalid head 8', 'Token 5 has invalid head 8', 'Token 7 has invalid head 8', 'Token 10 has invalid head 8', 'Token 13 has invalid head 8', 'Token 14 has invalid head 8']
63
+ [REMOVED] sent_id=3_4 ERRORS=['Multiple roots found: [1, 5]']
64
+ [REMOVED] sent_id=5_6 ERRORS=['Multiple roots found: [3, 24]']
65
+ [REMOVED] sent_id=11_12_13 ERRORS=['Multiple roots found: [5, 17, 25]']
66
+ [REMOVED] sent_id=119 ERRORS=['Token 2: Missing form']
67
+ [REMOVED] sent_id=179 ERRORS=['Token 2: Missing form']
68
+ [REMOVED] sent_id=188 ERRORS=['Token 2: Missing form']
69
+ [REMOVED] sent_id=223 ERRORS=['Token 2: Missing form']
70
+ [REMOVED] sent_id=268 ERRORS=['Token 2: Missing form']
71
+ [REMOVED] sent_id=325 ERRORS=['Token 2: Missing form']
72
+ [REMOVED] sent_id=388 ERRORS=['Token 2: Missing form']
73
+ [REMOVED] sent_id=399 ERRORS=['Token 2: Missing form']
74
+ [REMOVED] sent_id=475 ERRORS=['Token 2: Missing form']
75
+ [REMOVED] sent_id=505 ERRORS=['Token 2: Missing form']
76
+ [REMOVED] sent_id=520 ERRORS=['Token 2: Missing form']
77
+ [REMOVED] sent_id=562 ERRORS=['Token 2: Missing form']
78
+ [REMOVED] sent_id=669 ERRORS=['Token 2: Missing form']
79
+ [REMOVED] sent_id=711 ERRORS=['Token 2: Missing form']
80
+ [REMOVED] sent_id=731 ERRORS=['Token 2: Missing form']
81
+ [REMOVED] sent_id=867 ERRORS=['Token 2: Missing form']
82
+ [REMOVED] sent_id=884 ERRORS=['Token 2: Missing form']
83
+ [REMOVED] sent_id=923 ERRORS=['Token 2: Missing form']
84
+ [REMOVED] sent_id=939 ERRORS=['Token 2: Missing form']
85
+ [REMOVED] sent_id=1086 ERRORS=['Token 2: Missing form']
86
+ [REMOVED] sent_id=1179 ERRORS=['Token 2: Missing form']
87
+ [REMOVED] sent_id=1251 ERRORS=['Token 2: Missing form']
88
+ [REMOVED] sent_id=1345 ERRORS=['Token 2: Missing form']
89
+ [REMOVED] sent_id=1459 ERRORS=['Token 2: Missing form']
90
+ [REMOVED] sent_id=1656 ERRORS=['Token 2: Missing form']
91
+ [REMOVED] sent_id=1669 ERRORS=['Token 2: Missing form']
92
+ [REMOVED] sent_id=87_88 ERRORS=['Multiple roots found: [3, 6]']
93
+ [REMOVED] sent_id=65_split2_66_split2 ERRORS=['Line 4: Invalid token ID or head', 'Token 2 has invalid head 4', 'Token 3 has invalid head 4', 'Token 5 has invalid head 4']
94
+ [REMOVED] sent_id=25 ERRORS=['Token 2: Missing form']
95
+ [REMOVED] sent_id=136 ERRORS=['Token 2: Missing form']
96
+ [REMOVED] sent_id=208 ERRORS=['Token 2: Missing form']
97
+ [REMOVED] sent_id=230 ERRORS=['Token 2: Missing form']
98
+ [REMOVED] sent_id=245 ERRORS=['Token 2: Missing form']
99
+ [REMOVED] sent_id=276 ERRORS=['Token 2: Missing form']
100
+ [REMOVED] sent_id=320 ERRORS=['Token 2: Missing form']
101
+ [REMOVED] sent_id=366 ERRORS=['Token 2: Missing form']
102
+ [REMOVED] sent_id=519 ERRORS=['Token 2: Missing form']
103
+ [REMOVED] sent_id=569 ERRORS=['Token 2: Missing form']
104
+ [REMOVED] sent_id=50_split2 ERRORS=['Line 1: Invalid token ID or head', 'Line 6: Invalid token ID or head', 'No root found', 'Token 2 has invalid head 1']
105
+ [REMOVED] sent_id=53_54 ERRORS=['Multiple roots found: [27, 91]']
106
+ [REMOVED] sent_id=55_56_57 ERRORS=['Multiple roots found: [2, 4, 13]']
107
+ [REMOVED] sent_id=17_split1 ERRORS=['Multiple roots found: [2, 14, 17]']
108
+ [REMOVED] sent_id=17_split2 ERRORS=['Line 8: Invalid token ID or head', 'Line 25: Invalid token ID or head', 'Line 38: Invalid token ID or head', 'No root found', 'Token 3 has invalid head 8', 'Token 7 has invalid head 8', 'Token 9 has invalid head 8', 'Token 10 has invalid head 8', 'Token 17 has invalid head 8', 'Token 22 has invalid head 25', 'Token 23 has invalid head 25', 'Token 24 has invalid head 25', 'Token 26 has invalid head 25', 'Token 27 has invalid head 25', 'Token 28 has invalid head 25']
109
+ [REMOVED] sent_id=19_split1 ERRORS=['Multiple roots found: [3, 31]']
110
+ Cleaning DEV...
111
+ [REMOVED] sent_id=33 ERRORS=['Token 15: Missing deprel']
112
+ Cleaning TEST...
113
+ Writing TRAIN → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu (75787 valid sentences)
114
+ Writing DEV → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu (9 valid sentences)
115
+ Writing TEST → /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu (99 valid sentences)
116
+ Done.
117
+ Sourcing scripts/config_alvis.sh
118
+ Running stanza dataset preparation…
119
+ 2025-12-03 21:48:01 INFO: Datasets program called with:
120
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/datasets/prepare_depparse_treebank.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
121
+ 2025-12-03 21:48:01 DEBUG: Downloading resource file from https://raw.githubusercontent.com/stanfordnlp/stanza-resources/main/resources_1.11.0.json
122
+
123
+ 2025-12-03 21:48:01 INFO: Downloaded file to /cephyr/users/cleland/Alvis/stanza_resources/resources.json
124
+ 2025-12-03 21:48:01 DEBUG: Processing parameter "processors"...
125
+ 2025-12-03 21:48:01 WARNING: Can not find pos: diachronic from official model list. Ignoring it.
126
+ 2025-12-03 21:48:01 INFO: Downloading these customized packages for language: sv (Swedish)...
127
+ =======================
128
+ | Processor | Package |
129
+ -----------------------
130
+ =======================
131
+
132
+ 2025-12-03 21:48:01 INFO: Finished downloading models and saved to /cephyr/users/cleland/Alvis/stanza_resources
133
+ 2025-12-03 21:48:01 INFO: Using tagger model in /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt for sv_diachronic
134
+ 2025-12-03 21:48:01 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
135
+ 2025-12-03 21:48:01 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
136
+ Augmented 191 quotes: Counter({'„”': 29, '»«': 25, '「」': 25, '″″': 20, '«»': 17, '„“': 16, '””': 16, '""': 15, '《》': 15, '““': 13})
137
+ 2025-12-03 21:48:05 INFO: Running tagger to retag /local/tmp.5441282/tmp9_locfmj/sv_diachronic.train.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu
138
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmp9_locfmj/sv_diachronic.train.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu']
139
+ 2025-12-03 21:48:05 INFO: Running tagger in predict mode
140
+ 2025-12-03 21:48:05 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
141
+ 2025-12-03 21:48:08 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
142
+ 2025-12-03 21:48:08 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
143
+ 2025-12-03 21:48:08 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
144
+ 2025-12-03 21:48:08 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
145
+ 2025-12-03 21:48:08 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
146
+ 2025-12-03 21:48:11 INFO: Loading data with batch size 250...
147
+ 2025-12-03 21:48:58 INFO: Start evaluation...
148
+ 2025-12-03 21:53:09 INFO: UPOS XPOS UFeats AllTags
149
+ 2025-12-03 21:53:09 INFO: 85.39 67.66 79.30 66.13
150
+ 2025-12-03 21:53:09 INFO: POS Tagger score: sv_diachronic 66.13
151
+ 2025-12-03 21:53:10 INFO: Running tagger to retag /local/tmp.5441282/tmp9_locfmj/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
152
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmp9_locfmj/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
153
+ 2025-12-03 21:53:10 INFO: Running tagger in predict mode
154
+ 2025-12-03 21:53:10 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
155
+ 2025-12-03 21:53:13 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
156
+ 2025-12-03 21:53:13 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
157
+ 2025-12-03 21:53:13 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
158
+ 2025-12-03 21:53:13 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
159
+ 2025-12-03 21:53:13 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
160
+ 2025-12-03 21:53:13 INFO: Loading data with batch size 250...
161
+ 2025-12-03 21:53:13 INFO: Start evaluation...
162
+ 2025-12-03 21:53:13 INFO: UPOS XPOS UFeats AllTags
163
+ 2025-12-03 21:53:13 INFO: 88.61 79.46 83.42 74.75
164
+ 2025-12-03 21:53:13 INFO: POS Tagger score: sv_diachronic 74.75
165
+ 2025-12-03 21:53:13 INFO: Running tagger to retag /local/tmp.5441282/tmp9_locfmj/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
166
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5441282/tmp9_locfmj/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
167
+ 2025-12-03 21:53:13 INFO: Running tagger in predict mode
168
+ 2025-12-03 21:53:13 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
169
+ 2025-12-03 21:53:16 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
170
+ 2025-12-03 21:53:16 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
171
+ 2025-12-03 21:53:16 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
172
+ 2025-12-03 21:53:16 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
173
+ 2025-12-03 21:53:16 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
174
+ 2025-12-03 21:53:16 INFO: Loading data with batch size 250...
175
+ 2025-12-03 21:53:16 INFO: Start evaluation...
176
+ 2025-12-03 21:53:17 INFO: UPOS XPOS UFeats AllTags
177
+ 2025-12-03 21:53:17 INFO: 89.61 86.64 86.69 81.18
178
+ 2025-12-03 21:53:17 INFO: POS Tagger score: sv_diachronic 81.18
179
+ Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
180
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5441282/tmp9_locfmj/sv_diachronic.train.gold.conllu
181
+ Swapped 'w1, w2' for 'w1 ,w2' 216 times
182
+ Added 579 new sentences with asdf, zzzz -> asdf,zzzz
183
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5441282/tmp9_locfmj/sv_diachronic.dev.gold.conllu
184
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5441282/tmp9_locfmj/sv_diachronic.test.gold.conllu
185
+ Running stanza dependency parser training…
186
+ 2025-12-03 21:53:32 INFO: Training program called with:
187
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt --batch_size 32 --dropout 0.33
188
+ 2025-12-03 21:53:32 DEBUG: UD_Swedish-diachronic: sv_diachronic
189
+ 2025-12-03 21:53:32 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
190
+ 2025-12-03 21:53:32 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
191
+ 2025-12-03 21:53:32 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
192
+ 2025-12-03 21:53:32 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
193
+ 2025-12-03 21:53:32 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
194
+ 2025-12-03 21:53:32 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--batch_size', '32', '--dropout', '0.33']
195
+ 2025-12-03 21:53:32 INFO: Running parser in train mode
196
+ 2025-12-03 21:53:32 INFO: Using pretrained contextualized char embedding
197
+ 2025-12-03 21:53:32 INFO: Loading data with batch size 32...
198
+ 2025-12-03 21:53:41 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 76366
199
+ 2025-12-03 21:53:41 INFO: Original data size: 76366
200
+ 2025-12-03 21:53:42 INFO: Augmented data size: 77274
201
+ 2025-12-03 21:54:05 WARNING: sv_diachronic is not a known dataset. Examining the data to choose which xpos vocab to use
202
+ 2025-12-03 21:54:05 INFO: Original length = 77274
203
+ 2025-12-03 21:54:05 INFO: Filtered length = 77274
204
+ 2025-12-03 21:54:23 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
205
+ 2025-12-03 21:54:33 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
206
+ 2025-12-03 21:54:49 DEBUG: 46871 batches created.
207
+ 2025-12-03 21:54:49 DEBUG: 9 batches created.
208
+ 2025-12-03 21:54:49 INFO: Training parser...
209
+ 2025-12-03 21:54:49 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
210
+ 2025-12-03 21:54:49 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
211
+ 2025-12-03 21:54:49 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
212
+ 2025-12-03 21:54:50 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
213
+ 2025-12-03 21:54:58 INFO: Finished STEP 20/50000, loss = 6.937248 (0.206 sec/batch), lr: 0.003000
214
+ 2025-12-03 21:55:02 INFO: Finished STEP 40/50000, loss = 6.221194 (0.186 sec/batch), lr: 0.003000
215
+ 2025-12-03 21:55:05 INFO: Finished STEP 60/50000, loss = 6.004651 (0.157 sec/batch), lr: 0.003000
216
+ 2025-12-03 21:55:08 INFO: Finished STEP 80/50000, loss = 4.481259 (0.153 sec/batch), lr: 0.003000
217
+ 2025-12-03 21:55:11 INFO: Finished STEP 100/50000, loss = 4.621512 (0.146 sec/batch), lr: 0.003000
218
+ 2025-12-03 21:55:11 INFO: Evaluating on dev set...
219
+ 2025-12-03 21:55:12 INFO: LAS MLAS BLEX
220
+ 2025-12-03 21:55:12 INFO: 24.50 16.84 22.74
221
+ 2025-12-03 21:55:12 INFO: step 100: train_loss = 9.429485, dev_score = 0.2450
222
+ 2025-12-03 21:55:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
223
+ 2025-12-03 21:55:12 INFO: new best model saved.
224
+ 2025-12-03 21:55:12 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
225
+ 2025-12-03 21:55:12 INFO: new model checkpoint saved.
226
+ 2025-12-03 21:55:15 INFO: Finished STEP 120/50000, loss = 4.483292 (0.145 sec/batch), lr: 0.003000
227
+ 2025-12-03 21:55:18 INFO: Finished STEP 140/50000, loss = 5.362628 (0.139 sec/batch), lr: 0.003000
228
+ 2025-12-03 21:55:21 INFO: Finished STEP 160/50000, loss = 5.437283 (0.120 sec/batch), lr: 0.003000
229
+ 2025-12-03 21:55:23 INFO: Finished STEP 180/50000, loss = 4.825615 (0.122 sec/batch), lr: 0.003000
230
+ 2025-12-03 21:55:26 INFO: Finished STEP 200/50000, loss = 4.374323 (0.128 sec/batch), lr: 0.003000
231
+ 2025-12-03 21:55:26 INFO: Evaluating on dev set...
232
+ 2025-12-03 21:55:26 INFO: LAS MLAS BLEX
233
+ 2025-12-03 21:55:26 INFO: 41.83 28.75 34.17
234
+ 2025-12-03 21:55:26 INFO: step 200: train_loss = 5.169281, dev_score = 0.4183
235
+ 2025-12-03 21:55:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
236
+ 2025-12-03 21:55:26 INFO: new best model saved.
237
+ 2025-12-03 21:55:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
238
+ 2025-12-03 21:55:27 INFO: new model checkpoint saved.
239
+ 2025-12-03 21:55:30 INFO: Finished STEP 220/50000, loss = 4.692975 (0.123 sec/batch), lr: 0.003000
240
+ 2025-12-03 21:55:32 INFO: Finished STEP 240/50000, loss = 4.426068 (0.123 sec/batch), lr: 0.003000
241
+ 2025-12-03 21:55:34 INFO: Finished STEP 260/50000, loss = 5.457387 (0.119 sec/batch), lr: 0.003000
242
+ 2025-12-03 21:55:37 INFO: Finished STEP 280/50000, loss = 4.305553 (0.116 sec/batch), lr: 0.003000
243
+ 2025-12-03 21:55:39 INFO: Finished STEP 300/50000, loss = 4.359531 (0.117 sec/batch), lr: 0.003000
244
+ 2025-12-03 21:55:39 INFO: Evaluating on dev set...
245
+ 2025-12-03 21:55:40 INFO: LAS MLAS BLEX
246
+ 2025-12-03 21:55:40 INFO: 44.80 32.12 39.40
247
+ 2025-12-03 21:55:40 INFO: step 300: train_loss = 4.891226, dev_score = 0.4480
248
+ 2025-12-03 21:55:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
249
+ 2025-12-03 21:55:40 INFO: new best model saved.
250
+ 2025-12-03 21:55:41 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
251
+ 2025-12-03 21:55:41 INFO: new model checkpoint saved.
