Upload RenderSwitch Polish 67 evals to multiview-renderswitch-encoder-only/polish67-original-teacher-forced

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/manifest.json

@@ -0,0 +1,66 @@
+{
+  "checkpoint_count": 12,
+  "checkpoints": [
+    {
+      "label": "epoch_0",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_0.ckpt"
+    },
+    {
+      "label": "epoch_1",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_1.ckpt"
+    },
+    {
+      "label": "epoch_2",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_2.ckpt"
+    },
+    {
+      "label": "epoch_3",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_3.ckpt"
+    },
+    {
+      "label": "epoch_4",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_4.ckpt"
+    },
+    {
+      "label": "epoch_5",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_5.ckpt"
+    },
+    {
+      "label": "epoch_6",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_6.ckpt"
+    },
+    {
+      "label": "epoch_7",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_7.ckpt"
+    },
+    {
+      "label": "official59m",
+      "path": "/gpfs/scrubbed/qxli2/rescore/data/checkpoints/transcoda_59m_zeroshot_v1/transcoda-59M-zeroshot-v1.lightning-compatible.ckpt"
+    },
+    {
+      "label": "convnext2_300k_8ep",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-300k-8ep-modal-h200-4gpu/model.ckpt"
+    },
+    {
+      "label": "musvit_243k_8ep",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-synth-243k-8ep-musvit-frozen-explicit-modal-h200-4gpu/model.ckpt"
+    },
+    {
+      "label": "full_renderswitch_epoch6",
+      "path": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-unfrozen-8ep-h200x4/checkpoint_epoch_6.ckpt"
+    }
+  ],
+  "completed_count": 12,
+  "dataset": "cminst/transcoda-polish-filtered-eval",
+  "elapsed_seconds": 291.43735003471375,
+  "failure_count": 0,
+  "failures": [],
+  "image_column": "image_original_real",
+  "mode": "teacher_forced",
+  "output_dir": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced",
+  "row_count": 12,
+  "run_dir": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4",
+  "skipped_count": 0,
+  "split": "eval",
+  "transcription_column": "transcription"
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/metrics.csv

@@ -0,0 +1,13 @@
+checkpoint,mode,cer_percent,ser_percent,ler_percent,top1_error_rate,top1_errors,top1_errors_per_page,pages,tokens,skipped_existing,checkpoint_path,summary
+epoch_0,teacher_forced,9.614958763122559,10.225929260253906,42.33919143676758,0.10007381064215258,7457,111.29850746268657,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_0.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/summary.json
+epoch_1,teacher_forced,9.535521507263184,10.14130687713623,42.10032272338867,0.0992283432865866,7394,110.35820895522389,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_1.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/summary.json
+epoch_2,teacher_forced,9.481289863586426,10.14264965057373,42.202693939208984,0.09925518352009663,7396,110.38805970149254,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_2.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/summary.json
+epoch_3,teacher_forced,9.608084678649902,10.15070915222168,42.074729919433594,0.09933570422062672,7402,110.4776119402985,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_3.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/summary.json
+epoch_4,teacher_forced,9.550797462463379,10.130560874938965,42.2453498840332,0.0991478225860565,7388,110.26865671641791,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_4.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/summary.json
+epoch_5,teacher_forced,9.54239559173584,10.075489044189453,41.878517150878906,0.09858417768234584,7346,109.64179104477611,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_5.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_5/summary.json
+epoch_6,teacher_forced,9.598917961120605,10.117128372192383,41.86145782470703,0.09901362141850634,7378,110.11940298507463,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_6.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_6/summary.json
+epoch_7,teacher_forced,9.52482795715332,10.06877326965332,41.451969146728516,0.09855733744883581,7344,109.61194029850746,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_7.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_7/summary.json
+official59m,teacher_forced,9.20326042175293,11.272297859191895,41.59699630737305,0.0984768167483057,7338,109.5223880597015,67,74515,False,/gpfs/scrubbed/qxli2/rescore/data/checkpoints/transcoda_59m_zeroshot_v1/transcoda-59M-zeroshot-v1.lightning-compatible.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/official59m/summary.json
+convnext2_300k_8ep,teacher_forced,14.556106567382812,15.089727401733398,53.25029754638672,0.1396497349526941,10406,155.3134328358209,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-300k-8ep-modal-h200-4gpu/model.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/summary.json
+musvit_243k_8ep,teacher_forced,9.905973434448242,14.091714859008789,45.02644729614258,0.10283835469368584,7663,114.3731343283582,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-synth-243k-8ep-musvit-frozen-explicit-modal-h200-4gpu/model.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/musvit_243k_8ep/summary.json
+full_renderswitch_epoch6,teacher_forced,10.870677947998047,11.076187133789062,44.3951530456543,0.10868952559887271,8099,120.88059701492537,67,74515,False,/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-unfrozen-8ep-h200x4/checkpoint_epoch_6.ckpt,/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/full_renderswitch_epoch6/summary.json

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/command.log

@@ -0,0 +1,512 @@
+Skipping import of cpp extensions due to incompatible torch version 2.11.0+cu129 for torchao version 0.15.0             Please see https://github.com/pytorch/ao/issues/2919 for more info
+
+[transformers] ConvNextV2Model LOAD REPORT from: facebook/convnextv2-tiny-22k-224
+Key               | Status     |  | 
+------------------+------------+--+-
+classifier.bias   | UNEXPECTED |  | 
+classifier.weight | UNEXPECTED |  | 
+
+Notes:
+- UNEXPECTED:	can be ignored when loading from different task/architecture; not ok if you expect identical arch.
+2026-08-31 09:26:07.475 | INFO     | src.model.encoder:_load_transformers:81 - Loaded transformers encoder: facebook/convnextv2-tiny-22k-224, output_dim=768
+2026-08-31 09:26:07.475 | INFO     | src.model.encoder:load:49 - Loaded encoder 'facebook/convnextv2-tiny-22k-224' via transformers with output dimension 768
+2026-08-31 09:26:07.484 | WARNING  | src.model.modeling_transcoda:__init__:1041 - Encoder facebook/convnextv2-tiny-22k-224 does not support gradient checkpointing
+[eval] pages=32/67 mode=teacher_forced elapsed=12.4s
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-300k-8ep-modal-h200-4gpu/model.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.19623294253315396,
+      "errors": 2042,
+      "errors_per_1k_tokens": 27.4038784137422,
+      "errors_per_page_mean": 30.47761194029851,
+      "errors_per_page_median": 31.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.1619258120315203,
+      "errors": 1685,
+      "errors_per_1k_tokens": 22.61289673220157,
+      "errors_per_page_mean": 25.149253731343283,
+      "errors_per_page_median": 25.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.10676532769556026,
+      "errors": 1111,
+      "errors_per_1k_tokens": 14.909749714822519,
+      "errors_per_page_mean": 16.582089552238806,
+      "errors_per_page_median": 14.0,
+      "page_share": 0.9701492537313433,
+      "pages_with_error": 65
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.08062656159907745,
+      "errors": 839,
+      "errors_per_1k_tokens": 11.25947795745823,
+      "errors_per_page_mean": 12.522388059701493,
+      "errors_per_page_median": 7.0,
+      "page_share": 0.9701492537313433,
+      "pages_with_error": 65
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.06217566788391313,
+      "errors": 647,
+      "errors_per_1k_tokens": 8.682815540495202,
+      "errors_per_page_mean": 9.656716417910447,
+      "errors_per_page_median": 9.0,
+      "page_share": 0.9850746268656716,
+      "pages_with_error": 66
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.06102248702671536,
+      "errors": 635,
+      "errors_per_1k_tokens": 8.521774139435013,
+      "errors_per_page_mean": 9.477611940298507,
+      "errors_per_page_median": 9.0,
+      "page_share": 0.9850746268656716,
+      "pages_with_error": 66
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05362290985969633,
+      "errors": 558,
+      "errors_per_1k_tokens": 7.4884251492987985,
+      "errors_per_page_mean": 8.328358208955224,
+      "errors_per_page_median": 7.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.03478762252546608,
+      "errors": 362,
+      "errors_per_1k_tokens": 4.8580822653157085,
+      "errors_per_page_mean": 5.402985074626866,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.9552238805970149,
+      "pages_with_error": 64
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.0331539496444359,
+      "errors": 345,
+      "errors_per_1k_tokens": 4.62994028048044,
+      "errors_per_page_mean": 5.149253731343284,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.8507462686567164,
+      "pages_with_error": 57
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.032673457620603495,
+      "errors": 340,
+      "errors_per_1k_tokens": 4.562839696705361,
+      "errors_per_page_mean": 5.074626865671642,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.8059701492537313,
+      "pages_with_error": 54
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.023448010763021335,
+      "errors": 244,
+      "errors_per_1k_tokens": 3.2745084882238475,
+      "errors_per_page_mean": 3.6417910447761193,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.6865671641791045,
+      "pages_with_error": 46
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.019411877762829138,
+      "errors": 202,
+      "errors_per_1k_tokens": 2.710863584513185,
+      "errors_per_page_mean": 3.014925373134328,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.8059701492537313,
+      "pages_with_error": 54
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.017874303286565443,
+      "errors": 186,
+      "errors_per_1k_tokens": 2.496141716432933,
+      "errors_per_page_mean": 2.7761194029850746,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.7313432835820896,
+      "pages_with_error": 49
+    },
+    {
+      "category": "barline",
+      "error_share": 0.016336728810301748,
+      "errors": 170,
+      "errors_per_1k_tokens": 2.2814198483526806,
+      "errors_per_page_mean": 2.537313432835821,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.8507462686567164,
+      "pages_with_error": 57
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.013645973476840285,
+      "errors": 142,
+      "errors_per_1k_tokens": 1.905656579212239,
+      "errors_per_page_mean": 2.1194029850746268,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6865671641791045,
+      "pages_with_error": 46
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.011147414952911782,
+      "errors": 116,
+      "errors_per_1k_tokens": 1.5567335435818292,
+      "errors_per_page_mean": 1.7313432835820894,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5970149253731343,
+      "pages_with_error": 40
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.00999423409571401,
+      "errors": 104,
+      "errors_per_1k_tokens": 1.39569214252164,
+      "errors_per_page_mean": 1.5522388059701493,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.7313432835820896,
+      "pages_with_error": 49
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.00989813569094753,
+      "errors": 103,
+      "errors_per_1k_tokens": 1.3822720257666241,
+      "errors_per_page_mean": 1.537313432835821,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5223880597014925,
+      "pages_with_error": 35
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.008072266000384394,
+      "errors": 84,
+      "errors_per_1k_tokens": 1.1272898074213247,
+      "errors_per_page_mean": 1.2537313432835822,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5373134328358209,
+      "pages_with_error": 36
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.007207380357486066,
+      "errors": 75,
+      "errors_per_1k_tokens": 1.0065087566261826,
+      "errors_per_page_mean": 1.1194029850746268,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.44776119402985076,
+      "pages_with_error": 30
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.0065346915241206996,
+      "errors": 68,
+      "errors_per_1k_tokens": 0.9125679393410723,
+      "errors_per_page_mean": 1.0149253731343284,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5671641791044776,
+      "pages_with_error": 38
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005573707476455891,
+      "errors": 58,
+      "errors_per_1k_tokens": 0.7783667717909146,
+      "errors_per_page_mean": 0.8656716417910447,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.8208955223880597,
+      "pages_with_error": 55
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004708821833557563,
+      "errors": 49,
+      "errors_per_1k_tokens": 0.6575857209957726,
+      "errors_per_page_mean": 0.7313432835820896,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3582089552238806,
+      "pages_with_error": 24
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.00432442821449164,
+      "errors": 45,
+      "errors_per_1k_tokens": 0.6039052539757096,
+      "errors_per_page_mean": 0.6716417910447762,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.31343283582089554,
+      "pages_with_error": 21
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.003747837785892754,
+      "errors": 39,
+      "errors_per_1k_tokens": 0.523384553445615,
+      "errors_per_page_mean": 0.582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.29850746268656714,
+      "pages_with_error": 20
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.0028829521429944265,
+      "errors": 30,
+      "errors_per_1k_tokens": 0.40260350265047307,
+      "errors_per_page_mean": 0.44776119402985076,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.16417910447761194,
+      "pages_with_error": 11
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0027868537382279455,
+      "errors": 29,
+      "errors_per_1k_tokens": 0.3891833858954573,
+      "errors_per_page_mean": 0.43283582089552236,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0024024601191620218,
+      "errors": 25,
+      "errors_per_1k_tokens": 0.33550291887539424,
+      "errors_per_page_mean": 0.373134328358209,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.29850746268656714,
+      "pages_with_error": 20
+    },
+    {
+      "category": "clef",
+      "error_share": 0.0023063617143955412,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.19402985074626866,
+      "pages_with_error": 13
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.0020180665000960984,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.208955223880597,
+      "pages_with_error": 14
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.001537574476263694,
+      "errors": 16,
+      "errors_per_1k_tokens": 0.2147218680802523,
+      "errors_per_page_mean": 0.23880597014925373,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.13432835820895522,
+      "pages_with_error": 9
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0003843936190659235,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.04477611940298507,
+      "pages_with_error": 3
+    },
+    {
+      "category": "chord_structure_and_pitch",
+      "error_share": 0.00028829521429944266,
+      "errors": 3,
+      "errors_per_1k_tokens": 0.040260350265047304,
+      "errors_per_page_mean": 0.04477611940298507,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.029850746268656716,
+      "pages_with_error": 2
+    },
+    {
+      "category": "chord_structure_and_pitch_and_accidental",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "accidental_and_tie_slur_and_articulation",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "duration_and_accidental_and_tie_slur",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 1,
+    "accidental_and_tie_slur": 68,
+    "accidental_and_tie_slur_and_articulation": 1,
+    "accidental_only": 362,
+    "articulation_only": 30,
+    "barline": 170,
+    "chord_structure_and_accidental": 25,
+    "chord_structure_and_accidental_and_articulation": 4,
+    "chord_structure_and_accidental_and_tie_slur": 45,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 21,
+    "chord_structure_and_pitch": 3,
+    "chord_structure_and_pitch_and_accidental": 1,
+    "chord_structure_and_tie_slur": 202,
+    "chord_structure_and_tie_slur_and_articulation": 16,
+    "chord_structure_only": 142,
+    "clef": 24,
+    "column_or_line_break": 558,
+    "duration_and_accidental": 49,
+    "duration_and_accidental_and_tie_slur": 1,
+    "duration_and_pitch": 345,
+    "duration_and_pitch_and_accidental": 186,
+    "duration_and_pitch_and_accidental_and_tie_slur": 39,
+    "duration_and_pitch_and_tie_slur": 75,
+    "duration_and_tie_slur": 104,
+    "duration_only": 635,
+    "interpretation_or_control": 647,
+    "note_other": 2042,
+    "null_token": 116,
+    "pitch_and_accidental": 839,
+    "pitch_and_accidental_and_tie_slur": 103,
+    "pitch_and_tie_slur": 84,
+    "pitch_only": 1111,
+    "rest_note_confusion": 340,
+    "spine_operation": 244,
+    "termination": 58,
+    "tie_slur_and_articulation": 29,
+    "tie_slur_only": 1685
+  },
+  "component_error_groups": {
+    "structural": 1623,
+    "visual": 8783
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 16.40937467990443,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 4634,
+    "articulation": 25,
+    "barline": 153,
+    "clef": 26,
+    "column_or_line_break": 558,
+    "duration": 1654,
+    "interpretation_or_control": 142,
+    "note_other": 179,
+    "null_token": 116,
+    "pitch": 40,
+    "rest": 389,
+    "spine_operation": 335,
+    "termination": 55,
+    "tie_slur": 2100
+  },
+  "gold_error_groups": {
+    "structural": 1206,
+    "visual": 9200
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 14.556106567382812,
+      "LER": 53.25029754638672,
+      "SER": 15.089727401733398,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.1396497349526941,
+      "top1_errors": 10406,
+      "top1_errors_per_page": 155.3134328358209
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 15,
+    "predicted_unk_literal": 61
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/per_page_records.jsonl

@@ -0,0 +1,67 @@
+{"dataset_index": 0, "sample_id": "polish_00002", "source_ids": [], "target_token_count": 1725, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 59, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 55, "field_index": 0, "field_text": "2.AA[", "gold_field": "2.AA[", "gold_line": "2.AA[\t4r\t12r", "gold_note_token": "2.AA[", "gold_piece": "[", "gold_token_id": 43, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 67, "line_char_start": 55, "line_index": 4, "line_text": "2.AA[\t4r\t12r", "operation": "replace", "pred_field": "2.AA(", "pred_local_replacement_line": "2.AA(\t4r\t12r", "pred_note_token": "2.AA(", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 3.7895511405445177, "LER": 23.6842098236084, "SER": 3.7703016241299308}, "nll": 0.24922369420528412, "top1_error_count": 65, "top1_error_rate": 0.03768115942028986}}
+{"dataset_index": 1, "sample_id": "polish_00004", "source_ids": [], "target_token_count": 989, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 35, "field_char_start": 30, "field_index": 0, "field_text": "8AA(L", "gold_field": "8AA(L", "gold_line": "8AA(L\t4.ee(", "gold_note_token": null, "gold_piece": "8AA", "gold_token_id": 210, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 41, "line_char_start": 30, "line_index": 2, "line_text": "8AA(L\t4.ee(", "operation": "replace", "pred_field": "*M4/4(L", "pred_local_replacement_line": "*M4/4(L\t4.ee(", "pred_note_token": null, "pred_piece": "*M4/4", "pred_token_id": 229, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 16.666666666666664, "LER": 59.74842834472656, "SER": 16.09311740890688}, "nll": 0.9574042558670044, "top1_error_count": 159, "top1_error_rate": 0.16076845298281092}}
+{"dataset_index": 2, "sample_id": "polish_00005", "source_ids": [], "target_token_count": 742, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 35, "char_start": 30, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 35, "field_char_start": 30, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 41, "line_char_start": 30, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*M3", "pred_local_replacement_line": "*M3\t*M3/4", "pred_note_token": null, "pred_piece": "*M3", "pred_token_id": 259, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 7.702984038861901, "LER": 45.66929244995117, "SER": 9.7165991902834}, "nll": 0.5568324327468872, "top1_error_count": 72, "top1_error_rate": 0.09703504043126684}}
+{"dataset_index": 3, "sample_id": "polish_00007", "source_ids": [], "target_token_count": 888, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 31, "char_start": 26, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 31, "field_char_start": 26, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 37, "line_char_start": 26, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t*M3/4", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 18.946692357096982, "LER": 64.0, "SER": 17.587373167981962}, "nll": 0.9496373534202576, "top1_error_count": 156, "top1_error_rate": 0.17567567567567569}}
+{"dataset_index": 4, "sample_id": "polish_00008", "source_ids": [], "target_token_count": 1236, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 48, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 48, "field_char_start": 45, "field_index": 1, "field_text": "8d[", "gold_field": "8d[", "gold_line": "8r\t8d[", "gold_note_token": "8d[", "gold_piece": "8d[", "gold_token_id": 1247, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 48, "line_char_start": 42, "line_index": 3, "line_text": "8r\t8d[", "operation": "replace", "pred_field": "8d(", "pred_local_replacement_line": "8r\t8d(", "pred_note_token": "8d(", "pred_piece": "8d(", "pred_token_id": 1822, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 11.451844753234308, "LER": 50.26177978515625, "SER": 11.417004048582996}, "nll": 0.6420015692710876, "top1_error_count": 141, "top1_error_rate": 0.11407766990291263}}
+{"dataset_index": 5, "sample_id": "polish_00009", "source_ids": [], "target_token_count": 942, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 110, "char_start": 107, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 110, "field_char_start": 107, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "4BB", "gold_token_id": 274, "gold_token_index": 51, "gold_trimmed_index": 51, "group": "structural", "line_char_end": 120, "line_char_start": 107, "line_index": 11, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "*^", "pred_local_replacement_line": "*^\t4d 4f# 4b", "pred_note_token": null, "pred_piece": "*^", "pred_token_id": 191, "pred_trimmed_index": 51, "visually_grounded": false}, "metrics": {"CER": 10.236724248240563, "LER": 29.629629135131836, "SER": 9.032943676939425}, "nll": 0.5724238753318787, "top1_error_count": 85, "top1_error_rate": 0.09023354564755838}}
+{"dataset_index": 6, "sample_id": "polish_00011", "source_ids": [], "target_token_count": 1093, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#d#]", "pred_local_replacement_line": "*k[f#c#g#d#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#d#]", "pred_token_id": 688, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 10.475708502024291, "LER": 49.47368240356445, "SER": 12.179487179487179}, "nll": 0.6495988965034485, "top1_error_count": 133, "top1_error_rate": 0.12168344007319305}}
+{"dataset_index": 7, "sample_id": "polish_00013", "source_ids": [], "target_token_count": 1877, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 47, "char_start": 43, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 49, "field_char_start": 43, "field_index": 1, "field_text": "16.g(L", "gold_field": "16.g(L", "gold_line": "8CC 8GG 8CL\t16.g(L\t4en", "gold_note_token": "16.g(L", "gold_piece": "16.g", "gold_token_id": 1474, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 53, "line_char_start": 31, "line_index": 3, "line_text": "8CC 8GG 8CL\t16.g(L\t4en", "operation": "replace", "pred_field": "8.g(L", "pred_local_replacement_line": "8CC 8GG 8CL\t8.g(L\t4en", "pred_note_token": "8.g(L", "pred_piece": "8.g", "pred_token_id": 292, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 5.938634114153745, "LER": 40.52863311767578, "SER": 6.183368869936034}, "nll": 0.3528750240802765, "top1_error_count": 116, "top1_error_rate": 0.06180074587107086}}
+{"dataset_index": 8, "sample_id": "polish_00014", "source_ids": [], "target_token_count": 1099, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t16bn(L", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 57, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t16bn(L", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t16bn(L", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 12.031952071892162, "LER": 45.45454406738281, "SER": 13.387978142076504}, "nll": 0.768509566783905, "top1_error_count": 147, "top1_error_rate": 0.1337579617834395}}
+{"dataset_index": 9, "sample_id": "polish_00015", "source_ids": [], "target_token_count": 1086, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 38, "char_start": 37, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 38, "field_char_start": 37, "field_index": 1, "field_text": "*", "gold_field": "*", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "*^", "pred_local_replacement_line": "*^\t*^", "pred_note_token": null, "pred_piece": "*^", "pred_token_id": 191, "pred_trimmed_index": 10, "visually_grounded": false}, "metrics": {"CER": 13.455328310010763, "LER": 54.601226806640625, "SER": 13.640552995391705}, "nll": 0.7012812495231628, "top1_error_count": 148, "top1_error_rate": 0.13627992633517497}}
+{"dataset_index": 10, "sample_id": "polish_00017", "source_ids": [], "target_token_count": 1053, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 44, "field_char_start": 34, "field_index": 0, "field_text": "4BBB- 4BB-", "gold_field": "4BBB- 4BB-", "gold_line": "4BBB- 4BB-\t4E 4cn", "gold_note_token": null, "gold_piece": "4BBB-", "gold_token_id": 834, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 51, "line_char_start": 34, "line_index": 2, "line_text": "4BBB- 4BB-\t4E 4cn", "operation": "replace", "pred_field": "* 4BB-", "pred_local_replacement_line": "* 4BB-\t4E 4cn", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 6.015831134564644, "LER": 27.22222137451172, "SER": 8.365019011406844}, "nll": 0.47717782855033875, "top1_error_count": 88, "top1_error_rate": 0.0835707502374169}}
+{"dataset_index": 11, "sample_id": "polish_00019", "source_ids": [], "target_token_count": 1006, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 29, "char_start": 26, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 33, "field_char_start": 26, "field_index": 0, "field_text": "8AA 8AL", "gold_field": "8AA 8AL", "gold_line": "8AA 8AL\t8.cc 8.cccL", "gold_note_token": null, "gold_piece": "8AA", "gold_token_id": 210, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 45, "line_char_start": 26, "line_index": 2, "line_text": "8AA 8AL\t8.cc 8.cccL", "operation": "replace", "pred_field": "*M3/4 8AL", "pred_local_replacement_line": "*M3/4 8AL\t8.cc 8.cccL", "pred_note_token": null, "pred_piece": "*M3/4", "pred_token_id": 283, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 13.457556935817806, "LER": 50.588233947753906, "SER": 13.034825870646765}, "nll": 0.7551409006118774, "top1_error_count": 131, "top1_error_rate": 0.13021868787276342}}
+{"dataset_index": 12, "sample_id": "polish_00020", "source_ids": [], "target_token_count": 1183, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 8.504007123775601, "LER": 46.41148376464844, "SER": 11.16751269035533}, "nll": 0.5810454487800598, "top1_error_count": 132, "top1_error_rate": 0.11158072696534235}}
+{"dataset_index": 13, "sample_id": "polish_00022", "source_ids": [], "target_token_count": 695, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 35, "char_start": 30, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 35, "field_char_start": 30, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 41, "line_char_start": 30, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*M2/4", "pred_local_replacement_line": "*M2/4\t*M3/4", "pred_note_token": null, "pred_piece": "*M2/4", "pred_token_id": 304, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 22.13622291021672, "LER": 63.90977478027344, "SER": 20.7492795389049}, "nll": 1.1237390041351318, "top1_error_count": 144, "top1_error_rate": 0.20719424460431654}}
+{"dataset_index": 14, "sample_id": "polish_00023", "source_ids": [], "target_token_count": 1332, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 51, "field_index": 1, "field_text": "4gg)", "gold_field": "4gg)", "gold_line": "8GG 8GL\t4gg)\t8bnL", "gold_note_token": "4gg)", "gold_piece": "4gg)", "gold_token_id": 2154, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 60, "line_char_start": 43, "line_index": 3, "line_text": "8GG 8GL\t4gg)\t8bnL", "operation": "replace", "pred_field": "4gg]", "pred_local_replacement_line": "8GG 8GL\t4gg]\t8bnL", "pred_note_token": "4gg]", "pred_piece": "4gg]", "pred_token_id": 1531, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 10.304990757855823, "LER": 52.43902587890625, "SER": 10.293012772351616}, "nll": 0.5901538133621216, "top1_error_count": 137, "top1_error_rate": 0.10285285285285285}}
+{"dataset_index": 15, "sample_id": "polish_00024", "source_ids": [], "target_token_count": 1075, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#d#]", "pred_local_replacement_line": "*k[f#c#g#d#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#d#]", "pred_token_id": 688, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 11.829740187949144, "LER": 46.242774963378906, "SER": 12.383612662942271}, "nll": 0.6182152032852173, "top1_error_count": 133, "top1_error_rate": 0.12372093023255815}}
+{"dataset_index": 16, "sample_id": "polish_00025", "source_ids": [], "target_token_count": 879, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 24, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 24, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#]", "gold_field": "*k[f#c#]", "gold_line": "*k[f#c#]\t*k[f#c#]", "gold_note_token": null, "gold_piece": "*k[f#c#]", "gold_token_id": 294, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 33, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#]\t*k[f#c#]", "operation": "replace", "pred_field": "*k[]", "pred_local_replacement_line": "*k[]\t*k[f#c#]", "pred_note_token": null, "pred_piece": "*k[]", "pred_token_id": 302, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 15.214723926380369, "LER": 59.44055938720703, "SER": 15.717539863325742}, "nll": 0.809155285358429, "top1_error_count": 138, "top1_error_rate": 0.15699658703071673}}
+{"dataset_index": 17, "sample_id": "polish_00027", "source_ids": [], "target_token_count": 1825, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 51, "char_start": 46, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 56, "field_char_start": 46, "field_index": 0, "field_text": "8BBB- 8BB-", "gold_field": "8BBB- 8BB-", "gold_line": "8BBB- 8BB-\t8b- 8dd", "gold_note_token": null, "gold_piece": "8BBB-", "gold_token_id": 632, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 64, "line_char_start": 46, "line_index": 3, "line_text": "8BBB- 8BB-\t8b- 8dd", "operation": "replace", "pred_field": "*^ 8BB-", "pred_local_replacement_line": "*^ 8BB-\t8b- 8dd", "pred_note_token": null, "pred_piece": "*^", "pred_token_id": 191, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 16.05010195164579, "LER": 61.627906799316406, "SER": 15.021929824561402}, "nll": 0.7721052169799805, "top1_error_count": 274, "top1_error_rate": 0.15013698630136987}}
+{"dataset_index": 18, "sample_id": "polish_00029", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 30, "field_index": 0, "field_text": "8CC( 8C(L", "gold_field": "8CC( 8C(L", "gold_line": "8CC( 8C(L\t4gg([", "gold_note_token": null, "gold_piece": "8CC", "gold_token_id": 364, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 45, "line_char_start": 30, "line_index": 2, "line_text": "8CC( 8C(L\t4gg([", "operation": "replace", "pred_field": "*met(C( 8C(L", "pred_local_replacement_line": "*met(C( 8C(L\t4gg([", "pred_note_token": null, "pred_piece": "*met(C", "pred_token_id": 1348, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 10.702054794520548, "LER": 44.680850982666016, "SER": 12.228479485116653}, "nll": 0.662446141242981, "top1_error_count": 152, "top1_error_rate": 0.12218649517684887}}
+{"dataset_index": 19, "sample_id": "polish_00030", "source_ids": [], "target_token_count": 1263, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "*M4/4", "pred_local_replacement_line": "*M4/4\t*^", "pred_note_token": null, "pred_piece": "*M4/4", "pred_token_id": 229, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 13.96208969024503, "LER": 44.1860466003418, "SER": 12.044374009508717}, "nll": 0.6165903210639954, "top1_error_count": 152, "top1_error_rate": 0.12034837688044339}}