252
+ 2025-12-03 21:55:43 INFO: Finished STEP 320/50000, loss = 3.874385 (0.112 sec/batch), lr: 0.003000
253
+ 2025-12-03 21:55:45 INFO: Finished STEP 340/50000, loss = 4.105252 (0.120 sec/batch), lr: 0.003000
254
+ 2025-12-03 21:55:47 INFO: Finished STEP 360/50000, loss = 4.958949 (0.110 sec/batch), lr: 0.003000
255
+ 2025-12-03 21:55:49 INFO: Finished STEP 380/50000, loss = 4.490670 (0.114 sec/batch), lr: 0.003000
256
+ 2025-12-03 21:55:52 INFO: Finished STEP 400/50000, loss = 5.296941 (0.108 sec/batch), lr: 0.003000
257
+ 2025-12-03 21:55:52 INFO: Evaluating on dev set...
258
+ 2025-12-03 21:55:52 INFO: LAS MLAS BLEX
259
+ 2025-12-03 21:55:52 INFO: 53.47 42.37 45.76
260
+ 2025-12-03 21:55:52 INFO: step 400: train_loss = 4.699612, dev_score = 0.5347
261
+ 2025-12-03 21:55:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
262
+ 2025-12-03 21:55:52 INFO: new best model saved.
263
+ 2025-12-03 21:55:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
264
+ 2025-12-03 21:55:53 INFO: new model checkpoint saved.
265
+ 2025-12-03 21:55:55 INFO: Finished STEP 420/50000, loss = 4.217341 (0.103 sec/batch), lr: 0.003000
266
+ 2025-12-03 21:55:58 INFO: Finished STEP 440/50000, loss = 5.232607 (0.106 sec/batch), lr: 0.003000
267
+ 2025-12-03 21:56:00 INFO: Finished STEP 460/50000, loss = 4.069209 (0.101 sec/batch), lr: 0.003000
268
+ 2025-12-03 21:56:02 INFO: Finished STEP 480/50000, loss = 4.405638 (0.110 sec/batch), lr: 0.003000
269
+ 2025-12-03 21:56:04 INFO: Finished STEP 500/50000, loss = 3.847741 (0.097 sec/batch), lr: 0.003000
270
+ 2025-12-03 21:56:04 INFO: Evaluating on dev set...
271
+ 2025-12-03 21:56:04 INFO: LAS MLAS BLEX
272
+ 2025-12-03 21:56:04 INFO: 52.97 41.19 44.16
273
+ 2025-12-03 21:56:04 INFO: step 500: train_loss = 4.475364, dev_score = 0.5297
274
+ 2025-12-03 21:56:05 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
275
+ 2025-12-03 21:56:05 INFO: new model checkpoint saved.
276
+ 2025-12-03 21:56:07 INFO: Finished STEP 520/50000, loss = 4.622873 (0.095 sec/batch), lr: 0.003000
277
+ 2025-12-03 21:56:09 INFO: Finished STEP 540/50000, loss = 5.262753 (0.097 sec/batch), lr: 0.003000
278
+ 2025-12-03 21:56:11 INFO: Finished STEP 560/50000, loss = 3.362900 (0.101 sec/batch), lr: 0.003000
279
+ 2025-12-03 21:56:13 INFO: Finished STEP 580/50000, loss = 2.783157 (0.096 sec/batch), lr: 0.003000
280
+ 2025-12-03 21:56:15 INFO: Finished STEP 600/50000, loss = 4.463135 (0.086 sec/batch), lr: 0.003000
281
+ 2025-12-03 21:56:15 INFO: Evaluating on dev set...
282
+ 2025-12-03 21:56:16 INFO: LAS MLAS BLEX
283
+ 2025-12-03 21:56:16 INFO: 54.46 46.41 48.95
284
+ 2025-12-03 21:56:16 INFO: step 600: train_loss = 4.303635, dev_score = 0.5446
285
+ 2025-12-03 21:56:16 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
286
+ 2025-12-03 21:56:16 INFO: new best model saved.
287
+ 2025-12-03 21:56:17 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
288
+ 2025-12-03 21:56:17 INFO: new model checkpoint saved.
289
+ 2025-12-03 21:56:19 INFO: Finished STEP 620/50000, loss = 3.973504 (0.086 sec/batch), lr: 0.003000
290
+ 2025-12-03 21:56:21 INFO: Finished STEP 640/50000, loss = 4.144060 (0.088 sec/batch), lr: 0.003000
291
+ 2025-12-03 21:56:23 INFO: Finished STEP 660/50000, loss = 4.311639 (0.100 sec/batch), lr: 0.003000
292
+ 2025-12-03 21:56:25 INFO: Finished STEP 680/50000, loss = 3.482852 (0.108 sec/batch), lr: 0.003000
293
+ 2025-12-03 21:56:27 INFO: Finished STEP 700/50000, loss = 4.920451 (0.105 sec/batch), lr: 0.003000
294
+ 2025-12-03 21:56:27 INFO: Evaluating on dev set...
295
+ 2025-12-03 21:56:27 INFO: LAS MLAS BLEX
296
+ 2025-12-03 21:56:27 INFO: 54.95 46.09 51.16
297
+ 2025-12-03 21:56:27 INFO: step 700: train_loss = 4.357117, dev_score = 0.5495
298
+ 2025-12-03 21:56:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
299
+ 2025-12-03 21:56:27 INFO: new best model saved.
300
+ 2025-12-03 21:56:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
301
+ 2025-12-03 21:56:28 INFO: new model checkpoint saved.
302
+ 2025-12-03 21:56:30 INFO: Finished STEP 720/50000, loss = 4.037713 (0.097 sec/batch), lr: 0.003000
303
+ 2025-12-03 21:56:32 INFO: Finished STEP 740/50000, loss = 4.073860 (0.093 sec/batch), lr: 0.003000
304
+ 2025-12-03 21:56:34 INFO: Finished STEP 760/50000, loss = 3.492782 (0.087 sec/batch), lr: 0.003000
305
+ 2025-12-03 21:56:36 INFO: Finished STEP 780/50000, loss = 4.071904 (0.096 sec/batch), lr: 0.003000
306
+ 2025-12-03 21:56:38 INFO: Finished STEP 800/50000, loss = 3.010960 (0.094 sec/batch), lr: 0.003000
307
+ 2025-12-03 21:56:38 INFO: Evaluating on dev set...
308
+ 2025-12-03 21:56:38 INFO: LAS MLAS BLEX
309
+ 2025-12-03 21:56:38 INFO: 52.48 44.54 48.74
310
+ 2025-12-03 21:56:38 INFO: step 800: train_loss = 4.207058, dev_score = 0.5248
311
+ 2025-12-03 21:56:39 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
312
+ 2025-12-03 21:56:39 INFO: new model checkpoint saved.
313
+ 2025-12-03 21:56:41 INFO: Finished STEP 820/50000, loss = 4.195956 (0.092 sec/batch), lr: 0.003000
314
+ 2025-12-03 21:56:43 INFO: Finished STEP 840/50000, loss = 3.660316 (0.090 sec/batch), lr: 0.003000
315
+ 2025-12-03 21:56:44 INFO: Finished STEP 860/50000, loss = 3.070045 (0.107 sec/batch), lr: 0.003000
316
+ 2025-12-03 21:56:46 INFO: Finished STEP 880/50000, loss = 3.561736 (0.087 sec/batch), lr: 0.003000
317
+ 2025-12-03 21:56:48 INFO: Finished STEP 900/50000, loss = 4.851887 (0.078 sec/batch), lr: 0.003000
318
+ 2025-12-03 21:56:48 INFO: Evaluating on dev set...
319
+ 2025-12-03 21:56:49 INFO: LAS MLAS BLEX
320
+ 2025-12-03 21:56:49 INFO: 51.73 41.09 44.03
321
+ 2025-12-03 21:56:49 INFO: step 900: train_loss = 4.232460, dev_score = 0.5173
322
+ 2025-12-03 21:56:49 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
323
+ 2025-12-03 21:56:49 INFO: new model checkpoint saved.
324
+ 2025-12-03 21:56:51 INFO: Finished STEP 920/50000, loss = 4.294917 (0.088 sec/batch), lr: 0.003000
325
+ 2025-12-03 21:56:53 INFO: Finished STEP 940/50000, loss = 3.520856 (0.093 sec/batch), lr: 0.003000
326
+ 2025-12-03 21:56:55 INFO: Finished STEP 960/50000, loss = 3.940510 (0.095 sec/batch), lr: 0.003000
327
+ 2025-12-03 21:56:57 INFO: Finished STEP 980/50000, loss = 4.757580 (0.083 sec/batch), lr: 0.003000
328
+ 2025-12-03 21:56:58 INFO: Finished STEP 1000/50000, loss = 3.585855 (0.084 sec/batch), lr: 0.003000
329
+ 2025-12-03 21:56:58 INFO: Evaluating on dev set...
330
+ 2025-12-03 21:56:59 INFO: LAS MLAS BLEX
331
+ 2025-12-03 21:56:59 INFO: 52.72 44.96 46.64
332
+ 2025-12-03 21:56:59 INFO: step 1000: train_loss = 4.377043, dev_score = 0.5272
333
+ 2025-12-03 21:57:00 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
334
+ 2025-12-03 21:57:00 INFO: new model checkpoint saved.
335
+ 2025-12-03 21:57:01 INFO: Finished STEP 1020/50000, loss = 4.979126 (0.092 sec/batch), lr: 0.003000
336
+ 2025-12-03 21:57:03 INFO: Finished STEP 1040/50000, loss = 3.792933 (0.090 sec/batch), lr: 0.003000
337
+ 2025-12-03 21:57:05 INFO: Finished STEP 1060/50000, loss = 4.987607 (0.085 sec/batch), lr: 0.003000
338
+ 2025-12-03 21:57:07 INFO: Finished STEP 1080/50000, loss = 3.836903 (0.086 sec/batch), lr: 0.003000
339
+ 2025-12-03 21:57:09 INFO: Finished STEP 1100/50000, loss = 3.591379 (0.087 sec/batch), lr: 0.003000
340
+ 2025-12-03 21:57:09 INFO: Evaluating on dev set...
341
+ 2025-12-03 21:57:09 INFO: LAS MLAS BLEX
342
+ 2025-12-03 21:57:09 INFO: 55.45 45.67 49.05
343
+ 2025-12-03 21:57:09 INFO: step 1100: train_loss = 4.262863, dev_score = 0.5545
344
+ 2025-12-03 21:57:10 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
345
+ 2025-12-03 21:57:10 INFO: new best model saved.
346
+ 2025-12-03 21:57:10 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
347
+ 2025-12-03 21:57:10 INFO: new model checkpoint saved.
348
+ 2025-12-03 21:57:12 INFO: Finished STEP 1120/50000, loss = 4.292474 (0.093 sec/batch), lr: 0.003000
349
+ 2025-12-03 21:57:14 INFO: Finished STEP 1140/50000, loss = 3.930135 (0.075 sec/batch), lr: 0.003000
350
+ 2025-12-03 21:57:15 INFO: Finished STEP 1160/50000, loss = 3.263464 (0.085 sec/batch), lr: 0.003000
351
+ 2025-12-03 21:57:17 INFO: Finished STEP 1180/50000, loss = 3.390361 (0.088 sec/batch), lr: 0.003000
352
+ 2025-12-03 21:57:19 INFO: Finished STEP 1200/50000, loss = 3.816272 (0.094 sec/batch), lr: 0.003000
353
+ 2025-12-03 21:57:19 INFO: Evaluating on dev set...
354
+ 2025-12-03 21:57:19 INFO: LAS MLAS BLEX
355
+ 2025-12-03 21:57:19 INFO: 58.42 46.32 49.68
356
+ 2025-12-03 21:57:19 INFO: step 1200: train_loss = 4.076403, dev_score = 0.5842
357
+ 2025-12-03 21:57:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
358
+ 2025-12-03 21:57:20 INFO: new best model saved.
359
+ 2025-12-03 21:57:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
360
+ 2025-12-03 21:57:20 INFO: new model checkpoint saved.
361
+ 2025-12-03 21:57:22 INFO: Finished STEP 1220/50000, loss = 5.270017 (0.089 sec/batch), lr: 0.003000
362
+ 2025-12-03 21:57:24 INFO: Finished STEP 1240/50000, loss = 4.439935 (0.084 sec/batch), lr: 0.003000
363
+ 2025-12-03 21:57:26 INFO: Finished STEP 1260/50000, loss = 3.901986 (0.090 sec/batch), lr: 0.003000
364
+ 2025-12-03 21:57:27 INFO: Finished STEP 1280/50000, loss = 4.144909 (0.088 sec/batch), lr: 0.003000
365
+ 2025-12-03 21:57:29 INFO: Finished STEP 1300/50000, loss = 3.075126 (0.090 sec/batch), lr: 0.003000
366
+ 2025-12-03 21:57:29 INFO: Evaluating on dev set...
367
+ 2025-12-03 21:57:30 INFO: LAS MLAS BLEX
368
+ 2025-12-03 21:57:30 INFO: 58.42 48.84 51.37
369
+ 2025-12-03 21:57:30 INFO: step 1300: train_loss = 3.918128, dev_score = 0.5842
370
+ 2025-12-03 21:57:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
371
+ 2025-12-03 21:57:30 INFO: new best model saved.
372
+ 2025-12-03 21:57:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
373
+ 2025-12-03 21:57:30 INFO: new model checkpoint saved.
374
+ 2025-12-03 21:57:32 INFO: Finished STEP 1320/50000, loss = 2.861002 (0.084 sec/batch), lr: 0.003000
375
+ 2025-12-03 21:57:34 INFO: Finished STEP 1340/50000, loss = 4.137879 (0.082 sec/batch), lr: 0.003000
376
+ 2025-12-03 21:57:36 INFO: Finished STEP 1360/50000, loss = 6.432682 (0.084 sec/batch), lr: 0.003000
377
+ 2025-12-03 21:57:37 INFO: Finished STEP 1380/50000, loss = 5.857590 (0.083 sec/batch), lr: 0.003000
378
+ 2025-12-03 21:57:39 INFO: Finished STEP 1400/50000, loss = 4.340917 (0.079 sec/batch), lr: 0.003000
379
+ 2025-12-03 21:57:39 INFO: Evaluating on dev set...
380
+ 2025-12-03 21:57:40 INFO: LAS MLAS BLEX
381
+ 2025-12-03 21:57:40 INFO: 52.48 43.37 46.74
382
+ 2025-12-03 21:57:40 INFO: step 1400: train_loss = 4.200418, dev_score = 0.5248
383
+ 2025-12-03 21:57:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
384
+ 2025-12-03 21:57:40 INFO: new model checkpoint saved.
385
+ 2025-12-03 21:57:42 INFO: Finished STEP 1420/50000, loss = 4.600896 (0.079 sec/batch), lr: 0.003000
386
+ 2025-12-03 21:57:44 INFO: Finished STEP 1440/50000, loss = 3.618456 (0.088 sec/batch), lr: 0.003000
387
+ 2025-12-03 21:57:45 INFO: Finished STEP 1460/50000, loss = 6.108145 (0.080 sec/batch), lr: 0.003000
388
+ 2025-12-03 21:57:47 INFO: Finished STEP 1480/50000, loss = 4.806163 (0.084 sec/batch), lr: 0.003000
389
+ 2025-12-03 21:57:49 INFO: Finished STEP 1500/50000, loss = 3.163539 (0.087 sec/batch), lr: 0.003000
390
+ 2025-12-03 21:57:49 INFO: Evaluating on dev set...
391
+ 2025-12-03 21:57:49 INFO: LAS MLAS BLEX
392
+ 2025-12-03 21:57:49 INFO: 56.93 47.28 51.05
393
+ 2025-12-03 21:57:49 INFO: step 1500: train_loss = 4.250937, dev_score = 0.5693
394
+ 2025-12-03 21:57:50 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
395
+ 2025-12-03 21:57:50 INFO: new model checkpoint saved.
396
+ 2025-12-03 21:57:51 INFO: Finished STEP 1520/50000, loss = 5.288736 (0.080 sec/batch), lr: 0.003000
397
+ 2025-12-03 21:57:53 INFO: Finished STEP 1540/50000, loss = 6.437222 (0.080 sec/batch), lr: 0.003000
398
+ 2025-12-03 21:57:55 INFO: Finished STEP 1560/50000, loss = 6.026221 (0.081 sec/batch), lr: 0.003000
399
+ 2025-12-03 21:57:56 INFO: Finished STEP 1580/50000, loss = 3.159218 (0.069 sec/batch), lr: 0.003000
400
+ 2025-12-03 21:57:58 INFO: Finished STEP 1600/50000, loss = 4.267678 (0.076 sec/batch), lr: 0.003000
401
+ 2025-12-03 21:57:58 INFO: Evaluating on dev set...