+{"dataset_index": 20, "sample_id": "polish_00031", "source_ids": [], "target_token_count": 862, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "null_token", "char_end": 47, "char_start": 46, "component_category": "null_token", "component_flags": [], "component_group": "structural", "field_char_end": 47, "field_char_start": 46, "field_index": 0, "field_text": ".", "gold_field": ".", "gold_line": ".\t16g#q(", "gold_note_token": null, "gold_piece": ".", "gold_token_id": 14, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 54, "line_char_start": 46, "line_index": 3, "line_text": ".\t16g#q(", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t16g#q(", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 19.960861056751465, "LER": 62.237762451171875, "SER": 17.88617886178862}, "nll": 1.12758469581604, "top1_error_count": 154, "top1_error_rate": 0.17865429234338748}}
+{"dataset_index": 21, "sample_id": "polish_00032", "source_ids": [], "target_token_count": 1387, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 50, "char_start": 45, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 52, "field_char_start": 45, "field_index": 2, "field_text": "16ff#]L", "gold_field": "16ff#]L", "gold_line": "4G#(\t8G#L\t16ff#]L", "gold_note_token": "16ff#]L", "gold_piece": "16ff#", "gold_token_id": 180, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 52, "line_char_start": 35, "line_index": 3, "line_text": "4G#(\t8G#L\t16ff#]L", "operation": "replace", "pred_field": "24ff#]L", "pred_local_replacement_line": "4G#(\t8G#L\t24ff#]L", "pred_note_token": "24ff#]L", "pred_piece": "24ff#", "pred_token_id": 994, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 11.189979123173277, "LER": 47.05882263183594, "SER": 12.481962481962482}, "nll": 0.7121825814247131, "top1_error_count": 173, "top1_error_rate": 0.1247296322999279}}
+{"dataset_index": 22, "sample_id": "polish_00034", "source_ids": [], "target_token_count": 788, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 35, "char_start": 30, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 35, "field_char_start": 30, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 41, "line_char_start": 30, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*M3", "pred_local_replacement_line": "*M3\t*M3/4", "pred_note_token": null, "pred_piece": "*M3", "pred_token_id": 259, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 34.72750316856781, "LER": 82.60869598388672, "SER": 48.53875476493011}, "nll": 1.82101309299469, "top1_error_count": 240, "top1_error_rate": 0.30456852791878175}}
+{"dataset_index": 23, "sample_id": "polish_00036", "source_ids": [], "target_token_count": 1317, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 55, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 65, "field_char_start": 55, "field_index": 0, "field_text": "2.CC[ 2.C[", "gold_field": "2.CC[ 2.C[", "gold_line": "2.CC[ 2.C[\t4r\t8r", "gold_note_token": null, "gold_piece": "2.CC", "gold_token_id": 1965, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 71, "line_char_start": 55, "line_index": 4, "line_text": "2.CC[ 2.C[\t4r\t8r", "operation": "replace", "pred_field": "*[ 2.C[", "pred_local_replacement_line": "*[ 2.C[\t4r\t8r", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 15.708465368945223, "LER": 49.47916793823242, "SER": 12.386018237082066}, "nll": 0.795889139175415, "top1_error_count": 163, "top1_error_rate": 0.12376613515565679}}
+{"dataset_index": 24, "sample_id": "polish_00037", "source_ids": [], "target_token_count": 871, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 49, "char_start": 46, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 53, "field_char_start": 46, "field_index": 0, "field_text": "4Fn 4dn", "gold_field": "4Fn 4dn", "gold_line": "4Fn 4dn\t2f#] 2ff#]", "gold_note_token": null, "gold_piece": "4Fn", "gold_token_id": 1918, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 2, "line_text": "4Fn 4dn\t2f#] 2ff#]", "operation": "replace", "pred_field": "* 4dn", "pred_local_replacement_line": "* 4dn\t2f#] 2ff#]", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 8.502289077828646, "LER": 36.24161148071289, "SER": 9.770114942528735}, "nll": 0.5092135667800903, "top1_error_count": 85, "top1_error_rate": 0.09758897818599312}}
+{"dataset_index": 25, "sample_id": "polish_00038", "source_ids": [], "target_token_count": 1471, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 88, "char_start": 83, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 88, "field_char_start": 83, "field_index": 1, "field_text": "16a(L", "gold_field": "16a(L", "gold_line": "8CC 8FF# 8CJ\t16a(L\t8e-", "gold_note_token": "16a(L", "gold_piece": "16a(L", "gold_token_id": 2253, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 92, "line_char_start": 70, "line_index": 5, "line_text": "8CC 8FF# 8CJ\t16a(L\t8e-", "operation": "replace", "pred_field": "16a(LL", "pred_local_replacement_line": "8CC 8FF# 8CJ\t16a(LL\t8e-", "pred_note_token": "16a(LL", "pred_piece": "16a(LL", "pred_token_id": 1061, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 10.871302957633892, "LER": 55.86592102050781, "SER": 9.727891156462585}, "nll": 0.5586496591567993, "top1_error_count": 143, "top1_error_rate": 0.09721278042148199}}
+{"dataset_index": 26, "sample_id": "polish_00039", "source_ids": [], "target_token_count": 1503, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 93, "char_start": 88, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 93, "field_char_start": 88, "field_index": 0, "field_text": "16A-L", "gold_field": "16A-L", "gold_line": "16A-L\t2.AA-[\t16cL\t8ee-", "gold_note_token": null, "gold_piece": "16A-L", "gold_token_id": 2834, "gold_token_index": 24, "gold_trimmed_index": 24, "group": "visual", "line_char_end": 110, "line_char_start": 88, "line_index": 4, "line_text": "16A-L\t2.AA-[\t16cL\t8ee-", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t2.AA-[\t16cL\t8ee-", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 24, "visually_grounded": true}, "metrics": {"CER": 12.783424550430023, "LER": 50.264549255371094, "SER": 11.651131824234353}, "nll": 0.6559882760047913, "top1_error_count": 175, "top1_error_rate": 0.1164337990685296}}
+{"dataset_index": 27, "sample_id": "polish_00040", "source_ids": [], "target_token_count": 1109, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#d#]", "pred_local_replacement_line": "*k[f#c#g#d#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#d#]", "pred_token_id": 688, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 12.104743083003951, "LER": 51.1363639831543, "SER": 13.537906137184116}, "nll": 0.7346525192260742, "top1_error_count": 150, "top1_error_rate": 0.13525698827772767}}
+{"dataset_index": 28, "sample_id": "polish_00043", "source_ids": [], "target_token_count": 885, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "null_token", "char_end": 55, "char_start": 54, "component_category": "null_token", "component_flags": [], "component_group": "structural", "field_char_end": 55, "field_char_start": 54, "field_index": 0, "field_text": ".", "gold_field": ".", "gold_line": ".\t8ccq(", "gold_note_token": null, "gold_piece": ".", "gold_token_id": 14, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "insert", "pred_field": "*", "pred_local_replacement_line": "*\t8ccq(", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 25.657510227936882, "LER": 69.3641586303711, "SER": 25.1131221719457}, "nll": 1.4354969263076782, "top1_error_count": 222, "top1_error_rate": 0.25084745762711863}}
+{"dataset_index": 29, "sample_id": "polish_00044", "source_ids": [], "target_token_count": 677, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 46, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 51, "field_char_start": 46, "field_index": 0, "field_text": "8.B-L", "gold_field": "8.B-L", "gold_line": "8.B-L\t8.f 8.ddL", "gold_note_token": null, "gold_piece": "8.B-L", "gold_token_id": 1161, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 61, "line_char_start": 46, "line_index": 3, "line_text": "8.B-L\t8.f 8.ddL", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t8.f 8.ddL", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 11.869666408068271, "LER": 54.54545593261719, "SER": 14.644970414201183}, "nll": 0.7925536632537842, "top1_error_count": 99, "top1_error_rate": 0.14623338257016247}}
+{"dataset_index": 30, "sample_id": "polish_00046", "source_ids": [], "target_token_count": 1069, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 28, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 28, "field_char_start": 16, "field_index": 0, "field_text": "*k[b-e-a-d-]", "gold_field": "*k[b-e-a-d-]", "gold_line": "*k[b-e-a-d-]\t*k[b-e-a-d-]", "gold_note_token": null, "gold_piece": "*k[b-e-a-d-]", "gold_token_id": 637, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 41, "line_char_start": 16, "line_index": 1, "line_text": "*k[b-e-a-d-]\t*k[b-e-a-d-]", "operation": "replace", "pred_field": "*k[b-e-a-]", "pred_local_replacement_line": "*k[b-e-a-]\t*k[b-e-a-d-]", "pred_note_token": null, "pred_piece": "*k[b-e-a-]", "pred_token_id": 521, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 17.571059431524546, "LER": 65.3594741821289, "SER": 17.790262172284642}, "nll": 0.9108795523643494, "top1_error_count": 190, "top1_error_rate": 0.17773620205799812}}
+{"dataset_index": 31, "sample_id": "polish_00047", "source_ids": [], "target_token_count": 1243, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 42, "char_start": 39, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 42, "field_char_start": 39, "field_index": 0, "field_text": "8CL", "gold_field": "8CL", "gold_line": "8CL\t4r\t4ee", "gold_note_token": null, "gold_piece": "8CL", "gold_token_id": 442, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 49, "line_char_start": 39, "line_index": 2, "line_text": "8CL\t4r\t4ee", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t4r\t4ee", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 15.307101727447217, "LER": 59.75609588623047, "SER": 13.285024154589372}, "nll": 0.8548486232757568, "top1_error_count": 165, "top1_error_rate": 0.13274336283185842}}
+{"dataset_index": 32, "sample_id": "polish_00048", "source_ids": [], "target_token_count": 1384, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 44, "field_char_start": 42, "field_index": 0, "field_text": "4F", "gold_field": "4F", "gold_line": "4F\t4r", "gold_note_token": null, "gold_piece": "4F", "gold_token_id": 153, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 47, "line_char_start": 42, "line_index": 2, "line_text": "4F\t4r", "operation": "replace", "pred_field": "*M3/4", "pred_local_replacement_line": "*M3/4\t4r", "pred_note_token": null, "pred_piece": "*M3/4", "pred_token_id": 283, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 10.70578905630452, "LER": 31.460674285888672, "SER": 7.953723788864786}, "nll": 0.4545893967151642, "top1_error_count": 110, "top1_error_rate": 0.07947976878612717}}
+{"dataset_index": 33, "sample_id": "polish_00049", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 43, "char_start": 39, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 39, "field_index": 0, "field_text": "8AAL", "gold_field": "8AAL", "gold_line": "8AAL\t8.r\t4ee", "gold_note_token": null, "gold_piece": "8AAL", "gold_token_id": 512, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 51, "line_char_start": 39, "line_index": 2, "line_text": "8AAL\t8.r\t4ee", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t8.r\t4ee", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 19.26926926926927, "LER": 61.44578170776367, "SER": 16.170555108608205}, "nll": 0.9706302881240845, "top1_error_count": 201, "top1_error_rate": 0.16157556270096463}}
+{"dataset_index": 34, "sample_id": "polish_00050", "source_ids": [], "target_token_count": 1242, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 47, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 47, "field_char_start": 42, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 53, "line_char_start": 42, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*M3/2", "pred_local_replacement_line": "*M3/2\t*M3/4", "pred_note_token": null, "pred_piece": "*M3/2", "pred_token_id": 1252, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 20.99396191360892, "LER": 68.4491958618164, "SER": 18.77518130539887}, "nll": 0.9367584586143494, "top1_error_count": 233, "top1_error_rate": 0.18760064412238325}}
+{"dataset_index": 35, "sample_id": "polish_00052", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 47, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 59, "field_char_start": 47, "field_index": 0, "field_text": "2.DD-( 2.D-(", "gold_field": "2.DD-( 2.D-(", "gold_line": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "gold_note_token": null, "gold_piece": "2.DD", "gold_token_id": 1567, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 74, "line_char_start": 47, "line_index": 3, "line_text": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "operation": "replace", "pred_field": "*-( 2.D-(", "pred_local_replacement_line": "*-( 2.D-(\t8ff(L\t2.f 2.a-", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 19.263456090651555, "LER": 62.295082092285156, "SER": 20.16460905349794}, "nll": 1.1125969886779785, "top1_error_count": 196, "top1_error_rate": 0.2014388489208633}}
+{"dataset_index": 36, "sample_id": "polish_00053", "source_ids": [], "target_token_count": 1018, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "rest", "char_end": 36, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 36, "field_char_start": 34, "field_index": 0, "field_text": "4r", "gold_field": "4r", "gold_line": "4r\t8cn(L", "gold_note_token": null, "gold_piece": "4r", "gold_token_id": 97, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 42, "line_char_start": 34, "line_index": 2, "line_text": "4r\t8cn(L", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t8cn(L", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 6.6547831253713605, "LER": 20.44198989868164, "SER": 5.801376597836775}, "nll": 0.4127529561519623, "top1_error_count": 59, "top1_error_rate": 0.05795677799607073}}
+{"dataset_index": 37, "sample_id": "polish_00054", "source_ids": [], "target_token_count": 689, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 46, "field_index": 0, "field_text": "8B-( 8d(L", "gold_field": "8B-( 8d(L", "gold_line": "8B-( 8d(L\t8fL", "gold_note_token": "8d(L", "gold_piece": "8d(L", "gold_token_id": 1103, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 59, "line_char_start": 46, "line_index": 3, "line_text": "8B-( 8d(L\t8fL", "operation": "replace", "pred_field": "8B-( 8dL", "pred_local_replacement_line": "8B-( 8dL\t8fL", "pred_note_token": "8dL", "pred_piece": "8dL", "pred_token_id": 206, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 16.334355828220858, "LER": 59.55882263183594, "SER": 18.168604651162788}, "nll": 1.0013749599456787, "top1_error_count": 125, "top1_error_rate": 0.18142235123367198}}
+{"dataset_index": 38, "sample_id": "polish_00056", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 45, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 45, "field_char_start": 42, "field_index": 0, "field_text": "8FL", "gold_field": "8FL", "gold_line": "8FL\t4cc 4ccc", "gold_note_token": null, "gold_piece": "8FL", "gold_token_id": 433, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 54, "line_char_start": 42, "line_index": 2, "line_text": "8FL\t4cc 4ccc", "operation": "replace", "pred_field": "*M3/4", "pred_local_replacement_line": "*M3/4\t4cc 4ccc", "pred_note_token": null, "pred_piece": "*M3/4", "pred_token_id": 283, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 28.7248322147651, "LER": 69.12442016601562, "SER": 26.690391459074732}, "nll": 1.4108033180236816, "top1_error_count": 300, "top1_error_rate": 0.26666666666666666}}
+{"dataset_index": 39, "sample_id": "polish_00057", "source_ids": [], "target_token_count": 733, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 61, "char_start": 58, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 67, "field_char_start": 58, "field_index": 0, "field_text": "8e-( 8g(L", "gold_field": "8e-( 8g(L", "gold_line": "8e-( 8g(L\t4gg", "gold_note_token": null, "gold_piece": "8e-", "gold_token_id": 263, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 71, "line_char_start": 58, "line_index": 4, "line_text": "8e-( 8g(L\t4gg", "operation": "replace", "pred_field": "*( 8g(L", "pred_local_replacement_line": "*( 8g(L\t4gg", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 23.58108108108108, "LER": 64.33566284179688, "SER": 24.726775956284154}, "nll": 1.361873984336853, "top1_error_count": 181, "top1_error_rate": 0.24693042291950887}}