402
+ 2025-12-03 21:57:58 INFO: LAS MLAS BLEX
403
+ 2025-12-03 21:57:58 INFO: 56.44 49.58 52.94
404
+ 2025-12-03 21:57:58 INFO: step 1600: train_loss = 4.572977, dev_score = 0.5644
405
+ 2025-12-03 21:57:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
406
+ 2025-12-03 21:57:59 INFO: new model checkpoint saved.
407
+ 2025-12-03 21:58:01 INFO: Finished STEP 1620/50000, loss = 3.075799 (0.084 sec/batch), lr: 0.003000
408
+ 2025-12-03 21:58:02 INFO: Finished STEP 1640/50000, loss = 4.702259 (0.084 sec/batch), lr: 0.003000
409
+ 2025-12-03 21:58:04 INFO: Finished STEP 1660/50000, loss = 4.056483 (0.080 sec/batch), lr: 0.003000
410
+ 2025-12-03 21:58:06 INFO: Finished STEP 1680/50000, loss = 5.785334 (0.085 sec/batch), lr: 0.003000
411
+ 2025-12-03 21:58:07 INFO: Finished STEP 1700/50000, loss = 6.958914 (0.076 sec/batch), lr: 0.003000
412
+ 2025-12-03 21:58:07 INFO: Evaluating on dev set...
413
+ 2025-12-03 21:58:08 INFO: LAS MLAS BLEX
414
+ 2025-12-03 21:58:08 INFO: 58.42 47.90 51.68
415
+ 2025-12-03 21:58:08 INFO: step 1700: train_loss = 4.164190, dev_score = 0.5842
416
+ 2025-12-03 21:58:08 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
417
+ 2025-12-03 21:58:08 INFO: new best model saved.
418
+ 2025-12-03 21:58:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
419
+ 2025-12-03 21:58:09 INFO: new model checkpoint saved.
420
+ 2025-12-03 21:58:10 INFO: Finished STEP 1720/50000, loss = 4.814258 (0.076 sec/batch), lr: 0.003000
421
+ 2025-12-03 21:58:12 INFO: Finished STEP 1740/50000, loss = 4.181596 (0.079 sec/batch), lr: 0.003000
422
+ 2025-12-03 21:58:13 INFO: Finished STEP 1760/50000, loss = 2.630800 (0.074 sec/batch), lr: 0.003000
423
+ 2025-12-03 21:58:15 INFO: Finished STEP 1780/50000, loss = 4.297536 (0.083 sec/batch), lr: 0.003000
424
+ 2025-12-03 21:58:17 INFO: Finished STEP 1800/50000, loss = 4.589555 (0.085 sec/batch), lr: 0.003000
425
+ 2025-12-03 21:58:17 INFO: Evaluating on dev set...
426
+ 2025-12-03 21:58:17 INFO: LAS MLAS BLEX
427
+ 2025-12-03 21:58:17 INFO: 57.92 50.53 53.05
428
+ 2025-12-03 21:58:17 INFO: step 1800: train_loss = 4.259523, dev_score = 0.5792
429
+ 2025-12-03 21:58:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
430
+ 2025-12-03 21:58:18 INFO: new model checkpoint saved.
431
+ 2025-12-03 21:58:19 INFO: Finished STEP 1820/50000, loss = 5.325808 (0.080 sec/batch), lr: 0.003000
432
+ 2025-12-03 21:58:21 INFO: Finished STEP 1840/50000, loss = 3.579389 (0.082 sec/batch), lr: 0.003000
433
+ 2025-12-03 21:58:23 INFO: Finished STEP 1860/50000, loss = 4.758757 (0.085 sec/batch), lr: 0.003000
434
+ 2025-12-03 21:58:24 INFO: Finished STEP 1880/50000, loss = 3.579899 (0.075 sec/batch), lr: 0.003000
435
+ 2025-12-03 21:58:26 INFO: Finished STEP 1900/50000, loss = 3.342232 (0.080 sec/batch), lr: 0.003000
436
+ 2025-12-03 21:58:26 INFO: Evaluating on dev set...
437
+ 2025-12-03 21:58:26 INFO: LAS MLAS BLEX
438
+ 2025-12-03 21:58:26 INFO: 56.93 46.22 53.36
439
+ 2025-12-03 21:58:26 INFO: step 1900: train_loss = 4.426542, dev_score = 0.5693
440
+ 2025-12-03 21:58:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
441
+ 2025-12-03 21:58:27 INFO: new model checkpoint saved.
442
+ 2025-12-03 21:58:28 INFO: Finished STEP 1920/50000, loss = 5.643509 (0.079 sec/batch), lr: 0.003000
443
+ 2025-12-03 21:58:30 INFO: Finished STEP 1940/50000, loss = 2.522238 (0.074 sec/batch), lr: 0.003000
444
+ 2025-12-03 21:58:32 INFO: Finished STEP 1960/50000, loss = 3.243419 (0.077 sec/batch), lr: 0.003000
445
+ 2025-12-03 21:58:33 INFO: Finished STEP 1980/50000, loss = 3.340650 (0.077 sec/batch), lr: 0.003000
446
+ 2025-12-03 21:58:35 INFO: Finished STEP 2000/50000, loss = 7.267831 (0.079 sec/batch), lr: 0.003000
447
+ 2025-12-03 21:58:35 INFO: Evaluating on dev set...
448
+ 2025-12-03 21:58:35 INFO: LAS MLAS BLEX
449
+ 2025-12-03 21:58:35 INFO: 57.67 51.79 55.16
450
+ 2025-12-03 21:58:35 INFO: step 2000: train_loss = 4.196674, dev_score = 0.5767
451
+ 2025-12-03 21:58:36 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
452
+ 2025-12-03 21:58:36 INFO: new model checkpoint saved.
453
+ 2025-12-03 21:58:38 INFO: Finished STEP 2020/50000, loss = 3.868695 (0.077 sec/batch), lr: 0.003000
454
+ 2025-12-03 21:58:39 INFO: Finished STEP 2040/50000, loss = 5.761869 (0.065 sec/batch), lr: 0.003000
455
+ 2025-12-03 21:58:41 INFO: Finished STEP 2060/50000, loss = 5.211999 (0.085 sec/batch), lr: 0.003000
456
+ 2025-12-03 21:58:42 INFO: Finished STEP 2080/50000, loss = 4.174130 (0.076 sec/batch), lr: 0.003000
457
+ 2025-12-03 21:58:44 INFO: Finished STEP 2100/50000, loss = 3.859421 (0.080 sec/batch), lr: 0.003000
458
+ 2025-12-03 21:58:44 INFO: Evaluating on dev set...
459
+ 2025-12-03 21:58:44 INFO: LAS MLAS BLEX
460
+ 2025-12-03 21:58:44 INFO: 54.46 45.57 49.37
461
+ 2025-12-03 21:58:44 INFO: step 2100: train_loss = 4.484527, dev_score = 0.5446
462
+ 2025-12-03 21:58:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
463
+ 2025-12-03 21:58:45 INFO: new model checkpoint saved.
464
+ 2025-12-03 21:58:46 INFO: Finished STEP 2120/50000, loss = 4.043532 (0.066 sec/batch), lr: 0.003000
465
+ 2025-12-03 21:58:48 INFO: Finished STEP 2140/50000, loss = 4.347118 (0.075 sec/batch), lr: 0.003000
466
+ 2025-12-03 21:58:50 INFO: Finished STEP 2160/50000, loss = 3.567682 (0.076 sec/batch), lr: 0.003000
467
+ 2025-12-03 21:58:51 INFO: Finished STEP 2180/50000, loss = 4.363184 (0.078 sec/batch), lr: 0.003000
468
+ 2025-12-03 21:58:53 INFO: Finished STEP 2200/50000, loss = 6.490901 (0.081 sec/batch), lr: 0.003000
469
+ 2025-12-03 21:58:53 INFO: Evaluating on dev set...
470
+ 2025-12-03 21:58:53 INFO: LAS MLAS BLEX
471
+ 2025-12-03 21:58:53 INFO: 58.17 48.64 52.41
472
+ 2025-12-03 21:58:53 INFO: step 2200: train_loss = 4.329839, dev_score = 0.5817
473
+ 2025-12-03 21:58:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
474
+ 2025-12-03 21:58:54 INFO: new model checkpoint saved.
475
+ 2025-12-03 21:58:55 INFO: Finished STEP 2220/50000, loss = 4.196492 (0.077 sec/batch), lr: 0.003000
476
+ 2025-12-03 21:58:57 INFO: Finished STEP 2240/50000, loss = 5.155959 (0.078 sec/batch), lr: 0.003000
477
+ 2025-12-03 21:58:59 INFO: Finished STEP 2260/50000, loss = 3.478016 (0.074 sec/batch), lr: 0.003000
478
+ 2025-12-03 21:59:00 INFO: Finished STEP 2280/50000, loss = 5.644300 (0.081 sec/batch), lr: 0.003000
479
+ 2025-12-03 21:59:02 INFO: Finished STEP 2300/50000, loss = 3.756620 (0.087 sec/batch), lr: 0.003000
480
+ 2025-12-03 21:59:02 INFO: Evaluating on dev set...
481
+ 2025-12-03 21:59:02 INFO: LAS MLAS BLEX
482
+ 2025-12-03 21:59:02 INFO: 56.19 48.65 51.14
483
+ 2025-12-03 21:59:02 INFO: step 2300: train_loss = 4.281249, dev_score = 0.5619
484
+ 2025-12-03 21:59:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
485
+ 2025-12-03 21:59:03 INFO: new model checkpoint saved.
486
+ 2025-12-03 21:59:04 INFO: Finished STEP 2320/50000, loss = 3.944878 (0.074 sec/batch), lr: 0.003000
487
+ 2025-12-03 21:59:06 INFO: Finished STEP 2340/50000, loss = 6.044787 (0.082 sec/batch), lr: 0.003000
488
+ 2025-12-03 21:59:07 INFO: Finished STEP 2360/50000, loss = 6.580403 (0.074 sec/batch), lr: 0.003000
489
+ 2025-12-03 21:59:09 INFO: Finished STEP 2380/50000, loss = 3.113432 (0.077 sec/batch), lr: 0.003000
490
+ 2025-12-03 21:59:10 INFO: Finished STEP 2400/50000, loss = 3.035348 (0.077 sec/batch), lr: 0.003000
491
+ 2025-12-03 21:59:10 INFO: Evaluating on dev set...
492
+ 2025-12-03 21:59:11 INFO: LAS MLAS BLEX
493
+ 2025-12-03 21:59:11 INFO: 57.43 48.42 52.63
494
+ 2025-12-03 21:59:11 INFO: step 2400: train_loss = 4.313338, dev_score = 0.5743
495
+ 2025-12-03 21:59:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
496
+ 2025-12-03 21:59:11 INFO: new model checkpoint saved.
497
+ 2025-12-03 21:59:13 INFO: Finished STEP 2420/50000, loss = 4.012220 (0.076 sec/batch), lr: 0.003000
498
+ 2025-12-03 21:59:14 INFO: Finished STEP 2440/50000, loss = 4.620986 (0.070 sec/batch), lr: 0.003000
499
+ 2025-12-03 21:59:16 INFO: Finished STEP 2460/50000, loss = 3.839270 (0.071 sec/batch), lr: 0.003000
500
+ 2025-12-03 21:59:17 INFO: Finished STEP 2480/50000, loss = 5.325880 (0.070 sec/batch), lr: 0.003000
501
+ 2025-12-03 21:59:19 INFO: Finished STEP 2500/50000, loss = 4.642661 (0.077 sec/batch), lr: 0.003000
502
+ 2025-12-03 21:59:19 INFO: Evaluating on dev set...
503
+ 2025-12-03 21:59:19 INFO: LAS MLAS BLEX
504
+ 2025-12-03 21:59:19 INFO: 54.95 44.92 49.58
505
+ 2025-12-03 21:59:19 INFO: step 2500: train_loss = 4.171005, dev_score = 0.5495
506
+ 2025-12-03 21:59:20 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
507
+ 2025-12-03 21:59:20 INFO: new model checkpoint saved.
508
+ 2025-12-03 21:59:21 INFO: Finished STEP 2520/50000, loss = 2.329720 (0.079 sec/batch), lr: 0.003000
509
+ 2025-12-03 21:59:23 INFO: Finished STEP 2540/50000, loss = 4.466135 (0.067 sec/batch), lr: 0.003000
510
+ 2025-12-03 21:59:24 INFO: Finished STEP 2560/50000, loss = 6.058784 (0.072 sec/batch), lr: 0.003000
511
+ 2025-12-03 21:59:26 INFO: Finished STEP 2580/50000, loss = 9.061809 (0.090 sec/batch), lr: 0.003000
512
+ 2025-12-03 21:59:27 INFO: Finished STEP 2600/50000, loss = 5.426898 (0.076 sec/batch), lr: 0.003000
513
+ 2025-12-03 21:59:27 INFO: Evaluating on dev set...
514
+ 2025-12-03 21:59:28 INFO: LAS MLAS BLEX
515
+ 2025-12-03 21:59:28 INFO: 55.20 44.63 47.58
516
+ 2025-12-03 21:59:28 INFO: step 2600: train_loss = 4.241316, dev_score = 0.5520
517
+ 2025-12-03 21:59:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
518
+ 2025-12-03 21:59:28 INFO: new model checkpoint saved.
519
+ 2025-12-03 21:59:30 INFO: Finished STEP 2620/50000, loss = 4.473849 (0.077 sec/batch), lr: 0.003000
520
+ 2025-12-03 21:59:31 INFO: Finished STEP 2640/50000, loss = 3.683683 (0.074 sec/batch), lr: 0.003000
521
+ 2025-12-03 21:59:33 INFO: Finished STEP 2660/50000, loss = 3.074204 (0.078 sec/batch), lr: 0.003000
522
+ 2025-12-03 21:59:34 INFO: Finished STEP 2680/50000, loss = 3.974907 (0.073 sec/batch), lr: 0.003000
523
+ 2025-12-03 21:59:36 INFO: Finished STEP 2700/50000, loss = 2.102455 (0.074 sec/batch), lr: 0.003000
524
+ 2025-12-03 21:59:36 INFO: Evaluating on dev set...
525
+ 2025-12-03 21:59:36 INFO: LAS MLAS BLEX
526
+ 2025-12-03 21:59:36 INFO: 61.39 53.19 55.32
527
+ 2025-12-03 21:59:36 INFO: step 2700: train_loss = 4.094942, dev_score = 0.6139
528
+ 2025-12-03 21:59:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
529
+ 2025-12-03 21:59:37 INFO: new best model saved.
530
+ 2025-12-03 21:59:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
531
+ 2025-12-03 21:59:37 INFO: new model checkpoint saved.
532
+ 2025-12-03 21:59:39 INFO: Finished STEP 2720/50000, loss = 5.810821 (0.063 sec/batch), lr: 0.003000
533
+ 2025-12-03 21:59:40 INFO: Finished STEP 2740/50000, loss = 2.966178 (0.072 sec/batch), lr: 0.003000
534
+ 2025-12-03 21:59:42 INFO: Finished STEP 2760/50000, loss = 4.077942 (0.068 sec/batch), lr: 0.003000
535
+ 2025-12-03 21:59:43 INFO: Finished STEP 2780/50000, loss = 3.828436 (0.074 sec/batch), lr: 0.003000
536
+ 2025-12-03 21:59:45 INFO: Finished STEP 2800/50000, loss = 3.264549 (0.070 sec/batch), lr: 0.003000
537
+ 2025-12-03 21:59:45 INFO: Evaluating on dev set...
538
+ 2025-12-03 21:59:45 INFO: LAS MLAS BLEX
539
+ 2025-12-03 21:59:45 INFO: 58.91 51.37 54.74
540
+ 2025-12-03 21:59:45 INFO: step 2800: train_loss = 4.226886, dev_score = 0.5891
541
+ 2025-12-03 21:59:46 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
542
+ 2025-12-03 21:59:46 INFO: new model checkpoint saved.
543
+ 2025-12-03 21:59:47 INFO: Finished STEP 2820/50000, loss = 4.143430 (0.071 sec/batch), lr: 0.003000
544
+ 2025-12-03 21:59:49 INFO: Finished STEP 2840/50000, loss = 4.038007 (0.073 sec/batch), lr: 0.003000
545
+ 2025-12-03 21:59:50 INFO: Finished STEP 2860/50000, loss = 3.176973 (0.068 sec/batch), lr: 0.003000
546
+ 2025-12-03 21:59:52 INFO: Finished STEP 2880/50000, loss = 3.589462 (0.071 sec/batch), lr: 0.003000
547
+ 2025-12-03 21:59:53 INFO: Finished STEP 2900/50000, loss = 4.507996 (0.068 sec/batch), lr: 0.003000
548
+ 2025-12-03 21:59:53 INFO: Evaluating on dev set...
549
+ 2025-12-03 21:59:54 INFO: LAS MLAS BLEX
550
+ 2025-12-03 21:59:54 INFO: 57.67 49.37 51.48
551
+ 2025-12-03 21:59:54 INFO: step 2900: train_loss = 3.985535, dev_score = 0.5767
552
+ 2025-12-03 21:59:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
553
+ 2025-12-03 21:59:54 INFO: new model checkpoint saved.
554
+ 2025-12-03 21:59:56 INFO: Finished STEP 2920/50000, loss = 5.738523 (0.072 sec/batch), lr: 0.003000
555
+ 2025-12-03 21:59:57 INFO: Finished STEP 2940/50000, loss = 5.717393 (0.074 sec/batch), lr: 0.003000
556
+ 2025-12-03 21:59:59 INFO: Finished STEP 2960/50000, loss = 3.576367 (0.073 sec/batch), lr: 0.003000
557
+ 2025-12-03 22:00:00 INFO: Finished STEP 2980/50000, loss = 3.845478 (0.064 sec/batch), lr: 0.003000
558
+ 2025-12-03 22:00:01 INFO: Finished STEP 3000/50000, loss = 5.697991 (0.068 sec/batch), lr: 0.003000
559
+ 2025-12-03 22:00:01 INFO: Evaluating on dev set...
560
+ 2025-12-03 22:00:02 INFO: LAS MLAS BLEX
561
+ 2025-12-03 22:00:02 INFO: 58.17 48.64 52.41
562
+ 2025-12-03 22:00:02 INFO: step 3000: train_loss = 4.231318, dev_score = 0.5817
563
+ 2025-12-03 22:00:03 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
564
+ 2025-12-03 22:00:03 INFO: new model checkpoint saved.
565
+ 2025-12-03 22:00:04 INFO: Finished STEP 3020/50000, loss = 7.619741 (0.068 sec/batch), lr: 0.003000
566
+ 2025-12-03 22:00:05 INFO: Finished STEP 3040/50000, loss = 5.920513 (0.073 sec/batch), lr: 0.003000
567
+ 2025-12-03 22:00:07 INFO: Finished STEP 3060/50000, loss = 5.683625 (0.071 sec/batch), lr: 0.003000
568
+ 2025-12-03 22:00:08 INFO: Finished STEP 3080/50000, loss = 3.095924 (0.066 sec/batch), lr: 0.003000
569
+ 2025-12-03 22:00:10 INFO: Finished STEP 3100/50000, loss = 3.130288 (0.068 sec/batch), lr: 0.003000
570
+ 2025-12-03 22:00:10 INFO: Evaluating on dev set...
571
+ 2025-12-03 22:00:10 INFO: LAS MLAS BLEX
572
+ 2025-12-03 22:00:10 INFO: 58.66 50.00 53.39
573
+ 2025-12-03 22:00:10 INFO: step 3100: train_loss = 4.305735, dev_score = 0.5866
574
+ 2025-12-03 22:00:11 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
575
+ 2025-12-03 22:00:11 INFO: new model checkpoint saved.
576
+ 2025-12-03 22:00:12 INFO: Finished STEP 3120/50000, loss = 5.291130 (0.071 sec/batch), lr: 0.003000
577
+ 2025-12-03 22:00:14 INFO: Finished STEP 3140/50000, loss = 4.440284 (0.067 sec/batch), lr: 0.003000
578
+ 2025-12-03 22:00:15 INFO: Finished STEP 3160/50000, loss = 4.735047 (0.069 sec/batch), lr: 0.003000
579
+ 2025-12-03 22:00:17 INFO: Finished STEP 3180/50000, loss = 5.151906 (0.068 sec/batch), lr: 0.003000
580
+ 2025-12-03 22:00:18 INFO: Finished STEP 3200/50000, loss = 5.473666 (0.066 sec/batch), lr: 0.003000
581
+ 2025-12-03 22:00:18 INFO: Evaluating on dev set...
582
+ 2025-12-03 22:00:19 INFO: LAS MLAS BLEX
583
+ 2025-12-03 22:00:19 INFO: 58.17 49.58 52.97
584
+ 2025-12-03 22:00:19 INFO: step 3200: train_loss = 4.252154, dev_score = 0.5817
585
+ 2025-12-03 22:00:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
586
+ 2025-12-03 22:00:19 INFO: new model checkpoint saved.
587
+ 2025-12-03 22:00:21 INFO: Finished STEP 3220/50000, loss = 4.079478 (0.078 sec/batch), lr: 0.003000
588
+ 2025-12-03 22:00:22 INFO: Finished STEP 3240/50000, loss = 3.994006 (0.065 sec/batch), lr: 0.003000
589
+ 2025-12-03 22:00:23 INFO: Finished STEP 3260/50000, loss = 3.426797 (0.075 sec/batch), lr: 0.003000
590
+ 2025-12-03 22:00:25 INFO: Finished STEP 3280/50000, loss = 2.964750 (0.068 sec/batch), lr: 0.003000
591
+ 2025-12-03 22:00:26 INFO: Finished STEP 3300/50000, loss = 4.955059 (0.069 sec/batch), lr: 0.003000
592
+ 2025-12-03 22:00:26 INFO: Evaluating on dev set...
593
+ 2025-12-03 22:00:27 INFO: LAS MLAS BLEX
594
+ 2025-12-03 22:00:27 INFO: 57.92 45.67 52.43
595
+ 2025-12-03 22:00:27 INFO: step 3300: train_loss = 4.448298, dev_score = 0.5792
596
+ 2025-12-03 22:00:28 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
597
+ 2025-12-03 22:00:28 INFO: new model checkpoint saved.
598
+ 2025-12-03 22:00:29 INFO: Finished STEP 3320/50000, loss = 5.514875 (0.060 sec/batch), lr: 0.003000
599
+ 2025-12-03 22:00:30 INFO: Finished STEP 3340/50000, loss = 5.389848 (0.067 sec/batch), lr: 0.003000
600
+ 2025-12-03 22:00:32 INFO: Finished STEP 3360/50000, loss = 6.221178 (0.063 sec/batch), lr: 0.003000
601
+ 2025-12-03 22:00:33 INFO: Finished STEP 3380/50000, loss = 4.681024 (0.069 sec/batch), lr: 0.003000
602
+ 2025-12-03 22:00:35 INFO: Finished STEP 3400/50000, loss = 4.096542 (0.065 sec/batch), lr: 0.003000
603
+ 2025-12-03 22:00:35 INFO: Evaluating on dev set...
604
+ 2025-12-03 22:00:35 INFO: LAS MLAS BLEX
605
+ 2025-12-03 22:00:35 INFO: 57.92 49.47 53.73
606
+ 2025-12-03 22:00:35 INFO: step 3400: train_loss = 4.527486, dev_score = 0.5792
607
+ 2025-12-03 22:00:36 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
608
+ 2025-12-03 22:00:36 INFO: new model checkpoint saved.
609
+ 2025-12-03 22:00:37 INFO: Finished STEP 3420/50000, loss = 4.950140 (0.063 sec/batch), lr: 0.003000
610
+ 2025-12-03 22:00:38 INFO: Finished STEP 3440/50000, loss = 3.049980 (0.068 sec/batch), lr: 0.003000
611
+ 2025-12-03 22:00:40 INFO: Finished STEP 3460/50000, loss = 3.217714 (0.069 sec/batch), lr: 0.003000
612
+ 2025-12-03 22:00:41 INFO: Finished STEP 3480/50000, loss = 7.384990 (0.074 sec/batch), lr: 0.003000
613
+ 2025-12-03 22:00:43 INFO: Finished STEP 3500/50000, loss = 4.654266 (0.066 sec/batch), lr: 0.003000
614
+ 2025-12-03 22:00:43 INFO: Evaluating on dev set...
615
+ 2025-12-03 22:00:43 INFO: LAS MLAS BLEX
616
+ 2025-12-03 22:00:43 INFO: 57.92 51.28 53.42
617
+ 2025-12-03 22:00:43 INFO: step 3500: train_loss = 4.370241, dev_score = 0.5792
618
+ 2025-12-03 22:00:44 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
619
+ 2025-12-03 22:00:44 INFO: new model checkpoint saved.
620
+ 2025-12-03 22:00:45 INFO: Finished STEP 3520/50000, loss = 5.403185 (0.067 sec/batch), lr: 0.003000
621
+ 2025-12-03 22:00:47 INFO: Finished STEP 3540/50000, loss = 3.413640 (0.065 sec/batch), lr: 0.003000
622
+ 2025-12-03 22:00:48 INFO: Finished STEP 3560/50000, loss = 4.872176 (0.069 sec/batch), lr: 0.003000
623
+ 2025-12-03 22:00:49 INFO: Finished STEP 3580/50000, loss = 3.841375 (0.069 sec/batch), lr: 0.003000
624
+ 2025-12-03 22:00:51 INFO: Finished STEP 3600/50000, loss = 3.843983 (0.071 sec/batch), lr: 0.003000
625
+ 2025-12-03 22:00:51 INFO: Evaluating on dev set...
626
+ 2025-12-03 22:00:51 INFO: LAS MLAS BLEX
627
+ 2025-12-03 22:00:51 INFO: 57.43 50.32 53.28
628
+ 2025-12-03 22:00:51 INFO: step 3600: train_loss = 4.379443, dev_score = 0.5743
629
+ 2025-12-03 22:00:52 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
630
+ 2025-12-03 22:00:52 INFO: new model checkpoint saved.
631
+ 2025-12-03 22:00:53 INFO: Finished STEP 3620/50000, loss = 4.895670 (0.064 sec/batch), lr: 0.003000
632
+ 2025-12-03 22:00:55 INFO: Finished STEP 3640/50000, loss = 4.107656 (0.068 sec/batch), lr: 0.003000
633
+ 2025-12-03 22:00:56 INFO: Finished STEP 3660/50000, loss = 3.756389 (0.062 sec/batch), lr: 0.003000
634
+ 2025-12-03 22:00:58 INFO: Finished STEP 3680/50000, loss = 3.594301 (0.075 sec/batch), lr: 0.003000
635
+ 2025-12-03 22:00:59 INFO: Finished STEP 3700/50000, loss = 4.018555 (0.073 sec/batch), lr: 0.003000
636
+ 2025-12-03 22:00:59 INFO: Evaluating on dev set...
637
+ 2025-12-03 22:00:59 INFO: LAS MLAS BLEX
638
+ 2025-12-03 22:00:59 INFO: 58.42 48.83 56.05
639
+ 2025-12-03 22:00:59 INFO: step 3700: train_loss = 4.141694, dev_score = 0.5842
640
+ 2025-12-03 22:00:59 INFO: Training ended with 3700 steps.
641
+ 2025-12-03 22:00:59 INFO: Best dev F1 = 61.39, at iteration = 2700
642
+ 2025-12-03 22:01:01 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--batch_size', '32', '--dropout', '0.33']
643
+ 2025-12-03 22:01:01 INFO: Running parser in predict mode
644
+ 2025-12-03 22:01:01 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
645
+ 2025-12-03 22:01:04 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
646
+ 2025-12-03 22:01:04 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
647
+ 2025-12-03 22:01:04 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
648
+ 2025-12-03 22:01:04 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
649
+ 2025-12-03 22:01:04 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
650
+ 2025-12-03 22:01:04 INFO: Loading data with batch size 32...
651
+ 2025-12-03 22:01:04 DEBUG: 9 batches created.
652
+ 2025-12-03 22:01:04 INFO: F1 scores for each dependency:
653
+ Note that unlabeled attachment errors hurt the labeled attachment scores
654
+ acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
655
+ acl:relcl: p 0.1667 r 0.1429 f1 0.1538 (7 actual)
656
+ advcl: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
657
+ advmod: p 0.5000 r 0.6000 f1 0.5455 (25 actual)
658
+ amod: p 0.8000 r 0.7742 f1 0.7869 (31 actual)
659
+ appos: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
660
+ aux: p 0.9000 r 0.8182 f1 0.8571 (11 actual)
661
+ case: p 0.8596 r 0.8750 f1 0.8673 (56 actual)
662
+ cc: p 0.7143 r 0.7692 f1 0.7407 (13 actual)
663
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
664
+ conj: p 0.2188 r 0.5833 f1 0.3182 (12 actual)
665
+ cop: p 0.5000 r 0.3333 f1 0.4000 (3 actual)
666
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
667
+ det: p 0.7826 r 0.8182 f1 0.8000 (22 actual)
668
+ expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
669
+ iobj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
670
+ mark: p 0.8182 r 0.7500 f1 0.7826 (12 actual)
671
+ nmod: p 0.4615 r 0.4000 f1 0.4286 (15 actual)
672
+ nmod:poss: p 0.9444 r 0.8947 f1 0.9189 (19 actual)
673
+ nsubj: p 0.4583 r 0.6471 f1 0.5366 (17 actual)
674
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
675
+ obj: p 0.7391 r 0.7727 f1 0.7556 (22 actual)
676
+ obl: p 0.5208 r 0.6098 f1 0.5618 (41 actual)
677
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
678
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
679
+ parataxis: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
680
+ punct: p 0.4231 r 0.4231 f1 0.4231 (52 actual)
681
+ root: p 0.4444 r 0.4444 f1 0.4444 (9 actual)
682
+ xcomp: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
683
+ 2025-12-03 22:01:04 INFO: LAS MLAS BLEX
684
+ 2025-12-03 22:01:04 INFO: 61.39 53.19 55.32
685
+ 2025-12-03 22:01:04 INFO: Parser score:
686
+ 2025-12-03 22:01:04 INFO: sv_diachronic 61.39
687
+ 2025-12-03 22:01:05 INFO: Finished running dev set on
688
+ UD_Swedish-diachronic
689
+ UAS LAS CLAS MLAS BLEX
690
+ 68.81 61.39 55.32 53.19 55.32
691
+ 2025-12-03 22:01:05 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--batch_size', '32', '--dropout', '0.33']
692
+ 2025-12-03 22:01:05 INFO: Running parser in predict mode
693
+ 2025-12-03 22:01:05 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
694
+ 2025-12-03 22:01:07 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
695
+ 2025-12-03 22:01:07 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
696
+ 2025-12-03 22:01:07 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
697
+ 2025-12-03 22:01:07 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
698
+ 2025-12-03 22:01:08 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
699
+ 2025-12-03 22:01:08 INFO: Loading data with batch size 32...
700
+ 2025-12-03 22:01:08 DEBUG: 93 batches created.
701
+ 2025-12-03 22:01:12 INFO: F1 scores for each dependency:
702
+ Note that unlabeled attachment errors hurt the labeled attachment scores
703
+ acl: p 0.0000 r 0.0000 f1 0.0000 (32 actual)
704
+ acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
705
+ acl:relcl: p 0.2653 r 0.1733 f1 0.2097 (75 actual)
706
+ advcl: p 0.0923 r 0.1000 f1 0.0960 (60 actual)
707
+ advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
708
+ advmod: p 0.5579 r 0.5933 f1 0.5750 (268 actual)
709
+ amod: p 0.7639 r 0.7739 f1 0.7689 (230 actual)
710
+ appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
711
+ aux: p 0.8353 r 0.8452 f1 0.8402 (84 actual)
712
+ aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
713
+ case: p 0.8140 r 0.8097 f1 0.8118 (373 actual)
714
+ cc: p 0.6556 r 0.6387 f1 0.6471 (155 actual)
715
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
716
+ compound:prt: p 0.8421 r 0.7619 f1 0.8000 (21 actual)
717
+ conj: p 0.1406 r 0.2848 f1 0.1883 (158 actual)
718
+ cop: p 0.8519 r 0.5000 f1 0.6301 (46 actual)
719
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
720
+ dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
721
+ det: p 0.7664 r 0.7885 f1 0.7773 (208 actual)
722
+ discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
723
+ dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
724
+ expl: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
725
+ expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
726
+ fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
727
+ flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
728
+ flat:name: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
729
+ goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
730
+ iobj: p 0.1667 r 0.2857 f1 0.2105 (14 actual)
731
+ mark: p 0.6772 r 0.6993 f1 0.6881 (153 actual)
732
+ nmod: p 0.5128 r 0.3922 f1 0.4444 (102 actual)
733
+ nmod:poss: p 0.8601 r 0.8662 f1 0.8632 (142 actual)
734
+ nsubj: p 0.5777 r 0.6107 f1 0.5938 (280 actual)
735
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
736
+ nummod: p 0.0000 r 0.0000 f1 0.0000 (10 actual)
737
+ obj: p 0.5810 r 0.5683 f1 0.5746 (183 actual)
738
+ obl: p 0.4519 r 0.6259 f1 0.5249 (278 actual)
739
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
740
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
741
+ parataxis: p 0.0000 r 0.0000 f1 0.0000 (18 actual)
742
+ punct: p 0.4624 r 0.4635 f1 0.4630 (425 actual)
743
+ reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
744
+ root: p 0.5556 r 0.5556 f1 0.5556 (99 actual)
745
+ vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
746
+ xcomp: p 0.6364 r 0.1867 f1 0.2887 (75 actual)
747
+ 2025-12-03 22:01:12 INFO: LAS MLAS BLEX
748
+ 2025-12-03 22:01:12 INFO: 57.41 47.88 51.07
749
+ 2025-12-03 22:01:12 INFO: Parser score:
750
+ 2025-12-03 22:01:12 INFO: sv_diachronic 57.41
751
+ 2025-12-03 22:01:12 INFO: Finished running test set on
752
+ UD_Swedish-diachronic
753
+ UAS LAS CLAS MLAS BLEX
754
+ 66.78 57.41 51.07 47.88 51.07
755
+ DONE.