+{"dataset_index": 40, "sample_id": "polish_00058", "source_ids": [], "target_token_count": 900, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 46, "char_start": 42, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 46, "field_char_start": 42, "field_index": 0, "field_text": "8.DL", "gold_field": "8.DL", "gold_line": "8.DL\t8.dL", "gold_note_token": "8.DL", "gold_piece": "8.DL", "gold_token_id": 1624, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 51, "line_char_start": 42, "line_index": 3, "line_text": "8.DL\t8.dL", "operation": "replace", "pred_field": "8DL", "pred_local_replacement_line": "8DL\t8.dL", "pred_note_token": "8DL", "pred_piece": "8DL", "pred_token_id": 407, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 17.507598784194528, "LER": 85.71428680419922, "SER": 66.74082313681868}, "nll": 0.9576418995857239, "top1_error_count": 173, "top1_error_rate": 0.1922222222222222}}
+{"dataset_index": 41, "sample_id": "polish_00059", "source_ids": [], "target_token_count": 1838, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 34, "field_index": 0, "field_text": "*M3/8", "gold_field": "*M3/8", "gold_line": "*M3/8\t*M3/8", "gold_note_token": null, "gold_piece": "*M3/8", "gold_token_id": 592, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 45, "line_char_start": 34, "line_index": 2, "line_text": "*M3/8\t*M3/8", "operation": "replace", "pred_field": "*M2/4", "pred_local_replacement_line": "*M2/4\t*M3/8", "pred_note_token": null, "pred_piece": "*M2/4", "pred_token_id": 304, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 13.656783468104223, "LER": 63.70656204223633, "SER": 13.445835601524225}, "nll": 0.7079006433486938, "top1_error_count": 247, "top1_error_rate": 0.13438520130576714}}
+{"dataset_index": 42, "sample_id": "polish_00060", "source_ids": [], "target_token_count": 1528, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 62, "char_start": 59, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 66, "field_char_start": 59, "field_index": 0, "field_text": "8FF 8FL", "gold_field": "8FF 8FL", "gold_line": "8FF 8FL\t2.F\t4.f", "gold_note_token": null, "gold_piece": "8FF", "gold_token_id": 276, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 74, "line_char_start": 59, "line_index": 4, "line_text": "8FF 8FL\t2.F\t4.f", "operation": "replace", "pred_field": "* 8FL", "pred_local_replacement_line": "* 8FL\t2.F\t4.f", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 8.225108225108226, "LER": 51.86915969848633, "SER": 10.085134250163721}, "nll": 0.5479966998100281, "top1_error_count": 154, "top1_error_rate": 0.10078534031413612}}
+{"dataset_index": 43, "sample_id": "polish_00061", "source_ids": [], "target_token_count": 653, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 26, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 26, "field_char_start": 16, "field_index": 0, "field_text": "*k[b-e-a-]", "gold_field": "*k[b-e-a-]", "gold_line": "*k[b-e-a-]\t*k[b-e-a-]", "gold_note_token": null, "gold_piece": "*k[b-e-a-]", "gold_token_id": 521, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 37, "line_char_start": 16, "line_index": 1, "line_text": "*k[b-e-a-]\t*k[b-e-a-]", "operation": "replace", "pred_field": "*k[b-e-a-d-]", "pred_local_replacement_line": "*k[b-e-a-d-]\t*k[b-e-a-]", "pred_note_token": null, "pred_piece": "*k[b-e-a-d-]", "pred_token_id": 637, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 29.836309523809522, "LER": 75.51020050048828, "SER": 31.28834355828221}, "nll": 1.7003357410430908, "top1_error_count": 204, "top1_error_rate": 0.3124042879019908}}
+{"dataset_index": 44, "sample_id": "polish_00062", "source_ids": [], "target_token_count": 781, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 16.911764705882355, "LER": 64.4859848022461, "SER": 14.743589743589745}, "nll": 0.9174989461898804, "top1_error_count": 115, "top1_error_rate": 0.147247119078105}}
+{"dataset_index": 45, "sample_id": "polish_00063", "source_ids": [], "target_token_count": 963, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "null_token", "char_end": 47, "char_start": 46, "component_category": "null_token", "component_flags": [], "component_group": "structural", "field_char_end": 47, "field_char_start": 46, "field_index": 0, "field_text": ".", "gold_field": ".", "gold_line": ".\t16qqddnL", "gold_note_token": null, "gold_piece": ".", "gold_token_id": 14, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 56, "line_char_start": 46, "line_index": 2, "line_text": ".\t16qqddnL", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t16qqddnL", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 6.261510128913445, "LER": 32.23684310913086, "SER": 8.212058212058212}, "nll": 0.4657658338546753, "top1_error_count": 79, "top1_error_rate": 0.08203530633437175}}
+{"dataset_index": 46, "sample_id": "polish_00064", "source_ids": [], "target_token_count": 928, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 53, "char_start": 49, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 53, "field_char_start": 49, "field_index": 1, "field_text": "8dd[", "gold_field": "8dd[", "gold_line": "8r\t8dd[", "gold_note_token": "8dd[", "gold_piece": "8dd[", "gold_token_id": 1303, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 53, "line_char_start": 46, "line_index": 3, "line_text": "8r\t8dd[", "operation": "replace", "pred_field": "8dd(", "pred_local_replacement_line": "8r\t8dd(", "pred_note_token": "8dd(", "pred_piece": "8dd(", "pred_token_id": 1124, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 10.69182389937107, "LER": 45.76271057128906, "SER": 13.807982740021574}, "nll": 0.7600541114807129, "top1_error_count": 128, "top1_error_rate": 0.13793103448275862}}
+{"dataset_index": 47, "sample_id": "polish_00067", "source_ids": [], "target_token_count": 1138, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 43, "char_start": 39, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 43, "field_char_start": 39, "field_index": 1, "field_text": "16aL", "gold_field": "16aL", "gold_line": "8F(L\t16aL", "gold_note_token": "16aL", "gold_piece": "16aL", "gold_token_id": 602, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 43, "line_char_start": 34, "line_index": 2, "line_text": "8F(L\t16aL", "operation": "replace", "pred_field": "16a(LL", "pred_local_replacement_line": "8F(L\t16a(LL", "pred_note_token": "16a(LL", "pred_piece": "16a(LL", "pred_token_id": 1061, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 12.399811409712399, "LER": 47.91666793823242, "SER": 15.127528583992964}, "nll": 0.820811927318573, "top1_error_count": 172, "top1_error_rate": 0.15114235500878734}}
+{"dataset_index": 48, "sample_id": "polish_00068", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 57, "char_start": 53, "component_category": "rest_note_confusion", "component_flags": ["is_rest", "pitch", "accidental"], "component_group": "visual", "field_char_end": 57, "field_char_start": 53, "field_index": 1, "field_text": "4ff#", "gold_field": "4ff#", "gold_line": "4r\t4ff#", "gold_note_token": "4ff#", "gold_piece": "4ff#", "gold_token_id": 225, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": "4r\t4ff#", "operation": "replace", "pred_field": "4r", "pred_local_replacement_line": "4r\t4r", "pred_note_token": "4r", "pred_piece": "4r", "pred_token_id": 97, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 33.87909319899244, "LER": 75.13812255859375, "SER": 28.80658436213992}, "nll": 1.6439672708511353, "top1_error_count": 280, "top1_error_rate": 0.28776978417266186}}
+{"dataset_index": 49, "sample_id": "polish_00071", "source_ids": [], "target_token_count": 1433, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 48, "char_start": 46, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 48, "field_char_start": 46, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*^", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 51, "line_char_start": 46, "line_index": 2, "line_text": "*^\t*^", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t*^", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 13.142375737152484, "LER": 60.655738830566406, "SER": 12.988826815642456}, "nll": 0.7404247522354126, "top1_error_count": 186, "top1_error_rate": 0.1297976273551989}}
+{"dataset_index": 50, "sample_id": "polish_00072", "source_ids": [], "target_token_count": 1123, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 55, "char_start": 52, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "4BB", "gold_token_id": 274, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 65, "line_char_start": 52, "line_index": 3, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t4d 4f# 4b", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 5.950579929399899, "LER": 29.787233352661133, "SER": 6.951871657754011}, "nll": 0.41038915514945984, "top1_error_count": 78, "top1_error_rate": 0.06945681211041853}}
+{"dataset_index": 51, "sample_id": "polish_00074", "source_ids": [], "target_token_count": 1355, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 34, "char_start": 30, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 36, "field_char_start": 30, "field_index": 0, "field_text": "8BBnq(", "gold_field": "8BBnq(", "gold_line": "8BBnq(\t.", "gold_note_token": null, "gold_piece": "8BBn", "gold_token_id": 2042, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 38, "line_char_start": 30, "line_index": 2, "line_text": "8BBnq(\t.", "operation": "replace", "pred_field": "*met(Cq(", "pred_local_replacement_line": "*met(Cq(\t.", "pred_note_token": null, "pred_piece": "*met(C", "pred_token_id": 1348, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 16.333777185974256, "LER": 59.487178802490234, "SER": 15.731166912850814}, "nll": 0.8584330081939697, "top1_error_count": 213, "top1_error_rate": 0.15719557195571957}}
+{"dataset_index": 52, "sample_id": "polish_00078", "source_ids": [], "target_token_count": 1172, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 54, "char_start": 52, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 54, "field_char_start": 52, "field_index": 2, "field_text": "*^", "gold_field": "*^", "gold_line": "*\t*\t*^\t*", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "structural", "line_char_end": 56, "line_char_start": 48, "line_index": 3, "line_text": "*\t*\t*^\t*", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t*\t*\t*", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 16, "visually_grounded": false}, "metrics": {"CER": 14.606741573033707, "LER": 48.701297760009766, "SER": 11.357813834329633}, "nll": 0.5689952373504639, "top1_error_count": 133, "top1_error_rate": 0.11348122866894197}}
+{"dataset_index": 53, "sample_id": "polish_00082", "source_ids": [], "target_token_count": 1027, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "*M4/4", "pred_local_replacement_line": "*M4/4\t*^", "pred_note_token": null, "pred_piece": "*M4/4", "pred_token_id": 229, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 16.11295681063123, "LER": 53.94736862182617, "SER": 12.865497076023392}, "nll": 0.7035173773765564, "top1_error_count": 132, "top1_error_rate": 0.12852969814995133}}
+{"dataset_index": 54, "sample_id": "polish_00084", "source_ids": [], "target_token_count": 860, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 38, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 38, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 49, "line_char_start": 38, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*M3/2", "pred_local_replacement_line": "*M3/2\t*M3/4", "pred_note_token": null, "pred_piece": "*M3/2", "pred_token_id": 1252, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 19.695121951219512, "LER": 62.42038345336914, "SER": 20.13969732246799}, "nll": 1.1918102502822876, "top1_error_count": 173, "top1_error_rate": 0.20116279069767443}}
+{"dataset_index": 55, "sample_id": "polish_00085", "source_ids": [], "target_token_count": 1428, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 54, "char_start": 49, "component_category": "duration_and_tie_slur", "component_flags": ["duration", "tie_slur"], "component_group": "visual", "field_char_end": 54, "field_char_start": 49, "field_index": 2, "field_text": "8aa(L", "gold_field": "8aa(L", "gold_line": "4d\t4D 4A\t8aa(L\t4f# 4dd", "gold_note_token": "8aa(L", "gold_piece": "8aa(L", "gold_token_id": 1057, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 40, "line_index": 3, "line_text": "4d\t4D 4A\t8aa(L\t4f# 4dd", "operation": "replace", "pred_field": "8.aa", "pred_local_replacement_line": "4d\t4D 4A\t8.aa\t4f# 4dd", "pred_note_token": "8.aa", "pred_piece": "8.aa", "pred_token_id": 386, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 10.37037037037037, "LER": 46.5838508605957, "SER": 8.759635599159076}, "nll": 0.5472274422645569, "top1_error_count": 125, "top1_error_rate": 0.08753501400560224}}
+{"dataset_index": 56, "sample_id": "polish_00086", "source_ids": [], "target_token_count": 1170, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 44, "field_char_start": 42, "field_index": 0, "field_text": "4C", "gold_field": "4C", "gold_line": "4C\t4r", "gold_note_token": null, "gold_piece": "4C", "gold_token_id": 181, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 47, "line_char_start": 42, "line_index": 2, "line_text": "4C\t4r", "operation": "replace", "pred_field": "*M3/4", "pred_local_replacement_line": "*M3/4\t4r", "pred_note_token": null, "pred_piece": "*M3/4", "pred_token_id": 283, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 22.948473282442748, "LER": 71.34146118164062, "SER": 21.813515825491873}, "nll": 1.16622793674469, "top1_error_count": 255, "top1_error_rate": 0.21794871794871795}}
+{"dataset_index": 57, "sample_id": "polish_00087", "source_ids": [], "target_token_count": 1300, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "null_token", "char_end": 43, "char_start": 42, "component_category": "null_token", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": ".", "gold_field": ".", "gold_line": ".\t8ccq( 8ee-q(", "gold_note_token": null, "gold_piece": ".", "gold_token_id": 14, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 56, "line_char_start": 42, "line_index": 2, "line_text": ".\t8ccq( 8ee-q(", "operation": "insert", "pred_field": "*", "pred_local_replacement_line": "*\t8ccq( 8ee-q(", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 18.75543951261967, "LER": 63.58695602416992, "SER": 16.551193225558123}, "nll": 0.8573907017707825, "top1_error_count": 215, "top1_error_rate": 0.16538461538461538}}
+{"dataset_index": 58, "sample_id": "polish_00089", "source_ids": [], "target_token_count": 782, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 37, "char_start": 35, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 35, "field_index": 0, "field_text": "8A( 8c(L", "gold_field": "8A( 8c(L", "gold_line": "8A( 8c(L\t2.D\t16g#(L", "gold_note_token": null, "gold_piece": "8A", "gold_token_id": 109, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 35, "line_index": 3, "line_text": "8A( 8c(L\t2.D\t16g#(L", "operation": "replace", "pred_field": "*( 8c(L", "pred_local_replacement_line": "*( 8c(L\t2.D\t16g#(L", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 17.801753202966957, "LER": 76.39751434326172, "SER": 51.21638924455826}, "nll": 0.9439278841018677, "top1_error_count": 141, "top1_error_rate": 0.18030690537084398}}
+{"dataset_index": 59, "sample_id": "polish_00090", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 46, "field_index": 2, "field_text": "8a(L", "gold_field": "8a(L", "gold_line": "8BB\t8r\t8a(L", "gold_note_token": "8a(L", "gold_piece": "8a(L", "gold_token_id": 661, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 50, "line_char_start": 39, "line_index": 2, "line_text": "8BB\t8r\t8a(L", "operation": "replace", "pred_field": "8a", "pred_local_replacement_line": "8BB\t8r\t8a", "pred_note_token": "8a", "pred_piece": "8a", "pred_token_id": 76, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 17.052980132450333, "LER": 57.71812057495117, "SER": 13.87900355871886}, "nll": 0.8659428358078003, "top1_error_count": 156, "top1_error_rate": 0.13866666666666666}}
+{"dataset_index": 60, "sample_id": "polish_00092", "source_ids": [], "target_token_count": 1153, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 43, "char_start": 38, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 38, "field_index": 0, "field_text": "8EE-L", "gold_field": "8EE-L", "gold_line": "8EE-L\t8r", "gold_note_token": null, "gold_piece": "8EE-L", "gold_token_id": 1038, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 46, "line_char_start": 38, "line_index": 2, "line_text": "8EE-L\t8r", "operation": "replace", "pred_field": "*M4/4", "pred_local_replacement_line": "*M4/4\t8r", "pred_note_token": null, "pred_piece": "*M4/4", "pred_token_id": 229, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 9.09970958373669, "LER": 46.36871337890625, "SER": 9.288194444444445}, "nll": 0.5908276438713074, "top1_error_count": 107, "top1_error_rate": 0.09280138768430182}}