756
+ Full log saved to: logs/log_diachronic.pt_sv_diachron_is_20251203_214751.txt
757
+ Symlink updated: logs/latest.txt → log_diachronic.pt_sv_diachron_is_20251203_214751.txt
logs/log_diachronic.pt_sv_diachron_nn_20251203_003001.txt CHANGED
@@ -140,3 +140,580 @@ Augmented 291 quotes: Counter({'″″': 35, '„”': 34, '「」': 33, '””
140
  2025-12-03 00:30:18 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
141
  2025-12-03 00:30:18 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
142
  2025-12-03 00:30:21 INFO: Loading data with batch size 250...
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
140
  2025-12-03 00:30:18 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
141
  2025-12-03 00:30:18 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
142
  2025-12-03 00:30:21 INFO: Loading data with batch size 250...
143
+ 2025-12-03 00:31:10 INFO: Start evaluation...
144
+ 2025-12-03 00:35:33 INFO: UPOS XPOS UFeats AllTags
145
+ 2025-12-03 00:35:33 INFO: 87.98 62.42 77.22 61.01
146
+ 2025-12-03 00:35:33 INFO: POS Tagger score: sv_diachronic 61.01
147
+ 2025-12-03 00:35:34 INFO: Running tagger to retag /local/tmp.5437491/tmp6nhkre69/sv_diachronic.dev.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu
148
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5437491/tmp6nhkre69/sv_diachronic.dev.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu']
149
+ 2025-12-03 00:35:34 INFO: Running tagger in predict mode
150
+ 2025-12-03 00:35:34 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
151
+ 2025-12-03 00:35:36 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
152
+ 2025-12-03 00:35:36 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
153
+ 2025-12-03 00:35:36 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
154
+ 2025-12-03 00:35:36 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
155
+ 2025-12-03 00:35:37 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
156
+ 2025-12-03 00:35:37 INFO: Loading data with batch size 250...
157
+ 2025-12-03 00:35:37 INFO: Start evaluation...
158
+ 2025-12-03 00:35:37 INFO: UPOS XPOS UFeats AllTags
159
+ 2025-12-03 00:35:37 INFO: 88.61 79.46 83.42 74.75
160
+ 2025-12-03 00:35:37 INFO: POS Tagger score: sv_diachronic 74.75
161
+ 2025-12-03 00:35:37 INFO: Running tagger to retag /local/tmp.5437491/tmp6nhkre69/sv_diachronic.test.gold.conllu to /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu
162
+ Args: ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--save_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pos', '--save_name', 'diachronic.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--eval_file', '/local/tmp.5437491/tmp6nhkre69/sv_diachronic.test.gold.conllu', '--output_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu']
163
+ 2025-12-03 00:35:37 INFO: Running tagger in predict mode
164
+ 2025-12-03 00:35:37 INFO: Loading model from: /cephyr/users/cleland/Alvis/stanza_resources/sv/pos/diachronic.pt
165
+ 2025-12-03 00:35:40 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
166
+ 2025-12-03 00:35:40 DEBUG: POS model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
167
+ 2025-12-03 00:35:40 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
168
+ 2025-12-03 00:35:40 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
169
+ 2025-12-03 00:35:40 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
170
+ 2025-12-03 00:35:40 INFO: Loading data with batch size 250...
171
+ 2025-12-03 00:35:40 INFO: Start evaluation...
172
+ 2025-12-03 00:35:41 INFO: UPOS XPOS UFeats AllTags
173
+ 2025-12-03 00:35:41 INFO: 89.61 86.64 86.69 81.18
174
+ 2025-12-03 00:35:41 INFO: POS Tagger score: sv_diachronic 81.18
175
+ Preparing data for UD_Swedish-diachronic: sv_diachronic, sv
176
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-train.conllu and writing to /local/tmp.5437491/tmp6nhkre69/sv_diachronic.train.gold.conllu
177
+ Swapped 'w1, w2' for 'w1 ,w2' 287 times
178
+ Added 675 new sentences with asdf, zzzz -> asdf,zzzz
179
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-dev.conllu and writing to /local/tmp.5437491/tmp6nhkre69/sv_diachronic.dev.gold.conllu
180
+ Reading from /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/ud/UD_Swedish-diachronic/sv_diachronic-ud-test.conllu and writing to /local/tmp.5437491/tmp6nhkre69/sv_diachronic.test.gold.conllu
181
+ Running stanza dependency parser training…
182
+ 2025-12-03 00:35:58 INFO: Training program called with:
183
+ /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/stanza/utils/training/run_depparse.py UD_Swedish-diachronic --wordvec_pretrain_file /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt --batch_size 32 --dropout 0.33
184
+ 2025-12-03 00:35:58 DEBUG: UD_Swedish-diachronic: sv_diachronic
185
+ 2025-12-03 00:35:58 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
186
+ 2025-12-03 00:35:58 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
187
+ 2025-12-03 00:35:58 INFO: UD_Swedish-diachronic: saved_models/depparse/sv_diachronic_charlm_parser.pt does not exist, training new model
188
+ 2025-12-03 00:35:58 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt for forward charlm
189
+ 2025-12-03 00:35:58 INFO: Using model /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt for backward charlm
190
+ 2025-12-03 00:35:58 INFO: Running train depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--train_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--batch_size', '5000', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'train', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--batch_size', '32', '--dropout', '0.33']
191
+ 2025-12-03 00:35:58 INFO: Running parser in train mode
192
+ 2025-12-03 00:35:58 INFO: Using pretrained contextualized char embedding
193
+ 2025-12-03 00:35:58 INFO: Loading data with batch size 32...
194
+ 2025-12-03 00:36:07 INFO: Train File /mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.train.in.conllu, Data Size: 84502
195
+ 2025-12-03 00:36:07 INFO: Original data size: 84502
196
+ 2025-12-03 00:36:08 INFO: Augmented data size: 84708
197
+ 2025-12-03 00:36:31 WARNING: sv_diachronic is not a known dataset. Examining the data to choose which xpos vocab to use
198
+ 2025-12-03 00:36:31 INFO: Original length = 84708
199
+ 2025-12-03 00:36:31 INFO: Filtered length = 84708
200
+ 2025-12-03 00:36:50 WARNING: Chose XPOSDescription(xpos_type=<XPOSType.XPOS: 1>, sep='|') for the xpos factory for sv_diachronic
201
+ 2025-12-03 00:37:00 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
202
+ 2025-12-03 00:37:18 DEBUG: 50996 batches created.
203
+ 2025-12-03 00:37:18 DEBUG: 9 batches created.
204
+ 2025-12-03 00:37:18 INFO: Training parser...
205
+ 2025-12-03 00:37:18 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
206
+ 2025-12-03 00:37:18 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
207
+ 2025-12-03 00:37:18 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
208
+ 2025-12-03 00:37:18 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
209
+ 2025-12-03 00:37:22 INFO: Finished STEP 20/50000, loss = 0.000000 (0.032 sec/batch), lr: 0.003000
210
+ 2025-12-03 00:37:23 INFO: Finished STEP 40/50000, loss = 0.000000 (0.033 sec/batch), lr: 0.003000
211
+ 2025-12-03 00:37:24 INFO: Finished STEP 60/50000, loss = 0.000000 (0.033 sec/batch), lr: 0.003000
212
+ 2025-12-03 00:37:24 INFO: Finished STEP 80/50000, loss = 0.000000 (0.033 sec/batch), lr: 0.003000
213
+ 2025-12-03 00:37:25 INFO: Finished STEP 100/50000, loss = 2.869145 (0.033 sec/batch), lr: 0.003000
214
+ 2025-12-03 00:37:25 INFO: Evaluating on dev set...
215
+ 2025-12-03 00:37:26 INFO: LAS MLAS BLEX
216
+ 2025-12-03 00:37:26 INFO: 0.50 0.31 0.63
217
+ 2025-12-03 00:37:26 INFO: step 100: train_loss = 3721.223041, dev_score = 0.0050
218
+ 2025-12-03 00:37:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
219
+ 2025-12-03 00:37:26 INFO: new best model saved.
220
+ 2025-12-03 00:37:26 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
221
+ 2025-12-03 00:37:26 INFO: new model checkpoint saved.
222
+ 2025-12-03 00:37:27 INFO: Finished STEP 120/50000, loss = 1.472011 (0.032 sec/batch), lr: 0.003000
223
+ 2025-12-03 00:37:28 INFO: Finished STEP 140/50000, loss = 5173.653809 (0.033 sec/batch), lr: 0.003000
224
+ 2025-12-03 00:37:28 INFO: Finished STEP 160/50000, loss = 1.405938 (0.032 sec/batch), lr: 0.003000
225
+ 2025-12-03 00:37:29 INFO: Finished STEP 180/50000, loss = 0.777264 (0.033 sec/batch), lr: 0.003000
226
+ 2025-12-03 00:37:30 INFO: Finished STEP 200/50000, loss = 0.921184 (0.032 sec/batch), lr: 0.003000
227
+ 2025-12-03 00:37:30 INFO: Evaluating on dev set...
228
+ 2025-12-03 00:37:30 INFO: LAS MLAS BLEX
229
+ 2025-12-03 00:37:30 INFO: 2.97 1.44 1.80
230
+ 2025-12-03 00:37:30 INFO: step 200: train_loss = 53.301930, dev_score = 0.0297
231
+ 2025-12-03 00:37:30 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
232
+ 2025-12-03 00:37:30 INFO: new best model saved.
233
+ 2025-12-03 00:37:31 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
234
+ 2025-12-03 00:37:31 INFO: new model checkpoint saved.
235
+ 2025-12-03 00:37:32 INFO: Finished STEP 220/50000, loss = 1.159031 (0.032 sec/batch), lr: 0.003000
236
+ 2025-12-03 00:37:32 INFO: Finished STEP 240/50000, loss = 0.857107 (0.032 sec/batch), lr: 0.003000
237
+ 2025-12-03 00:37:33 INFO: Finished STEP 260/50000, loss = 1.147544 (0.033 sec/batch), lr: 0.003000
238
+ 2025-12-03 00:37:34 INFO: Finished STEP 280/50000, loss = 0.642448 (0.032 sec/batch), lr: 0.003000
239
+ 2025-12-03 00:37:34 INFO: Finished STEP 300/50000, loss = 0.877519 (0.032 sec/batch), lr: 0.003000
240
+ 2025-12-03 00:37:34 INFO: Evaluating on dev set...
241
+ 2025-12-03 00:37:35 INFO: LAS MLAS BLEX
242
+ 2025-12-03 00:37:35 INFO: 2.23 1.33 1.66
243
+ 2025-12-03 00:37:35 INFO: step 300: train_loss = 0.990781, dev_score = 0.0223
244
+ 2025-12-03 00:37:35 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
245
+ 2025-12-03 00:37:35 INFO: new model checkpoint saved.
246
+ 2025-12-03 00:37:36 INFO: Finished STEP 320/50000, loss = 4.604774 (0.033 sec/batch), lr: 0.003000
247
+ 2025-12-03 00:37:37 INFO: Finished STEP 340/50000, loss = 2.761129 (0.033 sec/batch), lr: 0.003000
248
+ 2025-12-03 00:37:37 INFO: Finished STEP 360/50000, loss = 1.883921 (0.033 sec/batch), lr: 0.003000
249
+ 2025-12-03 00:37:38 INFO: Finished STEP 380/50000, loss = 2.024679 (0.033 sec/batch), lr: 0.003000
250
+ 2025-12-03 00:37:39 INFO: Finished STEP 400/50000, loss = 3.201016 (0.033 sec/batch), lr: 0.003000
251
+ 2025-12-03 00:37:39 INFO: Evaluating on dev set...
252
+ 2025-12-03 00:37:40 INFO: LAS MLAS BLEX
253
+ 2025-12-03 00:37:40 INFO: 8.17 0.65 1.95
254
+ 2025-12-03 00:37:40 INFO: step 400: train_loss = 2.039984, dev_score = 0.0817
255
+ 2025-12-03 00:37:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
256
+ 2025-12-03 00:37:40 INFO: new best model saved.
257
+ 2025-12-03 00:37:40 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
258
+ 2025-12-03 00:37:40 INFO: new model checkpoint saved.
259
+ 2025-12-03 00:37:41 INFO: Finished STEP 420/50000, loss = 2.318334 (0.032 sec/batch), lr: 0.003000
260
+ 2025-12-03 00:37:42 INFO: Finished STEP 440/50000, loss = 2.347792 (0.033 sec/batch), lr: 0.003000
261
+ 2025-12-03 00:37:42 INFO: Finished STEP 460/50000, loss = 1.795001 (0.033 sec/batch), lr: 0.003000
262
+ 2025-12-03 00:37:43 INFO: Finished STEP 480/50000, loss = 2.403748 (0.038 sec/batch), lr: 0.003000
263
+ 2025-12-03 00:37:44 INFO: Finished STEP 500/50000, loss = 1.886348 (0.037 sec/batch), lr: 0.003000
264
+ 2025-12-03 00:37:44 INFO: Evaluating on dev set...
265
+ 2025-12-03 00:37:45 INFO: LAS MLAS BLEX
266
+ 2025-12-03 00:37:45 INFO: 6.68 0.35 3.50
267
+ 2025-12-03 00:37:45 INFO: step 500: train_loss = 1.796768, dev_score = 0.0668
268
+ 2025-12-03 00:37:45 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
269
+ 2025-12-03 00:37:45 INFO: new model checkpoint saved.
270
+ 2025-12-03 00:37:46 INFO: Finished STEP 520/50000, loss = 1.742396 (0.033 sec/batch), lr: 0.003000
271
+ 2025-12-03 00:37:47 INFO: Finished STEP 540/50000, loss = 1.576722 (0.038 sec/batch), lr: 0.003000
272
+ 2025-12-03 00:37:47 INFO: Finished STEP 560/50000, loss = 1.409795 (0.033 sec/batch), lr: 0.003000
273
+ 2025-12-03 00:37:48 INFO: Finished STEP 580/50000, loss = 1.341886 (0.033 sec/batch), lr: 0.003000
274
+ 2025-12-03 00:37:49 INFO: Finished STEP 600/50000, loss = 1.618135 (0.034 sec/batch), lr: 0.003000
275
+ 2025-12-03 00:37:49 INFO: Evaluating on dev set...
276
+ 2025-12-03 00:37:49 INFO: LAS MLAS BLEX
277
+ 2025-12-03 00:37:49 INFO: 3.96 0.43 0.87
278
+ 2025-12-03 00:37:49 INFO: step 600: train_loss = 1.672703, dev_score = 0.0396
279
+ 2025-12-03 00:37:50 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
280
+ 2025-12-03 00:37:50 INFO: new model checkpoint saved.
281
+ 2025-12-03 00:37:50 INFO: Finished STEP 620/50000, loss = 1.120005 (0.032 sec/batch), lr: 0.003000
282
+ 2025-12-03 00:37:51 INFO: Finished STEP 640/50000, loss = 1.549862 (0.034 sec/batch), lr: 0.003000
283
+ 2025-12-03 00:37:52 INFO: Finished STEP 660/50000, loss = 0.910717 (0.033 sec/batch), lr: 0.003000
284
+ 2025-12-03 00:37:52 INFO: Finished STEP 680/50000, loss = 3.636672 (0.035 sec/batch), lr: 0.003000
285
+ 2025-12-03 00:37:53 INFO: Finished STEP 700/50000, loss = 2.700411 (0.037 sec/batch), lr: 0.003000
286
+ 2025-12-03 00:37:53 INFO: Evaluating on dev set...