+{"dataset_index": 61, "sample_id": "polish_00093", "source_ids": [], "target_token_count": 1047, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 48, "char_start": 46, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 48, "field_char_start": 46, "field_index": 2, "field_text": "4b", "gold_field": "4b", "gold_line": "8EE\t8r\t4b", "gold_note_token": "4b", "gold_piece": "4b", "gold_token_id": 131, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 48, "line_char_start": 39, "line_index": 2, "line_text": "8EE\t8r\t4b", "operation": "replace", "pred_field": "4.b", "pred_local_replacement_line": "8EE\t8r\t4.b", "pred_note_token": "4.b", "pred_piece": "4.b", "pred_token_id": 344, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 13.305736250739209, "LER": 51.09489059448242, "SER": 12.045889101338432}, "nll": 0.7508025169372559, "top1_error_count": 126, "top1_error_rate": 0.12034383954154727}}
+{"dataset_index": 62, "sample_id": "polish_00094", "source_ids": [], "target_token_count": 773, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "pitch", "char_end": 46, "char_start": 44, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 47, "field_char_start": 40, "field_index": 1, "field_text": "32qqggL", "gold_field": "32qqggL", "gold_line": ".\t32qqggL", "gold_note_token": "32qqggL", "gold_piece": "gg", "gold_token_id": 88, "gold_token_index": 11, "gold_trimmed_index": 11, "group": "visual", "line_char_end": 47, "line_char_start": 38, "line_index": 2, "line_text": ".\t32qqggL", "operation": "replace", "pred_field": "32qqaaL", "pred_local_replacement_line": ".\t32qqaaL", "pred_note_token": "32qqaaL", "pred_piece": "aa", "pred_token_id": 197, "pred_trimmed_index": 11, "visually_grounded": true}, "metrics": {"CER": 23.31081081081081, "LER": 63.870967864990234, "SER": 22.53886010362694}, "nll": 1.2768330574035645, "top1_error_count": 174, "top1_error_rate": 0.22509702457956016}}
+{"dataset_index": 63, "sample_id": "polish_00095", "source_ids": [], "target_token_count": 1073, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 50, "component_category": "chord_structure_only", "component_flags": ["chord_structure"], "component_group": "visual", "field_char_end": 55, "field_char_start": 47, "field_index": 0, "field_text": "4D- 4B--", "gold_field": "4D- 4B--", "gold_line": "4D- 4B--\t12g-L\t12r", "gold_note_token": "4D-", "gold_piece": " ", "gold_token_id": 6, "gold_token_index": 13, "gold_trimmed_index": 13, "group": "visual", "line_char_end": 65, "line_char_start": 47, "line_index": 3, "line_text": "4D- 4B--\t12g-L\t12r", "operation": "replace", "pred_field": "4D-:4B--", "pred_local_replacement_line": "4D-:4B--\t12g-L\t12r", "pred_note_token": "4D-:4B--", "pred_piece": ":", "pred_token_id": 26, "pred_trimmed_index": 13, "visually_grounded": true}, "metrics": {"CER": 12.454212454212454, "LER": 42.26803970336914, "SER": 9.794776119402986}, "nll": 0.5109670162200928, "top1_error_count": 105, "top1_error_rate": 0.097856477166822}}
+{"dataset_index": 64, "sample_id": "polish_00096", "source_ids": [], "target_token_count": 935, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "clef", "char_end": 15, "char_start": 8, "component_category": "clef", "component_flags": [], "component_group": "visual", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2", "gold_note_token": null, "gold_piece": "*clefG2", "gold_token_id": 147, "gold_token_index": 2, "gold_trimmed_index": 2, "group": "visual", "line_char_end": 15, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2", "operation": "replace", "pred_field": "*clefG", "pred_local_replacement_line": "*clefF4\t*clefG", "pred_note_token": null, "pred_piece": "*clefG", "pred_token_id": 145, "pred_trimmed_index": 2, "visually_grounded": true}, "metrics": {"CER": 10.824742268041238, "LER": 47.85714340209961, "SER": 11.349036402569594}, "nll": 0.6287577152252197, "top1_error_count": 106, "top1_error_rate": 0.11336898395721925}}
+{"dataset_index": 65, "sample_id": "polish_00097", "source_ids": [], "target_token_count": 1050, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 63, "char_start": 61, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 63, "field_char_start": 57, "field_index": 1, "field_text": "16gg[L", "gold_field": "16gg[L", "gold_line": "8c 8eJ\t16gg[L\t16dd(L", "gold_note_token": "16gg[L", "gold_piece": "[L", "gold_token_id": 562, "gold_token_index": 23, "gold_trimmed_index": 23, "group": "visual", "line_char_end": 70, "line_char_start": 50, "line_index": 3, "line_text": "8c 8eJ\t16gg[L\t16dd(L", "operation": "replace", "pred_field": "16gg[LL", "pred_local_replacement_line": "8c 8eJ\t16gg[LL\t16dd(L", "pred_note_token": "16gg[LL", "pred_piece": "[LL", "pred_token_id": 469, "pred_trimmed_index": 23, "visually_grounded": true}, "metrics": {"CER": 12.687913409500903, "LER": 54.47761154174805, "SER": 12.392755004766444}, "nll": 0.7960441708564758, "top1_error_count": 130, "top1_error_rate": 0.12380952380952381}}
+{"dataset_index": 66, "sample_id": "polish_00100", "source_ids": [], "target_token_count": 1185, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2", "gold_note_token": null, "gold_piece": "\n", "gold_token_id": 5, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 15, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2", "operation": "replace", "pred_field": "*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\t", "pred_note_token": null, "pred_piece": "\t", "pred_token_id": 4, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 22.818791946308725, "LER": 64.02116394042969, "SER": 19.594594594594593}, "nll": 1.0010433197021484, "top1_error_count": 232, "top1_error_rate": 0.19578059071729959}}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/summary.json

@@ -0,0 +1,498 @@
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-300k-8ep-modal-h200-4gpu/model.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.19623294253315396,
+      "errors": 2042,
+      "errors_per_1k_tokens": 27.4038784137422,
+      "errors_per_page_mean": 30.47761194029851,
+      "errors_per_page_median": 31.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.1619258120315203,
+      "errors": 1685,
+      "errors_per_1k_tokens": 22.61289673220157,
+      "errors_per_page_mean": 25.149253731343283,
+      "errors_per_page_median": 25.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.10676532769556026,
+      "errors": 1111,
+      "errors_per_1k_tokens": 14.909749714822519,
+      "errors_per_page_mean": 16.582089552238806,
+      "errors_per_page_median": 14.0,
+      "page_share": 0.9701492537313433,
+      "pages_with_error": 65
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.08062656159907745,
+      "errors": 839,
+      "errors_per_1k_tokens": 11.25947795745823,
+      "errors_per_page_mean": 12.522388059701493,
+      "errors_per_page_median": 7.0,
+      "page_share": 0.9701492537313433,
+      "pages_with_error": 65
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.06217566788391313,
+      "errors": 647,
+      "errors_per_1k_tokens": 8.682815540495202,
+      "errors_per_page_mean": 9.656716417910447,
+      "errors_per_page_median": 9.0,
+      "page_share": 0.9850746268656716,
+      "pages_with_error": 66
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.06102248702671536,
+      "errors": 635,
+      "errors_per_1k_tokens": 8.521774139435013,
+      "errors_per_page_mean": 9.477611940298507,
+      "errors_per_page_median": 9.0,
+      "page_share": 0.9850746268656716,
+      "pages_with_error": 66
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05362290985969633,
+      "errors": 558,
+      "errors_per_1k_tokens": 7.4884251492987985,
+      "errors_per_page_mean": 8.328358208955224,
+      "errors_per_page_median": 7.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.03478762252546608,
+      "errors": 362,
+      "errors_per_1k_tokens": 4.8580822653157085,
+      "errors_per_page_mean": 5.402985074626866,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.9552238805970149,
+      "pages_with_error": 64
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.0331539496444359,
+      "errors": 345,
+      "errors_per_1k_tokens": 4.62994028048044,
+      "errors_per_page_mean": 5.149253731343284,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.8507462686567164,
+      "pages_with_error": 57
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.032673457620603495,
+      "errors": 340,
+      "errors_per_1k_tokens": 4.562839696705361,
+      "errors_per_page_mean": 5.074626865671642,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.8059701492537313,
+      "pages_with_error": 54
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.023448010763021335,
+      "errors": 244,
+      "errors_per_1k_tokens": 3.2745084882238475,
+      "errors_per_page_mean": 3.6417910447761193,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.6865671641791045,
+      "pages_with_error": 46
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.019411877762829138,
+      "errors": 202,
+      "errors_per_1k_tokens": 2.710863584513185,
+      "errors_per_page_mean": 3.014925373134328,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.8059701492537313,
+      "pages_with_error": 54
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.017874303286565443,
+      "errors": 186,
+      "errors_per_1k_tokens": 2.496141716432933,
+      "errors_per_page_mean": 2.7761194029850746,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.7313432835820896,
+      "pages_with_error": 49
+    },
+    {
+      "category": "barline",
+      "error_share": 0.016336728810301748,
+      "errors": 170,
+      "errors_per_1k_tokens": 2.2814198483526806,
+      "errors_per_page_mean": 2.537313432835821,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.8507462686567164,
+      "pages_with_error": 57
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.013645973476840285,
+      "errors": 142,
+      "errors_per_1k_tokens": 1.905656579212239,
+      "errors_per_page_mean": 2.1194029850746268,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6865671641791045,
+      "pages_with_error": 46
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.011147414952911782,
+      "errors": 116,
+      "errors_per_1k_tokens": 1.5567335435818292,
+      "errors_per_page_mean": 1.7313432835820894,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5970149253731343,
+      "pages_with_error": 40
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.00999423409571401,
+      "errors": 104,
+      "errors_per_1k_tokens": 1.39569214252164,
+      "errors_per_page_mean": 1.5522388059701493,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.7313432835820896,
+      "pages_with_error": 49
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.00989813569094753,
+      "errors": 103,
+      "errors_per_1k_tokens": 1.3822720257666241,
+      "errors_per_page_mean": 1.537313432835821,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5223880597014925,
+      "pages_with_error": 35
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.008072266000384394,
+      "errors": 84,
+      "errors_per_1k_tokens": 1.1272898074213247,
+      "errors_per_page_mean": 1.2537313432835822,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5373134328358209,
+      "pages_with_error": 36
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.007207380357486066,
+      "errors": 75,
+      "errors_per_1k_tokens": 1.0065087566261826,
+      "errors_per_page_mean": 1.1194029850746268,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.44776119402985076,
+      "pages_with_error": 30
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.0065346915241206996,
+      "errors": 68,
+      "errors_per_1k_tokens": 0.9125679393410723,
+      "errors_per_page_mean": 1.0149253731343284,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5671641791044776,
+      "pages_with_error": 38
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005573707476455891,
+      "errors": 58,
+      "errors_per_1k_tokens": 0.7783667717909146,
+      "errors_per_page_mean": 0.8656716417910447,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.8208955223880597,
+      "pages_with_error": 55
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004708821833557563,
+      "errors": 49,
+      "errors_per_1k_tokens": 0.6575857209957726,
+      "errors_per_page_mean": 0.7313432835820896,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3582089552238806,
+      "pages_with_error": 24
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.00432442821449164,
+      "errors": 45,
+      "errors_per_1k_tokens": 0.6039052539757096,
+      "errors_per_page_mean": 0.6716417910447762,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.31343283582089554,
+      "pages_with_error": 21
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.003747837785892754,
+      "errors": 39,
+      "errors_per_1k_tokens": 0.523384553445615,
+      "errors_per_page_mean": 0.582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.29850746268656714,
+      "pages_with_error": 20
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.0028829521429944265,
+      "errors": 30,
+      "errors_per_1k_tokens": 0.40260350265047307,
+      "errors_per_page_mean": 0.44776119402985076,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.16417910447761194,
+      "pages_with_error": 11
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0027868537382279455,
+      "errors": 29,
+      "errors_per_1k_tokens": 0.3891833858954573,
+      "errors_per_page_mean": 0.43283582089552236,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0024024601191620218,
+      "errors": 25,
+      "errors_per_1k_tokens": 0.33550291887539424,
+      "errors_per_page_mean": 0.373134328358209,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.29850746268656714,
+      "pages_with_error": 20
+    },
+    {
+      "category": "clef",
+      "error_share": 0.0023063617143955412,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.19402985074626866,
+      "pages_with_error": 13
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.0020180665000960984,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.208955223880597,
+      "pages_with_error": 14
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.001537574476263694,
+      "errors": 16,
+      "errors_per_1k_tokens": 0.2147218680802523,
+      "errors_per_page_mean": 0.23880597014925373,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.13432835820895522,
+      "pages_with_error": 9
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0003843936190659235,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.04477611940298507,
+      "pages_with_error": 3
+    },
+    {
+      "category": "chord_structure_and_pitch",
+      "error_share": 0.00028829521429944266,
+      "errors": 3,
+      "errors_per_1k_tokens": 0.040260350265047304,
+      "errors_per_page_mean": 0.04477611940298507,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.029850746268656716,
+      "pages_with_error": 2
+    },
+    {
+      "category": "chord_structure_and_pitch_and_accidental",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "accidental_and_tie_slur_and_articulation",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "duration_and_accidental_and_tie_slur",
+      "error_share": 9.609840476648087e-05,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 1,
+    "accidental_and_tie_slur": 68,
+    "accidental_and_tie_slur_and_articulation": 1,
+    "accidental_only": 362,
+    "articulation_only": 30,
+    "barline": 170,
+    "chord_structure_and_accidental": 25,
+    "chord_structure_and_accidental_and_articulation": 4,
+    "chord_structure_and_accidental_and_tie_slur": 45,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 21,
+    "chord_structure_and_pitch": 3,
+    "chord_structure_and_pitch_and_accidental": 1,
+    "chord_structure_and_tie_slur": 202,
+    "chord_structure_and_tie_slur_and_articulation": 16,
+    "chord_structure_only": 142,
+    "clef": 24,
+    "column_or_line_break": 558,
+    "duration_and_accidental": 49,
+    "duration_and_accidental_and_tie_slur": 1,
+    "duration_and_pitch": 345,
+    "duration_and_pitch_and_accidental": 186,
+    "duration_and_pitch_and_accidental_and_tie_slur": 39,
+    "duration_and_pitch_and_tie_slur": 75,
+    "duration_and_tie_slur": 104,
+    "duration_only": 635,
+    "interpretation_or_control": 647,
+    "note_other": 2042,
+    "null_token": 116,
+    "pitch_and_accidental": 839,
+    "pitch_and_accidental_and_tie_slur": 103,
+    "pitch_and_tie_slur": 84,
+    "pitch_only": 1111,
+    "rest_note_confusion": 340,
+    "spine_operation": 244,
+    "termination": 58,
+    "tie_slur_and_articulation": 29,
+    "tie_slur_only": 1685
+  },
+  "component_error_groups": {
+    "structural": 1623,
+    "visual": 8783
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 16.40937467990443,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 4634,
+    "articulation": 25,
+    "barline": 153,
+    "clef": 26,
+    "column_or_line_break": 558,
+    "duration": 1654,
+    "interpretation_or_control": 142,
+    "note_other": 179,
+    "null_token": 116,
+    "pitch": 40,
+    "rest": 389,
+    "spine_operation": 335,
+    "termination": 55,
+    "tie_slur": 2100
+  },
+  "gold_error_groups": {
+    "structural": 1206,
+    "visual": 9200
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 14.556106567382812,
+      "LER": 53.25029754638672,
+      "SER": 15.089727401733398,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.1396497349526941,
+      "top1_errors": 10406,
+      "top1_errors_per_page": 155.3134328358209
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 15,
+    "predicted_unk_literal": 61
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/summary.md

@@ -0,0 +1,43 @@
+# Transcoda Dataset Eval
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-300k-8ep-modal-h200-4gpu/model.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Dataset kind: `hf_parquet_split`
+- Mode: `teacher_forced`
+- Pages: `67`
+- Tokens: `74,515`
+- Elapsed: `16.4s`
+
+## Metrics
+
+| mode | CER | SER | LER | exact pages | top-1 error rate | top-1 errors/page |
+|---|---:|---:|---:|---:|---:|---:|
+| `teacher_forced` | 14.556107 | 15.089727 | 53.250298 | 0.000% | 13.9650% | 155.313 |
+
+## Teacher-Forced Error Categories
+
+| category | errors |
+|---|---:|
+| `accidental` | 4,634 |
+| `tie_slur` | 2,100 |
+| `duration` | 1,654 |
+| `column_or_line_break` | 558 |
+| `rest` | 389 |
+| `spine_operation` | 335 |
+| `note_other` | 179 |
+| `barline` | 153 |
+| `interpretation_or_control` | 142 |
+| `null_token` | 116 |
+| `termination` | 55 |
+| `pitch` | 40 |
+| `clef` | 26 |
+| `articulation` | 25 |
+
+## Outputs
+
+- `summary.json`
+- `per_page_records.jsonl`
+- `teacher_forced_predictions.jsonl` when teacher-forced mode is enabled
+- `free_generation_predictions.jsonl` when free-generation mode is enabled
+- `gold_prefix_errors.jsonl` when teacher-forced mode is enabled
+- `teacher_forced_error_report.md` when teacher-forced mode is enabled

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/convnext2_300k_8ep/teacher_forced_error_report.md

@@ -0,0 +1,530 @@
+# Teacher-Forced Error Report
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-300k-8ep-modal-h200-4gpu/model.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Split: `eval`
+- Pages: `67`
+- Tokens: `74,515`
+- Top-1 errors: `10,406`
+- Top-1 error rate: `13.965%`
+
+## Component Buckets
+
+| bucket | errors | error share | mean/page | median/page | pages | errors/1k tok |
+|---|---:|---:|---:|---:|---:|---:|
+| `note_other` | 2,042 | 19.62% | 30.478 | 31.000 | 67 (100.0%) | 27.404 |
+| `tie_slur_only` | 1,685 | 16.19% | 25.149 | 25.000 | 67 (100.0%) | 22.613 |
+| `pitch_only` | 1,111 | 10.68% | 16.582 | 14.000 | 65 (97.0%) | 14.910 |
+| `pitch_and_accidental` | 839 | 8.06% | 12.522 | 7.000 | 65 (97.0%) | 11.259 |
+| `interpretation_or_control` | 647 | 6.22% | 9.657 | 9.000 | 66 (98.5%) | 8.683 |
+| `duration_only` | 635 | 6.10% | 9.478 | 9.000 | 66 (98.5%) | 8.522 |
+| `column_or_line_break` | 558 | 5.36% | 8.328 | 7.000 | 67 (100.0%) | 7.488 |
+| `accidental_only` | 362 | 3.48% | 5.403 | 4.000 | 64 (95.5%) | 4.858 |
+| `duration_and_pitch` | 345 | 3.32% | 5.149 | 4.000 | 57 (85.1%) | 4.630 |
+| `rest_note_confusion` | 340 | 3.27% | 5.075 | 4.000 | 54 (80.6%) | 4.563 |
+| `spine_operation` | 244 | 2.34% | 3.642 | 2.000 | 46 (68.7%) | 3.275 |
+| `chord_structure_and_tie_slur` | 202 | 1.94% | 3.015 | 2.000 | 54 (80.6%) | 2.711 |
+| `duration_and_pitch_and_accidental` | 186 | 1.79% | 2.776 | 2.000 | 49 (73.1%) | 2.496 |
+| `barline` | 170 | 1.63% | 2.537 | 2.000 | 57 (85.1%) | 2.281 |
+| `chord_structure_only` | 142 | 1.36% | 2.119 | 1.000 | 46 (68.7%) | 1.906 |
+| `null_token` | 116 | 1.11% | 1.731 | 1.000 | 40 (59.7%) | 1.557 |
+| `duration_and_tie_slur` | 104 | 1.00% | 1.552 | 1.000 | 49 (73.1%) | 1.396 |
+| `pitch_and_accidental_and_tie_slur` | 103 | 0.99% | 1.537 | 1.000 | 35 (52.2%) | 1.382 |
+| `pitch_and_tie_slur` | 84 | 0.81% | 1.254 | 1.000 | 36 (53.7%) | 1.127 |
+| `duration_and_pitch_and_tie_slur` | 75 | 0.72% | 1.119 | 0.000 | 30 (44.8%) | 1.007 |
+| `accidental_and_tie_slur` | 68 | 0.65% | 1.015 | 1.000 | 38 (56.7%) | 0.913 |
+| `termination` | 58 | 0.56% | 0.866 | 1.000 | 55 (82.1%) | 0.778 |
+| `duration_and_accidental` | 49 | 0.47% | 0.731 | 0.000 | 24 (35.8%) | 0.658 |
+| `chord_structure_and_accidental_and_tie_slur` | 45 | 0.43% | 0.672 | 0.000 | 21 (31.3%) | 0.604 |
+| `duration_and_pitch_and_accidental_and_tie_slur` | 39 | 0.37% | 0.582 | 0.000 | 20 (29.9%) | 0.523 |
+| `articulation_only` | 30 | 0.29% | 0.448 | 0.000 | 11 (16.4%) | 0.403 |
+| `tie_slur_and_articulation` | 29 | 0.28% | 0.433 | 0.000 | 16 (23.9%) | 0.389 |
+| `chord_structure_and_accidental` | 25 | 0.24% | 0.373 | 0.000 | 20 (29.9%) | 0.336 |
+| `clef` | 24 | 0.23% | 0.358 | 0.000 | 13 (19.4%) | 0.322 |
+| `chord_structure_and_articulation` | 21 | 0.20% | 0.313 | 0.000 | 14 (20.9%) | 0.282 |
+| `chord_structure_and_tie_slur_and_articulation` | 16 | 0.15% | 0.239 | 0.000 | 9 (13.4%) | 0.215 |
+| `chord_structure_and_accidental_and_articulation` | 4 | 0.04% | 0.060 | 0.000 | 3 (4.5%) | 0.054 |
+| `chord_structure_and_pitch` | 3 | 0.03% | 0.045 | 0.000 | 2 (3.0%) | 0.040 |
+| `chord_structure_and_pitch_and_accidental` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+| `accidental_and_articulation` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+| `accidental_and_tie_slur_and_articulation` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+| `chord_structure_and_accidental_and_tie_slur_and_articulation` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+| `duration_and_accidental_and_tie_slur` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+
+## Legacy Regex Buckets
+
+These are retained for compatibility and should not be treated as component diffs.
+
+| bucket | errors |
+|---|---:|
+| `accidental` | 4,634 |
+| `tie_slur` | 2,100 |
+| `duration` | 1,654 |
+| `column_or_line_break` | 558 |
+| `rest` | 389 |
+| `spine_operation` | 335 |
+| `note_other` | 179 |
+| `barline` | 153 |
+| `interpretation_or_control` | 142 |
+| `null_token` | 116 |
+| `termination` | 55 |
+| `pitch` | 40 |
+| `clef` | 26 |
+| `articulation` | 25 |
+
+## Error Samples
+
+### Sample 1
+
+```text
+8BB-L	2.ddd-;[ 2.dddd-;[
+teacher-forced top-1 local replacement:
+8BB-L	4dd-d-;[ 2.dddd-;[
+```
+
+### Sample 2
+
+```text
+8ffL	4aa-) 4aaa-)
+teacher-forced top-1 local replacement:
+8ffL	4aa-) 4aaa-
+
+```
+
+### Sample 3
+
+```text
+4r	16aa]L
+teacher-forced top-1 local replacement:
+4r	8.aa]L
+```
+
+### Sample 4
+
+```text
+.	16e)J	.
+teacher-forced top-1 local replacement:
+.	16e)JJ	.
+```
+
+### Sample 5
+
+```text
+*^	*
+teacher-forced top-1 local replacement:
+*^	*^
+```
+
+### Sample 6
+
+```text
+8CC 8GG 8C	8cc)J	.
+teacher-forced top-1 local replacement:
+8CC 8GG 8BB	8cc)J	.
+```
+
+### Sample 7
+
+```text
+4F# 4d#	2b[
+teacher-forced top-1 local replacement:
+4F# 4d#	2b
+```
+
+### Sample 8
+
+```text
+16C#[	.
+teacher-forced top-1 local replacement:
+16C		.
+```
+
+### Sample 9
+
+```text
+.	12en
+teacher-forced top-1 local replacement:
+.	12e-n
+```
+
+### Sample 10
+
+```text
+8c 8e-	16dd(L
+teacher-forced top-1 local replacement:
+8c 8e-	8dd
+```
+
+### Sample 11
+
+```text
+*	*v	*v
+teacher-forced top-1 local replacement:
+=	*v	*v
+```
+
+### Sample 12
+
+```text
+4CC 4C	8g 8cc 8eenL
+teacher-forced top-1 local replacement:
+4CC 4C	8g(8cc 8eenL
+```
+
+### Sample 13
+
+```text
+.	.	16e-J
+teacher-forced top-1 local replacement:
+.	.	16e-JJ
+```
+
+### Sample 14
+
+```text
+.	32cc
+teacher-forced top-1 local replacement:
+.	32cc 
+```
+
+### Sample 15
+
+```text
+.	.	2A) 2c) 2f)
+teacher-forced top-1 local replacement:
+.	.	2A) 2c  2f)
+```
+
+### Sample 16
+
+```text
+8BBB	4dd#(
+teacher-forced top-1 local replacement:
+8BBB	4dd#)
+```
+
+
+## Sample Metadata
+
+### Sample 1
+
+- Sample: `polish_00056`
+- Row/token: `38` / `153`
+- Line/field: `26:1`
+- Component bucket: `duration_and_pitch`; flags: `duration, pitch`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `2.dd` -> `4dd-`
+- Gold note -> top-1 note: `2.ddd-;[` -> `4dd-d-;[`
+- Gold token likelihood: rank `5`, p `0.660%`
+- Predicted token likelihood: p `59.371%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 567 | `"4dd-"` | 59.371% |
+| 2 | 108 | `"4dd"` | 19.275% |
+| 3 | 1388 | `"2dd-"` | 6.258% |
+| 4 | 277 | `"2dd"` | 4.168% |
+| 5 | 588 | `"2.dd"` | 0.660% |
+
+### Sample 2
+
+- Sample: `polish_00095`
+- Row/token: `63` / `968`
+- Line/field: `172:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `)` -> `
+`
+- Gold note -> top-1 note: `4aaa-)` -> `4aaa-
+`
+- Gold token likelihood: rank `2`, p `0.042%`
+- Predicted token likelihood: p `92.013%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 5 | `"\n"` | 92.013% |
+| 2 | 10 | `")"` | 0.042% |
+| 3 | 6 | `" "` | 0.041% |
+| 4 | 55 | `"i"` | 0.022% |
+| 5 | 4 | `"\t"` | 0.016% |
+
+### Sample 3
+
+- Sample: `polish_00074`
+- Row/token: `51` / `1280`
+- Line/field: `183:1`
+- Component bucket: `duration_only`; flags: `duration`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16aa` -> `8.aa`
+- Gold note -> top-1 note: `16aa]L` -> `8.aa]L`
+- Gold token likelihood: rank `8`, p `0.044%`
+- Predicted token likelihood: p `98.092%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 386 | `"8.aa"` | 98.092% |
+| 2 | 484 | `"8.ff"` | 0.084% |
+| 3 | 1243 | `"8.bb"` | 0.077% |
+| 4 | 409 | `"8.ff#"` | 0.065% |
+| 5 | 272 | `"8.ee"` | 0.063% |
+
+### Sample 4
+
+- Sample: `polish_00032`
+- Row/token: `21` / `661`
+- Line/field: `122:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `)J` -> `)JJ`
+- Gold note -> top-1 note: `16e)J` -> `16e)JJ`
+- Gold token likelihood: rank `4`, p `0.516%`
+- Predicted token likelihood: p `84.181%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 305 | `")JJ"` | 84.181% |
+| 2 | 10 | `")"` | 1.542% |
+| 3 | 1570 | `")JJJ"` | 0.825% |
+| 4 | 186 | `")J"` | 0.516% |
+| 5 | 2754 | `")LL"` | 0.344% |
+
+### Sample 5
+
+- Sample: `polish_00015`
+- Row/token: `9` / `10`
+- Line/field: `2:1`
+- Component bucket: `spine_operation`; flags: `none`
+- Legacy bucket: `spine_operation`
+- Gold piece -> top-1 piece: `*` -> `*^`
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `2`, p `2.557%`
+- Predicted token likelihood: p `90.148%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 191 | `"*^"` | 90.148% |
+| 2 | 11 | `"*"` | 2.557% |
+| 3 | 692 | `"*met(c)"` | 0.033% |
+| 4 | 1794 | `"*8va"` | 0.027% |
+| 5 | 1245 | `"*met(c\|)"` | 0.014% |
+
+### Sample 6
+
+- Sample: `polish_00038`
+- Row/token: `25` / `472`
+- Line/field: `58:0`
+- Component bucket: `pitch_only`; flags: `pitch`
+- Legacy bucket: `duration`
+- Gold piece -> top-1 piece: `8C` -> `8BB`
+- Gold note -> top-1 note: `8C` -> `8BB`
+- Gold token likelihood: rank `2`, p `1.194%`
+- Predicted token likelihood: p `81.125%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 238 | `"8BB"` | 81.125% |
+| 2 | 166 | `"8C"` | 1.194% |
+| 3 | 275 | `"8BB-"` | 0.364% |
+| 4 | 152 | `"8D"` | 0.311% |
+| 5 | 221 | `"8GG"` | 0.293% |
+
+### Sample 7
+
+- Sample: `polish_00063`
+- Row/token: `45` / `369`
+- Line/field: `61:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `2b[` -> `2b`
+- Gold note -> top-1 note: `2b[` -> `2b`
+- Gold token likelihood: rank `5`, p `0.101%`
+- Predicted token likelihood: p `94.566%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 331 | `"2b"` | 94.566% |
+| 2 | 720 | `"2.b"` | 0.134% |
+| 3 | 277 | `"2dd"` | 0.134% |
+| 4 | 731 | `"2cc#"` | 0.104% |
+| 5 | 1657 | `"2b["` | 0.101% |
+
+### Sample 8
+
+- Sample: `polish_00100`
+- Row/token: `66` / `140`
+- Line/field: `21:0`
+- Component bucket: `accidental_and_tie_slur`; flags: `accidental, tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `#[` -> `	`
+- Gold note -> top-1 note: `16C#[` -> `16C`
+- Gold token likelihood: rank `5`, p `0.127%`
+- Predicted token likelihood: p `85.784%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 4 | `"\t"` | 85.784% |
+| 2 | 8 | `"#"` | 2.148% |
+| 3 | 59 | `"n"` | 0.423% |
+| 4 | 242 | `"#]"` | 0.340% |
+| 5 | 428 | `"#["` | 0.127% |
+
+### Sample 9
+
+- Sample: `polish_00030`
+- Row/token: `19` / `1087`
+- Line/field: `180:1`
+- Component bucket: `accidental_only`; flags: `accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `12e` -> `12e-`
+- Gold note -> top-1 note: `12en` -> `12e-n`
+- Gold token likelihood: rank `7`, p `0.119%`
+- Predicted token likelihood: p `95.720%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1359 | `"12e-"` | 95.720% |
+| 2 | 739 | `"12c"` | 0.197% |
+| 3 | 1363 | `"12ee-"` | 0.191% |
+| 4 | 2795 | `"6e-"` | 0.168% |
+| 5 | 1972 | `"12e-J"` | 0.158% |
+
+### Sample 10
+
+- Sample: `polish_00084`
+- Row/token: `54` / `527`
+- Line/field: `91:1`
+- Component bucket: `duration_and_tie_slur`; flags: `duration, tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16dd(L` -> `8dd`
+- Gold note -> top-1 note: `16dd(L` -> `8dd`
+- Gold token likelihood: rank `31`, p `0.436%`
+- Predicted token likelihood: p `19.448%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 79 | `"8dd"` | 19.448% |
+| 2 | 82 | `"8cc"` | 5.749% |
+| 3 | 112 | `"16dd"` | 4.339% |
+| 4 | 108 | `"4dd"` | 4.339% |
+| 5 | 330 | `"8ee-"` | 2.982% |
+
+### Sample 11
+
+- Sample: `polish_00072`
+- Row/token: `50` / `308`
+- Line/field: `54:0`
+- Component bucket: `interpretation_or_control`; flags: `none`
+- Legacy bucket: `interpretation_or_control`
+- Gold piece -> top-1 piece: `*` -> `=`
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `2`, p `0.233%`
+- Predicted token likelihood: p `91.289%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 28 | `"="` | 91.289% |
+| 2 | 11 | `"*"` | 0.233% |
+| 3 | 157 | `"*v"` | 0.023% |
+| 4 | 191 | `"*^"` | 0.019% |
+| 5 | 14 | `"."` | 0.018% |
+
+### Sample 12
+
+- Sample: `polish_00060`
+- Row/token: `42` / `1487`
+- Line/field: `207:1`
+- Component bucket: `chord_structure_and_tie_slur`; flags: `tie_slur, chord_structure`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: ` ` -> `(`
+- Gold note -> top-1 note: `8g` -> `8g(8cc`
+- Gold token likelihood: rank `2`, p `10.132%`
+- Predicted token likelihood: p `51.456%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 9 | `"("` | 51.456% |
+| 2 | 6 | `" "` | 10.132% |
+| 3 | 47 | `"\`"` | 2.995% |
+| 4 | 26 | `":"` | 2.562% |
+| 5 | 64 | `"t"` | 0.942% |
+
+### Sample 13
+
+- Sample: `polish_00023`
+- Row/token: `14` / `375`
+- Line/field: `48:2`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `16e-J` -> `16e-JJ`
+- Gold note -> top-1 note: `16e-J` -> `16e-JJ`
+- Gold token likelihood: rank `3`, p `0.154%`
+- Predicted token likelihood: p `96.296%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 711 | `"16e-JJ"` | 96.296% |
+| 2 | 394 | `"16e-"` | 0.306% |
+| 3 | 2425 | `"16e-J"` | 0.154% |
+| 4 | 1072 | `"32e-"` | 0.096% |
+| 5 | 687 | `"16ee-JJ"` | 0.091% |
+
+### Sample 14
+
+- Sample: `polish_00059`
+- Row/token: `41` / `1769`
+- Line/field: `248:1`
+- Component bucket: `column_or_line_break`; flags: `none`
+- Legacy bucket: `column_or_line_break`
+- Gold piece -> top-1 piece: `
+` -> ` `
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `27`, p `0.236%`
+- Predicted token likelihood: p `19.672%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 6 | `" "` | 19.672% |
+| 2 | 27 | `";"` | 15.321% |
+| 3 | 40 | `"T"` | 4.821% |
+| 4 | 39 | `"M"` | 2.924% |
+| 5 | 59 | `"n"` | 2.834% |
+
+### Sample 15
+
+- Sample: `polish_00074`
+- Row/token: `51` / `239`
+- Line/field: `32:2`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `)` -> ` `
+- Gold note -> top-1 note: `2c)` -> `2c`
+- Gold token likelihood: rank `2`, p `1.470%`
+- Predicted token likelihood: p `88.166%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 6 | `" "` | 88.166% |
+| 2 | 10 | `")"` | 1.470% |
+| 3 | 46 | `"_"` | 0.508% |
+| 4 | 44 | `"]"` | 0.463% |
+| 5 | 59 | `"n"` | 0.448% |
+
+### Sample 16
+
+- Sample: `polish_00011`
+- Row/token: `6` / `892`
+- Line/field: `156:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `(` -> `)`
+- Gold note -> top-1 note: `4dd#(` -> `4dd#)`
+- Gold token likelihood: rank `2`, p `9.564%`
+- Predicted token likelihood: p `77.619%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 10 | `")"` | 77.619% |
+| 2 | 9 | `"("` | 9.564% |
+| 3 | 2909 | `"(["` | 0.715% |
+| 4 | 5 | `"\n"` | 0.395% |
+| 5 | 27 | `";"` | 0.298% |

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/command.log

@@ -0,0 +1,502 @@
+Skipping import of cpp extensions due to incompatible torch version 2.11.0+cu129 for torchao version 0.15.0             Please see https://github.com/pytorch/ao/issues/2919 for more info
+[transformers] You are using a model of type `vit_mae` to instantiate a model of type `vit`. This may be expected if you are loading a checkpoint that shares a subset of the architecture (e.g., loading a `sam2_video` checkpoint into `Sam2Model`), but is otherwise not supported and can yield errors. Please verify that the checkpoint is compatible with the model you are instantiating.