287
+ 2025-12-03 00:37:54 INFO: LAS MLAS BLEX
288
+ 2025-12-03 00:37:54 INFO: 14.60 0.84 1.68
289
+ 2025-12-03 00:37:54 INFO: step 700: train_loss = 1.952326, dev_score = 0.1460
290
+ 2025-12-03 00:37:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
291
+ 2025-12-03 00:37:54 INFO: new best model saved.
292
+ 2025-12-03 00:37:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
293
+ 2025-12-03 00:37:54 INFO: new model checkpoint saved.
294
+ 2025-12-03 00:37:55 INFO: Finished STEP 720/50000, loss = 2.404730 (0.034 sec/batch), lr: 0.003000
295
+ 2025-12-03 00:37:56 INFO: Finished STEP 740/50000, loss = 2.597665 (0.035 sec/batch), lr: 0.003000
296
+ 2025-12-03 00:37:57 INFO: Finished STEP 760/50000, loss = 2.103006 (0.032 sec/batch), lr: 0.003000
297
+ 2025-12-03 00:37:57 INFO: Finished STEP 780/50000, loss = 2.304466 (0.034 sec/batch), lr: 0.003000
298
+ 2025-12-03 00:37:58 INFO: Finished STEP 800/50000, loss = 2.542548 (0.034 sec/batch), lr: 0.003000
299
+ 2025-12-03 00:37:58 INFO: Evaluating on dev set...
300
+ 2025-12-03 00:37:59 INFO: LAS MLAS BLEX
301
+ 2025-12-03 00:37:59 INFO: 14.60 4.79 6.26
302
+ 2025-12-03 00:37:59 INFO: step 800: train_loss = 2.437953, dev_score = 0.1460
303
+ 2025-12-03 00:37:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
304
+ 2025-12-03 00:37:59 INFO: new best model saved.
305
+ 2025-12-03 00:37:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
306
+ 2025-12-03 00:37:59 INFO: new model checkpoint saved.
307
+ 2025-12-03 00:38:00 INFO: Finished STEP 820/50000, loss = 2.790083 (0.035 sec/batch), lr: 0.003000
308
+ 2025-12-03 00:38:01 INFO: Finished STEP 840/50000, loss = 2.320239 (0.035 sec/batch), lr: 0.003000
309
+ 2025-12-03 00:38:02 INFO: Finished STEP 860/50000, loss = 1.994328 (0.034 sec/batch), lr: 0.003000
310
+ 2025-12-03 00:38:02 INFO: Finished STEP 880/50000, loss = 2.787316 (0.035 sec/batch), lr: 0.003000
311
+ 2025-12-03 00:38:03 INFO: Finished STEP 900/50000, loss = 2.700784 (0.034 sec/batch), lr: 0.003000
312
+ 2025-12-03 00:38:03 INFO: Evaluating on dev set...
313
+ 2025-12-03 00:38:04 INFO: LAS MLAS BLEX
314
+ 2025-12-03 00:38:04 INFO: 18.32 3.85 4.62
315
+ 2025-12-03 00:38:04 INFO: step 900: train_loss = 2.455402, dev_score = 0.1832
316
+ 2025-12-03 00:38:04 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
317
+ 2025-12-03 00:38:04 INFO: new best model saved.
318
+ 2025-12-03 00:38:04 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
319
+ 2025-12-03 00:38:04 INFO: new model checkpoint saved.
320
+ 2025-12-03 00:38:05 INFO: Finished STEP 920/50000, loss = 2.396327 (0.035 sec/batch), lr: 0.003000
321
+ 2025-12-03 00:38:06 INFO: Finished STEP 940/50000, loss = 1.992622 (0.034 sec/batch), lr: 0.003000
322
+ 2025-12-03 00:38:07 INFO: Finished STEP 960/50000, loss = 2.259057 (0.034 sec/batch), lr: 0.003000
323
+ 2025-12-03 00:38:07 INFO: Finished STEP 980/50000, loss = 2.550621 (0.034 sec/batch), lr: 0.003000
324
+ 2025-12-03 00:38:08 INFO: Finished STEP 1000/50000, loss = 2.292823 (0.034 sec/batch), lr: 0.003000
325
+ 2025-12-03 00:38:08 INFO: Evaluating on dev set...
326
+ 2025-12-03 00:38:09 INFO: LAS MLAS BLEX
327
+ 2025-12-03 00:38:09 INFO: 23.51 5.73 7.36
328
+ 2025-12-03 00:38:09 INFO: step 1000: train_loss = 2.272285, dev_score = 0.2351
329
+ 2025-12-03 00:38:09 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
330
+ 2025-12-03 00:38:09 INFO: new best model saved.
331
+ 2025-12-03 00:38:10 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
332
+ 2025-12-03 00:38:10 INFO: new model checkpoint saved.
333
+ 2025-12-03 00:38:10 INFO: Finished STEP 1020/50000, loss = 1.907463 (0.036 sec/batch), lr: 0.003000
334
+ 2025-12-03 00:38:11 INFO: Finished STEP 1040/50000, loss = 2.597835 (0.034 sec/batch), lr: 0.003000
335
+ 2025-12-03 00:38:12 INFO: Finished STEP 1060/50000, loss = 2.362857 (0.035 sec/batch), lr: 0.003000
336
+ 2025-12-03 00:38:12 INFO: Finished STEP 1080/50000, loss = 2.444307 (0.035 sec/batch), lr: 0.003000
337
+ 2025-12-03 00:38:13 INFO: Finished STEP 1100/50000, loss = 2.979832 (0.036 sec/batch), lr: 0.003000
338
+ 2025-12-03 00:38:13 INFO: Evaluating on dev set...
339
+ 2025-12-03 00:38:14 INFO: LAS MLAS BLEX
340
+ 2025-12-03 00:38:14 INFO: 25.74 6.15 8.20
341
+ 2025-12-03 00:38:14 INFO: step 1100: train_loss = 2.328531, dev_score = 0.2574
342
+ 2025-12-03 00:38:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
343
+ 2025-12-03 00:38:14 INFO: new best model saved.
344
+ 2025-12-03 00:38:14 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
345
+ 2025-12-03 00:38:14 INFO: new model checkpoint saved.
346
+ 2025-12-03 00:38:15 INFO: Finished STEP 1120/50000, loss = 1.734434 (0.033 sec/batch), lr: 0.003000
347
+ 2025-12-03 00:38:16 INFO: Finished STEP 1140/50000, loss = 2.508640 (0.034 sec/batch), lr: 0.003000
348
+ 2025-12-03 00:38:17 INFO: Finished STEP 1160/50000, loss = 2.418795 (0.035 sec/batch), lr: 0.003000
349
+ 2025-12-03 00:38:17 INFO: Finished STEP 1180/50000, loss = 2.594333 (0.037 sec/batch), lr: 0.003000
350
+ 2025-12-03 00:38:18 INFO: Finished STEP 1200/50000, loss = 1.854763 (0.035 sec/batch), lr: 0.003000
351
+ 2025-12-03 00:38:18 INFO: Evaluating on dev set...
352
+ 2025-12-03 00:38:19 INFO: LAS MLAS BLEX
353
+ 2025-12-03 00:38:19 INFO: 26.49 9.31 11.74
354
+ 2025-12-03 00:38:19 INFO: step 1200: train_loss = 2.502941, dev_score = 0.2649
355
+ 2025-12-03 00:38:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
356
+ 2025-12-03 00:38:19 INFO: new best model saved.
357
+ 2025-12-03 00:38:19 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
358
+ 2025-12-03 00:38:19 INFO: new model checkpoint saved.
359
+ 2025-12-03 00:38:20 INFO: Finished STEP 1220/50000, loss = 3.290319 (0.035 sec/batch), lr: 0.003000
360
+ 2025-12-03 00:38:21 INFO: Finished STEP 1240/50000, loss = 2.376419 (0.036 sec/batch), lr: 0.003000
361
+ 2025-12-03 00:38:22 INFO: Finished STEP 1260/50000, loss = 2.190226 (0.038 sec/batch), lr: 0.003000
362
+ 2025-12-03 00:38:22 INFO: Finished STEP 1280/50000, loss = 2.514861 (0.034 sec/batch), lr: 0.003000
363
+ 2025-12-03 00:38:23 INFO: Finished STEP 1300/50000, loss = 3.613056 (0.035 sec/batch), lr: 0.003000
364
+ 2025-12-03 00:38:23 INFO: Evaluating on dev set...
365
+ 2025-12-03 00:38:24 INFO: LAS MLAS BLEX
366
+ 2025-12-03 00:38:24 INFO: 27.72 12.24 14.29
367
+ 2025-12-03 00:38:24 INFO: step 1300: train_loss = 2.848149, dev_score = 0.2772
368
+ 2025-12-03 00:38:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
369
+ 2025-12-03 00:38:24 INFO: new best model saved.
370
+ 2025-12-03 00:38:24 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
371
+ 2025-12-03 00:38:24 INFO: new model checkpoint saved.
372
+ 2025-12-03 00:38:25 INFO: Finished STEP 1320/50000, loss = 2.592073 (0.039 sec/batch), lr: 0.003000
373
+ 2025-12-03 00:38:26 INFO: Finished STEP 1340/50000, loss = 1.771206 (0.035 sec/batch), lr: 0.003000
374
+ 2025-12-03 00:38:27 INFO: Finished STEP 1360/50000, loss = 2.209909 (0.034 sec/batch), lr: 0.003000
375
+ 2025-12-03 00:38:27 INFO: Finished STEP 1380/50000, loss = 2.989327 (0.036 sec/batch), lr: 0.003000
376
+ 2025-12-03 00:38:28 INFO: Finished STEP 1400/50000, loss = 2.774464 (0.034 sec/batch), lr: 0.003000
377
+ 2025-12-03 00:38:28 INFO: Evaluating on dev set...
378
+ 2025-12-03 00:38:29 INFO: LAS MLAS BLEX
379
+ 2025-12-03 00:38:29 INFO: 27.72 11.09 12.32
380
+ 2025-12-03 00:38:29 INFO: step 1400: train_loss = 2.722048, dev_score = 0.2772
381
+ 2025-12-03 00:38:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
382
+ 2025-12-03 00:38:29 INFO: new best model saved.
383
+ 2025-12-03 00:38:29 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
384
+ 2025-12-03 00:38:29 INFO: new model checkpoint saved.
385
+ 2025-12-03 00:38:30 INFO: Finished STEP 1420/50000, loss = 3.454481 (0.035 sec/batch), lr: 0.003000
386
+ 2025-12-03 00:38:31 INFO: Finished STEP 1440/50000, loss = 2.674970 (0.035 sec/batch), lr: 0.003000
387
+ 2025-12-03 00:38:32 INFO: Finished STEP 1460/50000, loss = 2.481898 (0.034 sec/batch), lr: 0.003000
388
+ 2025-12-03 00:38:32 INFO: Finished STEP 1480/50000, loss = 3.442808 (0.035 sec/batch), lr: 0.003000
389
+ 2025-12-03 00:38:33 INFO: Finished STEP 1500/50000, loss = 3.347362 (0.034 sec/batch), lr: 0.003000
390
+ 2025-12-03 00:38:33 INFO: Evaluating on dev set...
391
+ 2025-12-03 00:38:34 INFO: LAS MLAS BLEX
392
+ 2025-12-03 00:38:34 INFO: 25.99 14.31 15.86
393
+ 2025-12-03 00:38:34 INFO: step 1500: train_loss = 2.798998, dev_score = 0.2599
394
+ 2025-12-03 00:38:34 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
395
+ 2025-12-03 00:38:34 INFO: new model checkpoint saved.
396
+ 2025-12-03 00:38:35 INFO: Finished STEP 1520/50000, loss = 2.220331 (0.035 sec/batch), lr: 0.003000
397
+ 2025-12-03 00:38:36 INFO: Finished STEP 1540/50000, loss = 2.878744 (0.034 sec/batch), lr: 0.003000
398
+ 2025-12-03 00:38:36 INFO: Finished STEP 1560/50000, loss = 2.658022 (0.037 sec/batch), lr: 0.003000
399
+ 2025-12-03 00:38:37 INFO: Finished STEP 1580/50000, loss = 2.898341 (0.036 sec/batch), lr: 0.003000
400
+ 2025-12-03 00:38:38 INFO: Finished STEP 1600/50000, loss = 2.236073 (0.035 sec/batch), lr: 0.003000
401
+ 2025-12-03 00:38:38 INFO: Evaluating on dev set...
402
+ 2025-12-03 00:38:39 INFO: LAS MLAS BLEX
403
+ 2025-12-03 00:38:39 INFO: 31.68 17.54 19.21
404
+ 2025-12-03 00:38:39 INFO: step 1600: train_loss = 2.792263, dev_score = 0.3168
405
+ 2025-12-03 00:38:39 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
406
+ 2025-12-03 00:38:39 INFO: new best model saved.
407
+ 2025-12-03 00:38:39 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
408
+ 2025-12-03 00:38:39 INFO: new model checkpoint saved.
409
+ 2025-12-03 00:38:40 INFO: Finished STEP 1620/50000, loss = 3.108658 (0.035 sec/batch), lr: 0.003000
410
+ 2025-12-03 00:38:41 INFO: Finished STEP 1640/50000, loss = 2.404290 (0.035 sec/batch), lr: 0.003000
411
+ 2025-12-03 00:38:42 INFO: Finished STEP 1660/50000, loss = 2.833586 (0.034 sec/batch), lr: 0.003000
412
+ 2025-12-03 00:38:42 INFO: Finished STEP 1680/50000, loss = 2.168653 (0.034 sec/batch), lr: 0.003000
413
+ 2025-12-03 00:38:43 INFO: Finished STEP 1700/50000, loss = 2.581872 (0.035 sec/batch), lr: 0.003000
414
+ 2025-12-03 00:38:43 INFO: Evaluating on dev set...
415
+ 2025-12-03 00:38:44 INFO: LAS MLAS BLEX
416
+ 2025-12-03 00:38:44 INFO: 30.69 16.49 17.76
417
+ 2025-12-03 00:38:44 INFO: step 1700: train_loss = 2.642489, dev_score = 0.3069
418
+ 2025-12-03 00:38:44 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
419
+ 2025-12-03 00:38:44 INFO: new model checkpoint saved.
420
+ 2025-12-03 00:38:45 INFO: Finished STEP 1720/50000, loss = 2.661027 (0.036 sec/batch), lr: 0.003000
421
+ 2025-12-03 00:38:46 INFO: Finished STEP 1740/50000, loss = 1.947067 (0.035 sec/batch), lr: 0.003000
422
+ 2025-12-03 00:38:46 INFO: Finished STEP 1760/50000, loss = 2.557057 (0.037 sec/batch), lr: 0.003000
423
+ 2025-12-03 00:38:47 INFO: Finished STEP 1780/50000, loss = 2.333129 (0.034 sec/batch), lr: 0.003000
424
+ 2025-12-03 00:38:48 INFO: Finished STEP 1800/50000, loss = 2.481168 (0.034 sec/batch), lr: 0.003000
425
+ 2025-12-03 00:38:48 INFO: Evaluating on dev set...
426
+ 2025-12-03 00:38:48 INFO: LAS MLAS BLEX
427
+ 2025-12-03 00:38:48 INFO: 34.16 17.11 18.33
428
+ 2025-12-03 00:38:48 INFO: step 1800: train_loss = 2.576569, dev_score = 0.3416
429
+ 2025-12-03 00:38:49 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
430
+ 2025-12-03 00:38:49 INFO: new best model saved.
431
+ 2025-12-03 00:38:49 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
432
+ 2025-12-03 00:38:49 INFO: new model checkpoint saved.
433
+ 2025-12-03 00:38:50 INFO: Finished STEP 1820/50000, loss = 3.247551 (0.036 sec/batch), lr: 0.003000
434
+ 2025-12-03 00:38:51 INFO: Finished STEP 1840/50000, loss = 2.598015 (0.036 sec/batch), lr: 0.003000
435
+ 2025-12-03 00:38:51 INFO: Finished STEP 1860/50000, loss = 2.779014 (0.034 sec/batch), lr: 0.003000
436
+ 2025-12-03 00:38:52 INFO: Finished STEP 1880/50000, loss = 3.907140 (0.036 sec/batch), lr: 0.003000
437
+ 2025-12-03 00:38:53 INFO: Finished STEP 1900/50000, loss = 2.739502 (0.034 sec/batch), lr: 0.003000
438
+ 2025-12-03 00:38:53 INFO: Evaluating on dev set...