+
+[transformers] ViTModel LOAD REPORT from: PRAIG/musvit
+Key                                                                     | Status     | 
+------------------------------------------------------------------------+------------+-
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.weight  | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.weight | UNEXPECTED | 
+decoder.decoder_pred.weight                                             | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.bias            | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.weight          | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.bias    | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.bias            | UNEXPECTED | 
+decoder.decoder_norm.bias                                               | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.weight          | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.bias   | UNEXPECTED | 
+decoder.decoder_pred.bias                                               | UNEXPECTED | 
+decoder.decoder_norm.weight                                             | UNEXPECTED | 
+decoder.decoder_embed.weight                                            | UNEXPECTED | 
+decoder.decoder_pos_embed                                               | UNEXPECTED | 
+decoder.mask_token                                                      | UNEXPECTED | 
+decoder.decoder_embed.bias                                              | UNEXPECTED | 
+pooler.dense.bias                                                       | MISSING    | 
+pooler.dense.weight                                                     | MISSING    | 
+
+Notes:
+- UNEXPECTED:	can be ignored when loading from different task/architecture; not ok if you expect identical arch.
+- MISSING:	those params were newly initialized because missing from the checkpoint. Consider training on your downstream task.
+[eval] pages=32/67 mode=teacher_forced elapsed=9.3s
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_0.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.3116534799517232,
+      "errors": 2324,
+      "errors_per_1k_tokens": 31.188351338656645,
+      "errors_per_page_mean": 34.6865671641791,
+      "errors_per_page_median": 36.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.22475526351079522,
+      "errors": 1676,
+      "errors_per_1k_tokens": 22.49211568140643,
+      "errors_per_page_mean": 25.01492537313433,
+      "errors_per_page_median": 24.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.09762639131017836,
+      "errors": 728,
+      "errors_per_1k_tokens": 9.769844997651479,
+      "errors_per_page_mean": 10.865671641791044,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.04867909346922355,
+      "errors": 363,
+      "errors_per_1k_tokens": 4.871502382070724,
+      "errors_per_page_mean": 5.417910447761194,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.8507462686567164,
+      "pages_with_error": 57
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04787448035402977,
+      "errors": 357,
+      "errors_per_1k_tokens": 4.79098168154063,
+      "errors_per_page_mean": 5.3283582089552235,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.043851414778060885,
+      "errors": 327,
+      "errors_per_1k_tokens": 4.388378178890156,
+      "errors_per_page_mean": 4.880597014925373,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03312323990881051,
+      "errors": 247,
+      "errors_per_1k_tokens": 3.3147688384888947,
+      "errors_per_page_mean": 3.6865671641791047,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.7910447761194029,
+      "pages_with_error": 53
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.030307094005632293,
+      "errors": 226,
+      "errors_per_1k_tokens": 3.0329463866335638,
+      "errors_per_page_mean": 3.373134328358209,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.6865671641791045,
+      "pages_with_error": 46
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.024272495641678958,
+      "errors": 181,
+      "errors_per_1k_tokens": 2.4290411326578543,
+      "errors_per_page_mean": 2.701492537313433,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6417910447761194,
+      "pages_with_error": 43
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.021590451924366368,
+      "errors": 161,
+      "errors_per_1k_tokens": 2.1606387975575387,
+      "errors_per_page_mean": 2.4029850746268657,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5671641791044776,
+      "pages_with_error": 38
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.01622636448974118,
+      "errors": 121,
+      "errors_per_1k_tokens": 1.623834127356908,
+      "errors_per_page_mean": 1.8059701492537314,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6567164179104478,
+      "pages_with_error": 44
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.012337401099637924,
+      "errors": 92,
+      "errors_per_1k_tokens": 1.2346507414614507,
+      "errors_per_page_mean": 1.373134328358209,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5970149253731343,
+      "pages_with_error": 40
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.009789459568190962,
+      "errors": 73,
+      "errors_per_1k_tokens": 0.9796685231161512,
+      "errors_per_page_mean": 1.0895522388059702,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.44776119402985076,
+      "pages_with_error": 30
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.008448437709534665,
+      "errors": 63,
+      "errors_per_1k_tokens": 0.8454673555659934,
+      "errors_per_page_mean": 0.9402985074626866,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.008180233337803405,
+      "errors": 61,
+      "errors_per_1k_tokens": 0.8186271220559619,
+      "errors_per_page_mean": 0.9104477611940298,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.34328358208955223,
+      "pages_with_error": 23
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.0077779267802065176,
+      "errors": 58,
+      "errors_per_1k_tokens": 0.7783667717909146,
+      "errors_per_page_mean": 0.8656716417910447,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.31343283582089554,
+      "pages_with_error": 21
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.006705109293281481,
+      "errors": 50,
+      "errors_per_1k_tokens": 0.6710058377507885,
+      "errors_per_page_mean": 0.746268656716418,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.417910447761194,
+      "pages_with_error": 28
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.005632291806356444,
+      "errors": 42,
+      "errors_per_1k_tokens": 0.5636449037106623,
+      "errors_per_page_mean": 0.6268656716417911,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3283582089552239,
+      "pages_with_error": 22
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005364087434625185,
+      "errors": 40,
+      "errors_per_1k_tokens": 0.5368046702006307,
+      "errors_per_page_mean": 0.5970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.4925373134328358,
+      "pages_with_error": 33
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004961780877028295,
+      "errors": 37,
+      "errors_per_1k_tokens": 0.49654431993558346,
+      "errors_per_page_mean": 0.5522388059701493,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.0046935765052970365,
+      "errors": 35,
+      "errors_per_1k_tokens": 0.4697040864255519,
+      "errors_per_page_mean": 0.5223880597014925,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.19402985074626866,
+      "pages_with_error": 13
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0036207590183719995,
+      "errors": 27,
+      "errors_per_1k_tokens": 0.36234315238542575,
+      "errors_per_page_mean": 0.40298507462686567,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.29850746268656714,
+      "pages_with_error": 20
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0034866568325063697,
+      "errors": 26,
+      "errors_per_1k_tokens": 0.34892303563041,
+      "errors_per_page_mean": 0.3880597014925373,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0029502480890438512,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0029502480890438512,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.0025479415314469626,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.13432835820895522,
+      "pages_with_error": 9
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.0025479415314469626,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "clef",
+      "error_share": 0.0025479415314469626,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0014751240445219256,
+      "errors": 11,
+      "errors_per_1k_tokens": 0.14762128430517346,
+      "errors_per_page_mean": 0.16417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.08955223880597014,
+      "pages_with_error": 6
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0013410218586562962,
+      "errors": 10,
+      "errors_per_1k_tokens": 0.13420116755015768,
+      "errors_per_page_mean": 0.14925373134328357,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1044776119402985,
+      "pages_with_error": 7
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.0013410218586562962,
+      "errors": 10,
+      "errors_per_1k_tokens": 0.13420116755015768,
+      "errors_per_page_mean": 0.14925373134328357,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0006705109293281481,
+      "errors": 5,
+      "errors_per_1k_tokens": 0.06710058377507884,
+      "errors_per_page_mean": 0.07462686567164178,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0004023065575968888,
+      "errors": 3,
+      "errors_per_1k_tokens": 0.040260350265047304,
+      "errors_per_page_mean": 0.04477611940298507,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.04477611940298507,
+      "pages_with_error": 3
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.00013410218586562962,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "duration_and_accidental_and_tie_slur",
+      "error_share": 0.00013410218586562962,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 3,
+    "accidental_and_tie_slur": 50,
+    "accidental_only": 357,
+    "articulation_only": 35,
+    "barline": 27,
+    "chord_structure_and_accidental": 10,
+    "chord_structure_and_accidental_and_articulation": 11,
+    "chord_structure_and_accidental_and_tie_slur": 19,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 61,
+    "chord_structure_and_tie_slur": 92,
+    "chord_structure_and_tie_slur_and_articulation": 10,
+    "chord_structure_only": 42,
+    "clef": 19,
+    "column_or_line_break": 363,
+    "duration_and_accidental": 37,
+    "duration_and_accidental_and_tie_slur": 1,
+    "duration_and_pitch": 121,
+    "duration_and_pitch_and_accidental": 63,
+    "duration_and_pitch_and_accidental_and_tie_slur": 5,
+    "duration_and_pitch_and_tie_slur": 19,
+    "duration_and_tie_slur": 58,
+    "duration_only": 728,
+    "interpretation_or_control": 247,
+    "note_other": 2324,
+    "null_token": 73,
+    "pitch_and_accidental": 226,
+    "pitch_and_accidental_and_tie_slur": 26,
+    "pitch_and_tie_slur": 22,
+    "pitch_only": 327,
+    "rest_note_confusion": 161,
+    "spine_operation": 181,
+    "termination": 40,
+    "tie_slur_and_articulation": 22,
+    "tie_slur_only": 1676
+  },
+  "component_error_groups": {
+    "structural": 904,
+    "visual": 6553
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 12.985633342992514,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3211,
+    "articulation": 39,
+    "barline": 28,
+    "clef": 19,
+    "column_or_line_break": 364,
+    "duration": 999,
+    "interpretation_or_control": 149,
+    "note_other": 204,
+    "null_token": 74,
+    "pitch": 20,
+    "rest": 270,
+    "spine_operation": 216,
+    "termination": 31,
+    "tie_slur": 1833
+  },
+  "gold_error_groups": {
+    "structural": 834,
+    "visual": 6623
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.614958763122559,
+      "LER": 42.33919143676758,
+      "SER": 10.225929260253906,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.10007381064215258,
+      "top1_errors": 7457,
+      "top1_errors_per_page": 111.29850746268657
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 38,
+    "predicted_unk_literal": 77
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/per_page_records.jsonl

@@ -0,0 +1,67 @@
+{"dataset_index": 0, "sample_id": "polish_00002", "source_ids": [], "target_token_count": 1725, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 59, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 55, "field_index": 0, "field_text": "2.AA[", "gold_field": "2.AA[", "gold_line": "2.AA[\t4r\t12r", "gold_note_token": "2.AA[", "gold_piece": "[", "gold_token_id": 43, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 67, "line_char_start": 55, "line_index": 4, "line_text": "2.AA[\t4r\t12r", "operation": "replace", "pred_field": "2.AA(", "pred_local_replacement_line": "2.AA(\t4r\t12r", "pred_note_token": "2.AA(", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 14.091243561442237, "LER": 69.29824829101562, "SER": 10.382830626450117}, "nll": 0.44311535358428955, "top1_error_count": 179, "top1_error_rate": 0.10376811594202899}}
+{"dataset_index": 1, "sample_id": "polish_00004", "source_ids": [], "target_token_count": 989, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 35, "field_char_start": 30, "field_index": 0, "field_text": "8AA(L", "gold_field": "8AA(L", "gold_line": "8AA(L\t4.ee(", "gold_note_token": "8AA(L", "gold_piece": "8AA", "gold_token_id": 210, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 41, "line_char_start": 30, "line_index": 2, "line_text": "8AA(L\t4.ee(", "operation": "replace", "pred_field": "=!|:(L", "pred_local_replacement_line": "=!|:(L\t4.ee(", "pred_note_token": "=!|:(L", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 7.380952380952381, "LER": 35.22012710571289, "SER": 7.793522267206478}, "nll": 0.46089398860931396, "top1_error_count": 77, "top1_error_rate": 0.07785642062689585}}
+{"dataset_index": 2, "sample_id": "polish_00005", "source_ids": [], "target_token_count": 742, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t2cc(", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 55, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t2cc(", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t2cc(", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 6.037473976405274, "LER": 37.7952766418457, "SER": 9.041835357624832}, "nll": 0.5307933688163757, "top1_error_count": 67, "top1_error_rate": 0.09029649595687332}}
+{"dataset_index": 3, "sample_id": "polish_00007", "source_ids": [], "target_token_count": 888, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 85, "char_start": 81, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 87, "field_char_start": 81, "field_index": 1, "field_text": "16dd#L", "gold_field": "16dd#L", "gold_line": "8E 8A 8cJ\t16dd#L", "gold_note_token": "16dd#L", "gold_piece": "16dd", "gold_token_id": 112, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 87, "line_char_start": 71, "line_index": 6, "line_text": "8E 8A 8cJ\t16dd#L", "operation": "replace", "pred_field": "16dd#LL#L", "pred_local_replacement_line": "8E 8A 8cJ\t16dd#LL#L", "pred_note_token": "16dd#LL#L", "pred_piece": "16dd#LL", "pred_token_id": 1164, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 7.321772639691715, "LER": 32.66666793823242, "SER": 8.229988726042842}, "nll": 0.4590685963630676, "top1_error_count": 73, "top1_error_rate": 0.08220720720720721}}