439
+ 2025-12-03 00:38:53 INFO: LAS MLAS BLEX
440
+ 2025-12-03 00:38:53 INFO: 34.90 21.25 23.75
441
+ 2025-12-03 00:38:53 INFO: step 1900: train_loss = 2.978315, dev_score = 0.3490
442
+ 2025-12-03 00:38:53 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
443
+ 2025-12-03 00:38:53 INFO: new best model saved.
444
+ 2025-12-03 00:38:54 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
445
+ 2025-12-03 00:38:54 INFO: new model checkpoint saved.
446
+ 2025-12-03 00:38:55 INFO: Finished STEP 1920/50000, loss = 2.824293 (0.035 sec/batch), lr: 0.003000
447
+ 2025-12-03 00:38:56 INFO: Finished STEP 1940/50000, loss = 3.076251 (0.035 sec/batch), lr: 0.003000
448
+ 2025-12-03 00:38:56 INFO: Finished STEP 1960/50000, loss = 2.556714 (0.034 sec/batch), lr: 0.003000
449
+ 2025-12-03 00:38:57 INFO: Finished STEP 1980/50000, loss = 3.574326 (0.037 sec/batch), lr: 0.003000
450
+ 2025-12-03 00:38:58 INFO: Finished STEP 2000/50000, loss = 3.166797 (0.036 sec/batch), lr: 0.003000
451
+ 2025-12-03 00:38:58 INFO: Evaluating on dev set...
452
+ 2025-12-03 00:38:58 INFO: LAS MLAS BLEX
453
+ 2025-12-03 00:38:58 INFO: 41.34 25.67 28.57
454
+ 2025-12-03 00:38:58 INFO: step 2000: train_loss = 2.939474, dev_score = 0.4134
455
+ 2025-12-03 00:38:58 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
456
+ 2025-12-03 00:38:58 INFO: new best model saved.
457
+ 2025-12-03 00:38:59 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
458
+ 2025-12-03 00:38:59 INFO: new model checkpoint saved.
459
+ 2025-12-03 00:39:00 INFO: Finished STEP 2020/50000, loss = 2.706799 (0.036 sec/batch), lr: 0.003000
460
+ 2025-12-03 00:39:00 INFO: Finished STEP 2040/50000, loss = 2.405847 (0.036 sec/batch), lr: 0.003000
461
+ 2025-12-03 00:39:01 INFO: Finished STEP 2060/50000, loss = 3.432140 (0.035 sec/batch), lr: 0.003000
462
+ 2025-12-03 00:39:02 INFO: Finished STEP 2080/50000, loss = 3.064786 (0.035 sec/batch), lr: 0.003000
463
+ 2025-12-03 00:39:03 INFO: Finished STEP 2100/50000, loss = 2.427642 (0.035 sec/batch), lr: 0.003000
464
+ 2025-12-03 00:39:03 INFO: Evaluating on dev set...
465
+ 2025-12-03 00:39:03 INFO: LAS MLAS BLEX
466
+ 2025-12-03 00:39:03 INFO: 35.15 20.65 22.37
467
+ 2025-12-03 00:39:03 INFO: step 2100: train_loss = 3.057586, dev_score = 0.3515
468
+ 2025-12-03 00:39:04 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
469
+ 2025-12-03 00:39:04 INFO: new model checkpoint saved.
470
+ 2025-12-03 00:39:04 INFO: Finished STEP 2120/50000, loss = 4.900553 (0.035 sec/batch), lr: 0.003000
471
+ 2025-12-03 00:39:05 INFO: Finished STEP 2140/50000, loss = 3.500584 (0.037 sec/batch), lr: 0.003000
472
+ 2025-12-03 00:39:06 INFO: Finished STEP 2160/50000, loss = 2.639127 (0.036 sec/batch), lr: 0.003000
473
+ 2025-12-03 00:39:07 INFO: Finished STEP 2180/50000, loss = 2.110754 (0.037 sec/batch), lr: 0.003000
474
+ 2025-12-03 00:39:07 INFO: Finished STEP 2200/50000, loss = 4.028395 (0.035 sec/batch), lr: 0.003000
475
+ 2025-12-03 00:39:07 INFO: Evaluating on dev set...
476
+ 2025-12-03 00:39:08 INFO: LAS MLAS BLEX
477
+ 2025-12-03 00:39:08 INFO: 34.16 23.29 26.10
478
+ 2025-12-03 00:39:08 INFO: step 2200: train_loss = 2.984515, dev_score = 0.3416
479
+ 2025-12-03 00:39:08 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
480
+ 2025-12-03 00:39:08 INFO: new model checkpoint saved.
481
+ 2025-12-03 00:39:09 INFO: Finished STEP 2220/50000, loss = 2.188152 (0.037 sec/batch), lr: 0.003000
482
+ 2025-12-03 00:39:10 INFO: Finished STEP 2240/50000, loss = 2.980751 (0.035 sec/batch), lr: 0.003000
483
+ 2025-12-03 00:39:11 INFO: Finished STEP 2260/50000, loss = 2.396068 (0.036 sec/batch), lr: 0.003000
484
+ 2025-12-03 00:39:11 INFO: Finished STEP 2280/50000, loss = 2.121481 (0.036 sec/batch), lr: 0.003000
485
+ 2025-12-03 00:39:12 INFO: Finished STEP 2300/50000, loss = 2.107288 (0.035 sec/batch), lr: 0.003000
486
+ 2025-12-03 00:39:12 INFO: Evaluating on dev set...
487
+ 2025-12-03 00:39:13 INFO: LAS MLAS BLEX
488
+ 2025-12-03 00:39:13 INFO: 42.33 26.05 29.83
489
+ 2025-12-03 00:39:13 INFO: step 2300: train_loss = 2.953812, dev_score = 0.4233
490
+ 2025-12-03 00:39:13 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser.pt
491
+ 2025-12-03 00:39:13 INFO: new best model saved.
492
+ 2025-12-03 00:39:13 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
493
+ 2025-12-03 00:39:13 INFO: new model checkpoint saved.
494
+ 2025-12-03 00:39:14 INFO: Finished STEP 2320/50000, loss = 3.074659 (0.034 sec/batch), lr: 0.003000
495
+ 2025-12-03 00:39:15 INFO: Finished STEP 2340/50000, loss = 2.844486 (0.035 sec/batch), lr: 0.003000
496
+ 2025-12-03 00:39:16 INFO: Finished STEP 2360/50000, loss = 1.841562 (0.035 sec/batch), lr: 0.003000
497
+ 2025-12-03 00:39:16 INFO: Finished STEP 2380/50000, loss = 3.023133 (0.035 sec/batch), lr: 0.003000
498
+ 2025-12-03 00:39:17 INFO: Finished STEP 2400/50000, loss = 2.847883 (0.034 sec/batch), lr: 0.003000
499
+ 2025-12-03 00:39:17 INFO: Evaluating on dev set...
500
+ 2025-12-03 00:39:17 INFO: LAS MLAS BLEX
501
+ 2025-12-03 00:39:17 INFO: 36.14 21.43 23.95
502
+ 2025-12-03 00:39:17 INFO: step 2400: train_loss = 3.008838, dev_score = 0.3614
503
+ 2025-12-03 00:39:18 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
504
+ 2025-12-03 00:39:18 INFO: new model checkpoint saved.
505
+ 2025-12-03 00:39:19 INFO: Finished STEP 2420/50000, loss = 2.635043 (0.034 sec/batch), lr: 0.003000
506
+ 2025-12-03 00:39:20 INFO: Finished STEP 2440/50000, loss = 2.504678 (0.036 sec/batch), lr: 0.003000
507
+ 2025-12-03 00:39:20 INFO: Finished STEP 2460/50000, loss = 2.440725 (0.035 sec/batch), lr: 0.003000
508
+ 2025-12-03 00:39:21 INFO: Finished STEP 2480/50000, loss = 2.583835 (0.036 sec/batch), lr: 0.003000
509
+ 2025-12-03 00:39:22 INFO: Finished STEP 2500/50000, loss = 2.080993 (0.034 sec/batch), lr: 0.003000
510
+ 2025-12-03 00:39:22 INFO: Evaluating on dev set...
511
+ 2025-12-03 00:39:22 INFO: LAS MLAS BLEX
512
+ 2025-12-03 00:39:22 INFO: 33.66 19.83 23.14
513
+ 2025-12-03 00:39:22 INFO: step 2500: train_loss = 2.975335, dev_score = 0.3366
514
+ 2025-12-03 00:39:23 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
515
+ 2025-12-03 00:39:23 INFO: new model checkpoint saved.
516
+ 2025-12-03 00:39:23 INFO: Finished STEP 2520/50000, loss = 3.098147 (0.035 sec/batch), lr: 0.003000
517
+ 2025-12-03 00:39:24 INFO: Finished STEP 2540/50000, loss = 2.705004 (0.034 sec/batch), lr: 0.003000
518
+ 2025-12-03 00:39:25 INFO: Finished STEP 2560/50000, loss = 2.812203 (0.035 sec/batch), lr: 0.003000
519
+ 2025-12-03 00:39:26 INFO: Finished STEP 2580/50000, loss = 2.498108 (0.035 sec/batch), lr: 0.003000
520
+ 2025-12-03 00:39:26 INFO: Finished STEP 2600/50000, loss = 2.534167 (0.036 sec/batch), lr: 0.003000
521
+ 2025-12-03 00:39:26 INFO: Evaluating on dev set...
522
+ 2025-12-03 00:39:27 INFO: LAS MLAS BLEX
523
+ 2025-12-03 00:39:27 INFO: 35.40 23.95 26.47
524
+ 2025-12-03 00:39:27 INFO: step 2600: train_loss = 2.867347, dev_score = 0.3540
525
+ 2025-12-03 00:39:27 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
526
+ 2025-12-03 00:39:27 INFO: new model checkpoint saved.
527
+ 2025-12-03 00:39:28 INFO: Finished STEP 2620/50000, loss = 1.976379 (0.036 sec/batch), lr: 0.003000
528
+ 2025-12-03 00:39:29 INFO: Finished STEP 2640/50000, loss = 4.797091 (0.035 sec/batch), lr: 0.003000
529
+ 2025-12-03 00:39:30 INFO: Finished STEP 2660/50000, loss = 2.363913 (0.034 sec/batch), lr: 0.003000
530
+ 2025-12-03 00:39:30 INFO: Finished STEP 2680/50000, loss = 1.629785 (0.035 sec/batch), lr: 0.003000
531
+ 2025-12-03 00:39:31 INFO: Finished STEP 2700/50000, loss = 2.052314 (0.036 sec/batch), lr: 0.003000
532
+ 2025-12-03 00:39:31 INFO: Evaluating on dev set...
533
+ 2025-12-03 00:39:32 INFO: LAS MLAS BLEX
534
+ 2025-12-03 00:39:32 INFO: 34.16 23.26 26.64
535
+ 2025-12-03 00:39:32 INFO: step 2700: train_loss = 2.857439, dev_score = 0.3416
536
+ 2025-12-03 00:39:32 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
537
+ 2025-12-03 00:39:32 INFO: new model checkpoint saved.
538
+ 2025-12-03 00:39:33 INFO: Finished STEP 2720/50000, loss = 2.368320 (0.036 sec/batch), lr: 0.003000
539
+ 2025-12-03 00:39:34 INFO: Finished STEP 2740/50000, loss = 4.389025 (0.037 sec/batch), lr: 0.003000
540
+ 2025-12-03 00:39:34 INFO: Finished STEP 2760/50000, loss = 3.079744 (0.035 sec/batch), lr: 0.003000
541
+ 2025-12-03 00:39:35 INFO: Finished STEP 2780/50000, loss = 2.286925 (0.035 sec/batch), lr: 0.003000
542
+ 2025-12-03 00:39:36 INFO: Finished STEP 2800/50000, loss = 2.271642 (0.034 sec/batch), lr: 0.003000
543
+ 2025-12-03 00:39:36 INFO: Evaluating on dev set...
544
+ 2025-12-03 00:39:36 INFO: LAS MLAS BLEX
545
+ 2025-12-03 00:39:36 INFO: 38.37 22.87 26.20
546
+ 2025-12-03 00:39:36 INFO: step 2800: train_loss = 3.000191, dev_score = 0.3837
547
+ 2025-12-03 00:39:37 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
548
+ 2025-12-03 00:39:37 INFO: new model checkpoint saved.
549
+ 2025-12-03 00:39:38 INFO: Finished STEP 2820/50000, loss = 4.063986 (0.035 sec/batch), lr: 0.003000
550
+ 2025-12-03 00:39:38 INFO: Finished STEP 2840/50000, loss = 2.836863 (0.036 sec/batch), lr: 0.003000
551
+ 2025-12-03 00:39:39 INFO: Finished STEP 2860/50000, loss = 2.274727 (0.034 sec/batch), lr: 0.003000
552
+ 2025-12-03 00:39:40 INFO: Finished STEP 2880/50000, loss = 2.046604 (0.034 sec/batch), lr: 0.003000
553
+ 2025-12-03 00:39:41 INFO: Finished STEP 2900/50000, loss = 3.225588 (0.034 sec/batch), lr: 0.003000
554
+ 2025-12-03 00:39:41 INFO: Evaluating on dev set...
555
+ 2025-12-03 00:39:41 INFO: LAS MLAS BLEX
556
+ 2025-12-03 00:39:41 INFO: 40.84 25.83 28.75
557
+ 2025-12-03 00:39:41 INFO: step 2900: train_loss = 2.866668, dev_score = 0.4084
558
+ 2025-12-03 00:39:42 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
559
+ 2025-12-03 00:39:42 INFO: new model checkpoint saved.
560
+ 2025-12-03 00:39:42 INFO: Finished STEP 2920/50000, loss = 2.608736 (0.035 sec/batch), lr: 0.003000
561
+ 2025-12-03 00:39:43 INFO: Finished STEP 2940/50000, loss = 3.828312 (0.036 sec/batch), lr: 0.003000
562
+ 2025-12-03 00:39:44 INFO: Finished STEP 2960/50000, loss = 4.258009 (0.035 sec/batch), lr: 0.003000
563
+ 2025-12-03 00:39:45 INFO: Finished STEP 2980/50000, loss = 2.643698 (0.035 sec/batch), lr: 0.003000
564
+ 2025-12-03 00:39:45 INFO: Finished STEP 3000/50000, loss = 3.561427 (0.036 sec/batch), lr: 0.003000
565
+ 2025-12-03 00:39:45 INFO: Evaluating on dev set...
566
+ 2025-12-03 00:39:46 INFO: LAS MLAS BLEX
567
+ 2025-12-03 00:39:46 INFO: 39.85 28.33 32.56
568
+ 2025-12-03 00:39:46 INFO: step 3000: train_loss = 3.231410, dev_score = 0.3985
569
+ 2025-12-03 00:39:47 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
570
+ 2025-12-03 00:39:47 INFO: new model checkpoint saved.
571
+ 2025-12-03 00:39:47 INFO: Finished STEP 3020/50000, loss = 2.717970 (0.040 sec/batch), lr: 0.003000
572
+ 2025-12-03 00:39:48 INFO: Finished STEP 3040/50000, loss = 3.476188 (0.036 sec/batch), lr: 0.003000
573
+ 2025-12-03 00:39:49 INFO: Finished STEP 3060/50000, loss = 3.224195 (0.038 sec/batch), lr: 0.003000
574
+ 2025-12-03 00:39:50 INFO: Finished STEP 3080/50000, loss = 3.752515 (0.036 sec/batch), lr: 0.003000
575
+ 2025-12-03 00:39:50 INFO: Finished STEP 3100/50000, loss = 3.824567 (0.036 sec/batch), lr: 0.003000
576
+ 2025-12-03 00:39:50 INFO: Evaluating on dev set...
577
+ 2025-12-03 00:39:51 INFO: LAS MLAS BLEX
578
+ 2025-12-03 00:39:51 INFO: 40.10 27.31 31.09
579
+ 2025-12-03 00:39:51 INFO: step 3100: train_loss = 3.311538, dev_score = 0.4010
580
+ 2025-12-03 00:39:51 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
581
+ 2025-12-03 00:39:51 INFO: new model checkpoint saved.
582
+ 2025-12-03 00:39:52 INFO: Finished STEP 3120/50000, loss = 3.155392 (0.037 sec/batch), lr: 0.003000
583
+ 2025-12-03 00:39:53 INFO: Finished STEP 3140/50000, loss = 2.654355 (0.035 sec/batch), lr: 0.003000
584
+ 2025-12-03 00:39:54 INFO: Finished STEP 3160/50000, loss = 4.152210 (0.040 sec/batch), lr: 0.003000
585
+ 2025-12-03 00:39:54 INFO: Finished STEP 3180/50000, loss = 3.310773 (0.036 sec/batch), lr: 0.003000
586
+ 2025-12-03 00:39:55 INFO: Finished STEP 3200/50000, loss = 4.537406 (0.036 sec/batch), lr: 0.003000
587
+ 2025-12-03 00:39:55 INFO: Evaluating on dev set...
588
+ 2025-12-03 00:39:56 INFO: LAS MLAS BLEX
589
+ 2025-12-03 00:39:56 INFO: 41.09 26.67 29.17
590
+ 2025-12-03 00:39:56 INFO: step 3200: train_loss = 3.171517, dev_score = 0.4109
591
+ 2025-12-03 00:39:56 INFO: Model saved to saved_models/depparse/sv_diachronic_charlm_parser_checkpoint.pt
592
+ 2025-12-03 00:39:56 INFO: new model checkpoint saved.
593
+ 2025-12-03 00:39:57 INFO: Finished STEP 3220/50000, loss = 3.289989 (0.037 sec/batch), lr: 0.003000
594
+ 2025-12-03 00:39:58 INFO: Finished STEP 3240/50000, loss = 3.196197 (0.037 sec/batch), lr: 0.003000
595
+ 2025-12-03 00:39:58 INFO: Finished STEP 3260/50000, loss = 1.874357 (0.039 sec/batch), lr: 0.003000
596
+ 2025-12-03 00:39:59 INFO: Finished STEP 3280/50000, loss = 2.109042 (0.036 sec/batch), lr: 0.003000
597
+ 2025-12-03 00:40:00 INFO: Finished STEP 3300/50000, loss = 1.826021 (0.036 sec/batch), lr: 0.003000
598
+ 2025-12-03 00:40:00 INFO: Evaluating on dev set...
599
+ 2025-12-03 00:40:01 INFO: LAS MLAS BLEX
600
+ 2025-12-03 00:40:01 INFO: 40.35 27.85 30.38
601
+ 2025-12-03 00:40:01 INFO: step 3300: train_loss = 2.779594, dev_score = 0.4035
602
+ 2025-12-03 00:40:01 INFO: Training ended with 3300 steps.
603
+ 2025-12-03 00:40:01 INFO: Best dev F1 = 42.33, at iteration = 2300
604
+ 2025-12-03 00:40:02 INFO: Running dev depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.dev.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--batch_size', '32', '--dropout', '0.33']
605
+ 2025-12-03 00:40:02 INFO: Running parser in predict mode
606
+ 2025-12-03 00:40:02 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
607
+ 2025-12-03 00:40:05 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
608
+ 2025-12-03 00:40:05 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
609
+ 2025-12-03 00:40:05 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
610
+ 2025-12-03 00:40:05 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
611
+ 2025-12-03 00:40:05 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
612
+ 2025-12-03 00:40:05 INFO: Loading data with batch size 32...
613
+ 2025-12-03 00:40:05 DEBUG: 9 batches created.
614
+ 2025-12-03 00:40:06 INFO: F1 scores for each dependency:
615
+ Note that unlabeled attachment errors hurt the labeled attachment scores
616
+ acl: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
617
+ acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
618
+ advcl: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
619
+ advmod: p 0.6000 r 0.6000 f1 0.6000 (25 actual)
620
+ amod: p 0.7000 r 0.6774 f1 0.6885 (31 actual)
621
+ appos: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
622
+ aux: p 0.6154 r 0.7273 f1 0.6667 (11 actual)
623
+ case: p 0.8136 r 0.8571 f1 0.8348 (56 actual)
624
+ cc: p 0.5833 r 0.5385 f1 0.5600 (13 actual)
625
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
626
+ conj: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
627
+ cop: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
628
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
629
+ det: p 0.7273 r 0.7273 f1 0.7273 (22 actual)
630
+ expl: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
631
+ iobj: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
632
+ mark: p 0.2000 r 0.0833 f1 0.1176 (12 actual)
633
+ nmod: p 0.0000 r 0.0000 f1 0.0000 (15 actual)
634
+ nmod:poss: p 0.8235 r 0.7368 f1 0.7778 (19 actual)
635
+ nsubj: p 0.0604 r 0.5294 f1 0.1084 (17 actual)
636
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
637
+ obj: p 0.0000 r 0.0000 f1 0.0000 (22 actual)
638
+ obl: p 0.3636 r 0.0976 f1 0.1538 (41 actual)
639
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
640
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
641
+ parataxis: p 0.0000 r 0.0000 f1 0.0000 (3 actual)
642
+ punct: p 0.3846 r 0.3846 f1 0.3846 (52 actual)
643
+ root: p 0.5556 r 0.5556 f1 0.5556 (9 actual)
644
+ xcomp: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
645
+ 2025-12-03 00:40:06 INFO: LAS MLAS BLEX
646
+ 2025-12-03 00:40:06 INFO: 42.33 26.05 29.83
647
+ 2025-12-03 00:40:06 INFO: Parser score:
648
+ 2025-12-03 00:40:06 INFO: sv_diachronic 42.33
649
+ 2025-12-03 00:40:06 INFO: Finished running dev set on
650
+ UD_Swedish-diachronic
651
+ UAS LAS CLAS MLAS BLEX
652
+ 57.43 42.33 29.83 26.05 29.83
653
+ 2025-12-03 00:40:06 INFO: Running test depparse for UD_Swedish-diachronic with args ['--wordvec_dir', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain', '--eval_file', '/mimer/NOBACKUP/groups/dionysus/cleland/stanza-digphil/data/depparse/sv_diachronic.test.in.conllu', '--lang', 'sv', '--shorthand', 'sv_diachronic', '--mode', 'predict', '--charlm', '--charlm_shorthand', 'sv_conll17', '--charlm_forward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt', '--charlm_backward_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt', '--wordvec_pretrain_file', '/cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt', '--batch_size', '32', '--dropout', '0.33']
654
+ 2025-12-03 00:40:06 INFO: Running parser in predict mode
655
+ 2025-12-03 00:40:06 INFO: Loading model from: saved_models/depparse/sv_diachronic_charlm_parser.pt
656
+ 2025-12-03 00:40:09 DEBUG: Loaded pretrain from /cephyr/users/cleland/Alvis/stanza_resources/sv/pretrain/diachronic.pt
657
+ 2025-12-03 00:40:09 DEBUG: Depparse model loading charmodels: /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt and /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
658
+ 2025-12-03 00:40:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/forward_charlm/conll17.pt
659
+ 2025-12-03 00:40:09 DEBUG: Loading charlm from /cephyr/users/cleland/Alvis/stanza_resources/sv/backward_charlm/conll17.pt
660
+ 2025-12-03 00:40:09 DEBUG: Building Adam with lr=0.003000, betas=(0.9, 0.95), eps=0.000001
661
+ 2025-12-03 00:40:09 INFO: Loading data with batch size 32...
662
+ 2025-12-03 00:40:09 DEBUG: 93 batches created.
663
+ 2025-12-03 00:40:14 INFO: F1 scores for each dependency:
664
+ Note that unlabeled attachment errors hurt the labeled attachment scores
665
+ acl: p 0.0000 r 0.0000 f1 0.0000 (32 actual)
666
+ acl:cleft: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
667
+ acl:relcl: p 0.0000 r 0.0000 f1 0.0000 (75 actual)
668
+ advcl: p 0.0000 r 0.0000 f1 0.0000 (60 actual)
669
+ advcl:relcl: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
670
+ advmod: p 0.5227 r 0.5149 f1 0.5188 (268 actual)
671
+ amod: p 0.6515 r 0.6826 f1 0.6667 (230 actual)
672
+ appos: p 0.0000 r 0.0000 f1 0.0000 (13 actual)
673
+ aux: p 0.5877 r 0.7976 f1 0.6768 (84 actual)
674
+ aux:pass: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
675
+ case: p 0.6162 r 0.7748 f1 0.6865 (373 actual)
676
+ cc: p 0.5036 r 0.4516 f1 0.4762 (155 actual)
677
+ ccomp: p 0.0000 r 0.0000 f1 0.0000 (35 actual)
678
+ compound:prt: p 0.0000 r 0.0000 f1 0.0000 (21 actual)
679
+ conj: p 0.0000 r 0.0000 f1 0.0000 (158 actual)
680
+ cop: p 1.0000 r 0.0217 f1 0.0426 (46 actual)
681
+ csubj: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
682
+ dep: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
683
+ det: p 0.8146 r 0.8029 f1 0.8087 (208 actual)
684
+ discourse: p 0.0000 r 0.0000 f1 0.0000 (7 actual)
685
+ dislocated: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
686
+ expl: p 0.0000 r 0.0000 f1 0.0000 (11 actual)
687
+ expl:pv: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
688
+ fixed: p 0.0000 r 0.0000 f1 0.0000 (8 actual)
689
+ flat: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
690
+ flat:name: p 0.0000 r 0.0000 f1 0.0000 (12 actual)
691
+ goeswith: p 0.0000 r 0.0000 f1 0.0000 (2 actual)
692
+ iobj: p 0.0000 r 0.0000 f1 0.0000 (14 actual)
693
+ mark: p 0.4255 r 0.1307 f1 0.2000 (153 actual)
694
+ nmod: p 0.0000 r 0.0000 f1 0.0000 (102 actual)
695
+ nmod:poss: p 0.6319 r 0.8099 f1 0.7099 (142 actual)
696
+ nsubj: p 0.1188 r 0.5607 f1 0.1960 (280 actual)
697
+ nsubj:pass: p 0.0000 r 0.0000 f1 0.0000 (25 actual)
698
+ nummod: p 0.0000 r 0.0000 f1 0.0000 (10 actual)
699
+ obj: p 0.0000 r 0.0000 f1 0.0000 (183 actual)
700
+ obl: p 0.2909 r 0.1151 f1 0.1649 (278 actual)
701
+ obl:agent: p 0.0000 r 0.0000 f1 0.0000 (4 actual)
702
+ orphan: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
703
+ parataxis: p 0.0000 r 0.0000 f1 0.0000 (18 actual)
704
+ punct: p 0.3647 r 0.3647 f1 0.3647 (425 actual)
705
+ reparandum: p 0.0000 r 0.0000 f1 0.0000 (1 actual)
706
+ root: p 0.3838 r 0.3838 f1 0.3838 (99 actual)
707
+ vocative: p 0.0000 r 0.0000 f1 0.0000 (5 actual)
708
+ xcomp: p 0.0000 r 0.0000 f1 0.0000 (75 actual)
709
+ 2025-12-03 00:40:14 INFO: LAS MLAS BLEX
710
+ 2025-12-03 00:40:14 INFO: 39.12 26.14 29.41
711
+ 2025-12-03 00:40:14 INFO: Parser score:
712
+ 2025-12-03 00:40:14 INFO: sv_diachronic 39.12
713
+ 2025-12-03 00:40:14 INFO: Finished running test set on
714
+ UD_Swedish-diachronic
715
+ UAS LAS CLAS MLAS BLEX
716
+ 56.56 39.12 29.41 26.14 29.41
717
+ DONE.
718
+ Full log saved to: logs/log_diachronic.pt_sv_diachron_nn_20251203_003001.txt
719
+ Symlink updated: logs/latest.txt → log_diachronic.pt_sv_diachron_nn_20251203_003001.txt
saved_models/depparse/conll17_de_lit/sv_diachronic_charlm_parser.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b84862d0f728ba53e2076208a94d5e90bd1dc7e0f2884a12db3f21ffcaf073dc
3
+ size 146969172
saved_models/depparse/conll17_de_lit/sv_diachronic_charlm_parser_checkpoint.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5303a079697eaf9aa75caffc06be07b9b0acb1aaa1153af103a0b74adf847207
3
+ size 439729196
saved_models/depparse/conll17_icepahc/sv_diachronic_charlm_parser.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f59e82a5e94514bdf086d19931a26f5cd6274c6ad29c667f3366e93be0bec585
3
+ size 151645170
saved_models/depparse/conll17_icepahc/sv_diachronic_charlm_parser_checkpoint.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d2ef405f2a3a2fc6b376a51c7a93a5d5a9b9a186a8b0bce1e79dd5ab9c2e5572
3
+ size 453029994
saved_models/depparse/conll17_isPUD-pahc-gc/sv_diachronic_charlm_parser.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5fe5f0e97a6a4e257b124f8b229c59cc61d58124512421d56795ac6ffb67359e
3
+ size 146969174
saved_models/depparse/conll17_isPUD-pahc-gc/sv_diachronic_charlm_parser_checkpoint.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1742a1e312cb0260f04d6aaf52cda85e277d6a2d3f342d182dc486a871cbe932
3
+ size 439729194
saved_models/depparse/conll17_isPUD/sv_diachronic_charlm_parser.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dcdc23474d269b8f4dba7d246f184b42a446cae7294f1134121cf5da8511efa6
3
+ size 146117067
saved_models/depparse/conll17_isPUD/sv_diachronic_charlm_parser_checkpoint.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:66d8f413e99482f3b539677f46a83b30302e52d94fad185fdacb21a92c0f12bb
3
+ size 437206479
saved_models/depparse/conll17_nn/sv_diachronic_charlm_parser.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d74ed335831d36ee6d4569d1496a62860779c378ed391a693a8aa09159f388e
3
+ size 147986243
saved_models/depparse/conll17_nn/sv_diachronic_charlm_parser_checkpoint.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c816bf81174ee44f8bc937d29e25b5b650e9c387c9f6352f25337a2c4c11552b
3
+ size 442521655
saved_models/depparse/kubhist2-isPUD/sv_diachronic_charlm_parser.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a7fac12b102f25be203e20d2867a091faecb8f09dfe9ac5d9c616edc6e7f645f
3
+ size 146127839
saved_models/depparse/kubhist2-isPUD/sv_diachronic_charlm_parser_checkpoint.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b5647ebca89cffce00d03cb931c459ce61156734205ebccb25e8b421cf18462
3
+ size 437234658
saved_models/depparse/kubhist2_de_lit/sv_diachronic_charlm_parser.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:71163ccb4e2658d56c77323d9bb5a8b62fad59b62c9860b38e04d84361f427c1
3
+ size 146975967
saved_models/depparse/kubhist2_de_lit/sv_diachronic_charlm_parser_checkpoint.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25a5bb319d850ad88f27b4b5e8609604a11ba686da46a01ca17f180ca9d2d4ef
3
+ size 439746396
saved_models/depparse/kubhist2_is-modern/sv_diachronic_charlm_parser.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:335e4c614f54f5925012a6db8aad316b469f99abe4f0945375026a37fb4fe44e
3
+ size 148247396
saved_models/depparse/kubhist2_is-modern/sv_diachronic_charlm_parser_checkpoint.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:80a6097e3b359f248fbe75ba42974a9f47f243da51e1a6c261bdd5b73ba00e33
3
+ size 443361619
saved_models/depparse/kubhist2_nn/sv_diachronic_charlm_parser.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da6fcca1488acc8616fa50b8f41f91208d4fe2a0a09bef0366c772078e096093
3
+ size 147994017
saved_models/depparse/kubhist2_nn/sv_diachronic_charlm_parser_checkpoint.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ee738b46e517df3291da336189e3aa8578eda2aa67991aea3f76399ac09d40c3
3
+ size 442542038
ud-treebanks-is/is_gc-ud-dev.conllu ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:82bc5d1acabd7bea13503675ae4b4c2b76aa2167573b3509889db00fd1ce984b
3
+ size 972038
ud-treebanks-is/is_gc-ud-test.conllu ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d7593fba5a12600417e392aafdb81cbe1ff588f86f5ed1ad0cf0f17b4bd0e0da
3
+ size 936991
ud-treebanks-is/is_gc-ud-train.conllu ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:622205dfee16f16b3cc7cdd2e62f5c6d52c68ffcbe2733ec627c85cc56258fd9
3
+ size 7135578
ud-treebanks-is/is_modern-ud-dev.conllu ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69f65bfeec6467af1875ef2520d9b2b2749815dcc11e6ddadaa3eb3c31c4f51b
3
+ size 717538
ud-treebanks-is/is_modern-ud-test.conllu ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d746fdc1c4ea1aea83e9136293aca2b57028e8fbb43229d3d57ba5cb4fcbc412
3
+ size 852084
ud-treebanks-is/is_modern-ud-train.conllu ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd8f3f9aeaef70a46b392392739375a25ea1638550087c34081bfcca43a8484e
3
+ size 5277378