+{"dataset_index": 4, "sample_id": "polish_00008", "source_ids": [], "target_token_count": 1236, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 48, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 48, "field_char_start": 45, "field_index": 1, "field_text": "8d[", "gold_field": "8d[", "gold_line": "8r\t8d[", "gold_note_token": "8d[", "gold_piece": "8d[", "gold_token_id": 1247, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 48, "line_char_start": 42, "line_index": 3, "line_text": "8r\t8d[", "operation": "replace", "pred_field": "8d(", "pred_local_replacement_line": "8r\t8d(", "pred_note_token": "8d(", "pred_piece": "8d(", "pred_token_id": 1822, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 9.24772400574988, "LER": 45.549739837646484, "SER": 9.7165991902834}, "nll": 0.5033981800079346, "top1_error_count": 120, "top1_error_rate": 0.0970873786407767}}
+{"dataset_index": 5, "sample_id": "polish_00009", "source_ids": [], "target_token_count": 942, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 111, "char_start": 110, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 110, "field_char_start": 107, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 52, "gold_trimmed_index": 52, "group": "structural", "line_char_end": 120, "line_char_start": 107, "line_index": 11, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 52, "visually_grounded": false}, "metrics": {"CER": 5.566218809980806, "LER": 21.604938507080078, "SER": 5.844845908607864}, "nll": 0.33280929923057556, "top1_error_count": 55, "top1_error_rate": 0.058386411889596604}}
+{"dataset_index": 6, "sample_id": "polish_00011", "source_ids": [], "target_token_count": 1093, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 7.742914979757084, "LER": 42.6315803527832, "SER": 9.24908424908425}, "nll": 0.47925251722335815, "top1_error_count": 101, "top1_error_rate": 0.09240622140896615}}
+{"dataset_index": 7, "sample_id": "polish_00013", "source_ids": [], "target_token_count": 1877, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 47, "char_start": 43, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 49, "field_char_start": 43, "field_index": 1, "field_text": "16.g(L", "gold_field": "16.g(L", "gold_line": "8CC 8GG 8CL\t16.g(L\t4en", "gold_note_token": "16.g(L", "gold_piece": "16.g", "gold_token_id": 1474, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 53, "line_char_start": 31, "line_index": 3, "line_text": "8CC 8GG 8CL\t16.g(L\t4en", "operation": "replace", "pred_field": "16.a(L", "pred_local_replacement_line": "8CC 8GG 8CL\t16.a(L\t4en", "pred_note_token": "16.a(L", "pred_piece": "16.a", "pred_token_id": 1313, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 7.45628505443748, "LER": 47.13656234741211, "SER": 7.995735607675907}, "nll": 0.4330143928527832, "top1_error_count": 150, "top1_error_rate": 0.07991475759190197}}
+{"dataset_index": 8, "sample_id": "polish_00014", "source_ids": [], "target_token_count": 1099, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t16bn(L", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 57, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t16bn(L", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t16bn(L", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 7.338991512730904, "LER": 30.735931396484375, "SER": 9.562841530054644}, "nll": 0.4915309548377991, "top1_error_count": 105, "top1_error_rate": 0.09554140127388536}}
+{"dataset_index": 9, "sample_id": "polish_00015", "source_ids": [], "target_token_count": 1086, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 38, "char_start": 37, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 38, "field_char_start": 37, "field_index": 1, "field_text": "*", "gold_field": "*", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "*^", "pred_local_replacement_line": "*^\t*^", "pred_note_token": null, "pred_piece": "*^", "pred_token_id": 191, "pred_trimmed_index": 10, "visually_grounded": false}, "metrics": {"CER": 8.019375672766415, "LER": 39.87730026245117, "SER": 9.493087557603687}, "nll": 0.48127126693725586, "top1_error_count": 103, "top1_error_rate": 0.09484346224677716}}
+{"dataset_index": 10, "sample_id": "polish_00017", "source_ids": [], "target_token_count": 1053, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 40, "char_start": 39, "component_category": "chord_structure_and_tie_slur", "component_flags": ["tie_slur", "chord_structure"], "component_group": "visual", "field_char_end": 44, "field_char_start": 34, "field_index": 0, "field_text": "4BBB- 4BB-", "gold_field": "4BBB- 4BB-", "gold_line": "4BBB- 4BB-\t4E 4cn", "gold_note_token": "4BBB-", "gold_piece": " ", "gold_token_id": 6, "gold_token_index": 9, "gold_trimmed_index": 9, "group": "visual", "line_char_end": 51, "line_char_start": 34, "line_index": 2, "line_text": "4BBB- 4BB-\t4E 4cn", "operation": "replace", "pred_field": "4BBB-(4BB-", "pred_local_replacement_line": "4BBB-(4BB-\t4E 4cn", "pred_note_token": "4BBB-(4BB-", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 9, "visually_grounded": true}, "metrics": {"CER": 2.638522427440633, "LER": 15.0, "SER": 4.847908745247148}, "nll": 0.27339133620262146, "top1_error_count": 51, "top1_error_rate": 0.04843304843304843}}
+{"dataset_index": 11, "sample_id": "polish_00019", "source_ids": [], "target_token_count": 1006, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 177, "char_start": 175, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 177, "field_char_start": 170, "field_index": 1, "field_text": "16eee)J", "gold_field": "16eee)J", "gold_line": ".\t16eee)J", "gold_note_token": "16eee)J", "gold_piece": ")J", "gold_token_id": 186, "gold_token_index": 90, "gold_trimmed_index": 90, "group": "visual", "line_char_end": 177, "line_char_start": 168, "line_index": 12, "line_text": ".\t16eee)J", "operation": "replace", "pred_field": "16eee)JJ", "pred_local_replacement_line": ".\t16eee)JJ", "pred_note_token": "16eee)JJ", "pred_piece": ")JJ", "pred_token_id": 305, "pred_trimmed_index": 90, "visually_grounded": true}, "metrics": {"CER": 5.331262939958592, "LER": 34.70588302612305, "SER": 8.059701492537313}, "nll": 0.505367636680603, "top1_error_count": 81, "top1_error_rate": 0.0805168986083499}}
+{"dataset_index": 12, "sample_id": "polish_00020", "source_ids": [], "target_token_count": 1183, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 5.6099732858414955, "LER": 35.88516616821289, "SER": 7.868020304568528}, "nll": 0.43884968757629395, "top1_error_count": 93, "top1_error_rate": 0.07861369399830938}}
+{"dataset_index": 13, "sample_id": "polish_00022", "source_ids": [], "target_token_count": 695, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 56, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 52, "field_index": 0, "field_text": "8G( 8B(L", "gold_field": "8G( 8B(L", "gold_line": "8G( 8B(L\t8ddd)L", "gold_note_token": "8B(L", "gold_piece": "8B(L", "gold_token_id": 2168, "gold_token_index": 20, "gold_trimmed_index": 20, "group": "visual", "line_char_end": 67, "line_char_start": 52, "line_index": 4, "line_text": "8G( 8B(L\t8ddd)L", "operation": "replace", "pred_field": "8G( 8BL", "pred_local_replacement_line": "8G( 8BL\t8ddd)L", "pred_note_token": "8BL", "pred_piece": "8BL", "pred_token_id": 396, "pred_trimmed_index": 20, "visually_grounded": true}, "metrics": {"CER": 8.978328173374612, "LER": 42.85714340209961, "SER": 10.951008645533141}, "nll": 0.6354540586471558, "top1_error_count": 76, "top1_error_rate": 0.1093525179856115}}
+{"dataset_index": 14, "sample_id": "polish_00023", "source_ids": [], "target_token_count": 1332, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 51, "field_index": 1, "field_text": "4gg)", "gold_field": "4gg)", "gold_line": "8GG 8GL\t4gg)\t8bnL", "gold_note_token": "4gg)", "gold_piece": "4gg)", "gold_token_id": 2154, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 60, "line_char_start": 43, "line_index": 3, "line_text": "8GG 8GL\t4gg)\t8bnL", "operation": "replace", "pred_field": "4gg", "pred_local_replacement_line": "8GG 8GL\t4gg\t8bnL", "pred_note_token": "4gg", "pred_piece": "4gg", "pred_token_id": 164, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 7.070240295748614, "LER": 46.34146499633789, "SER": 8.264462809917356}, "nll": 0.4163506031036377, "top1_error_count": 110, "top1_error_rate": 0.08258258258258258}}
+{"dataset_index": 15, "sample_id": "polish_00024", "source_ids": [], "target_token_count": 1075, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 8.181315644002211, "LER": 38.72832489013672, "SER": 9.404096834264433}, "nll": 0.47295889258384705, "top1_error_count": 101, "top1_error_rate": 0.09395348837209302}}
+{"dataset_index": 16, "sample_id": "polish_00025", "source_ids": [], "target_token_count": 879, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 34, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 45, "line_char_start": 34, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*M2/4", "pred_local_replacement_line": "*M2/4\t*M3/4", "pred_note_token": null, "pred_piece": "*M2/4", "pred_token_id": 304, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 7.85276073619632, "LER": 48.9510498046875, "SER": 11.275626423690206}, "nll": 0.536614716053009, "top1_error_count": 99, "top1_error_rate": 0.11262798634812286}}
+{"dataset_index": 17, "sample_id": "polish_00027", "source_ids": [], "target_token_count": 1825, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 46, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 46, "field_index": 0, "field_text": "8BBB- 8BB-", "gold_field": "8BBB- 8BB-", "gold_line": "8BBB- 8BB-\t8b- 8dd", "gold_note_token": "8BBB-", "gold_piece": "8BBB-", "gold_token_id": 632, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 3, "line_text": "8BBB- 8BB-\t8b- 8dd", "operation": "replace", "pred_field": "8AAA 8BB-", "pred_local_replacement_line": "8AAA 8BB-\t8b- 8dd", "pred_note_token": "8AAA", "pred_piece": "8AAA", "pred_token_id": 490, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 9.146519079522283, "LER": 51.937984466552734, "SER": 10.087719298245613}, "nll": 0.47878822684288025, "top1_error_count": 184, "top1_error_rate": 0.10082191780821918}}
+{"dataset_index": 18, "sample_id": "polish_00029", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 39, "field_char_start": 30, "field_index": 0, "field_text": "8CC( 8C(L", "gold_field": "8CC( 8C(L", "gold_line": "8CC( 8C(L\t4gg([", "gold_note_token": "8CC(", "gold_piece": "8CC", "gold_token_id": 364, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 45, "line_char_start": 30, "line_index": 2, "line_text": "8CC( 8C(L\t4gg([", "operation": "replace", "pred_field": "=!|:( 8C(L", "pred_local_replacement_line": "=!|:( 8C(L\t4gg([", "pred_note_token": "=!|:(", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 7.705479452054795, "LER": 35.319149017333984, "SER": 9.895414320193082}, "nll": 0.5735674500465393, "top1_error_count": 123, "top1_error_rate": 0.09887459807073955}}
+{"dataset_index": 19, "sample_id": "polish_00030", "source_ids": [], "target_token_count": 1263, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "=!|:", "pred_local_replacement_line": "=!|:\t*^", "pred_note_token": null, "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 11.927877947295423, "LER": 49.30232620239258, "SER": 9.508716323296355}, "nll": 0.43322452902793884, "top1_error_count": 120, "top1_error_rate": 0.09501187648456057}}
+{"dataset_index": 20, "sample_id": "polish_00031", "source_ids": [], "target_token_count": 862, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "null_token", "char_end": 47, "char_start": 46, "component_category": "null_token", "component_flags": [], "component_group": "structural", "field_char_end": 47, "field_char_start": 46, "field_index": 0, "field_text": ".", "gold_field": ".", "gold_line": ".\t16g#q(", "gold_note_token": null, "gold_piece": ".", "gold_token_id": 14, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 54, "line_char_start": 46, "line_index": 3, "line_text": ".\t16g#q(", "operation": "replace", "pred_field": "4.r", "pred_local_replacement_line": "4.r\t16g#q(", "pred_note_token": null, "pred_piece": "4.r", "pred_token_id": 426, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 11.219830397912588, "LER": 42.65734100341797, "SER": 10.569105691056912}, "nll": 0.5548887848854065, "top1_error_count": 91, "top1_error_rate": 0.10556844547563805}}
+{"dataset_index": 21, "sample_id": "polish_00032", "source_ids": [], "target_token_count": 1387, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 50, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 52, "field_char_start": 45, "field_index": 2, "field_text": "16ff#]L", "gold_field": "16ff#]L", "gold_line": "4G#(\t8G#L\t16ff#]L", "gold_note_token": "16ff#]L", "gold_piece": "16ff#", "gold_token_id": 180, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 52, "line_char_start": 35, "line_index": 3, "line_text": "4G#(\t8G#L\t16ff#]L", "operation": "replace", "pred_field": "16ff#]LL]L", "pred_local_replacement_line": "4G#(\t8G#L\t16ff#]LL]L", "pred_note_token": "16ff#]LL]L", "pred_piece": "16ff#]LL", "pred_token_id": 2839, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 11.064718162839249, "LER": 44.117645263671875, "SER": 11.688311688311687}, "nll": 0.659920871257782, "top1_error_count": 162, "top1_error_rate": 0.11679884643114635}}
+{"dataset_index": 22, "sample_id": "polish_00034", "source_ids": [], "target_token_count": 788, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 45, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8f( 8a(L", "gold_field": "8f( 8a(L", "gold_line": "8f( 8a(L\t16r", "gold_note_token": null, "gold_piece": "8f(", "gold_token_id": 2288, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 42, "line_index": 3, "line_text": "8f( 8a(L\t16r", "operation": "replace", "pred_field": "* 8a(L", "pred_local_replacement_line": "* 8a(L\t16r", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 24.271229404309253, "LER": 57.76397705078125, "SER": 39.00889453621347}, "nll": 1.3930959701538086, "top1_error_count": 180, "top1_error_rate": 0.22842639593908629}}
+{"dataset_index": 23, "sample_id": "polish_00036", "source_ids": [], "target_token_count": 1317, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 451, "char_start": 447, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 451, "field_char_start": 447, "field_index": 2, "field_text": "16gJ", "gold_field": "16gJ", "gold_line": ".\t.\t16gJ", "gold_note_token": "16gJ", "gold_piece": "16gJ", "gold_token_id": 859, "gold_token_index": 280, "gold_trimmed_index": 280, "group": "visual", "line_char_end": 451, "line_char_start": 443, "line_index": 41, "line_text": ".\t.\t16gJ", "operation": "replace", "pred_field": "16gJJ", "pred_local_replacement_line": ".\t.\t16gJJ", "pred_note_token": "16gJJ", "pred_piece": "16gJJ", "pred_token_id": 351, "pred_trimmed_index": 280, "visually_grounded": true}, "metrics": {"CER": 5.975554549569941, "LER": 23.4375, "SER": 4.86322188449848}, "nll": 0.29727262258529663, "top1_error_count": 64, "top1_error_rate": 0.04859529233105543}}
+{"dataset_index": 24, "sample_id": "polish_00037", "source_ids": [], "target_token_count": 871, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 49, "char_start": 46, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 46, "field_index": 0, "field_text": "4Fn 4dn", "gold_field": "4Fn 4dn", "gold_line": "4Fn 4dn\t2f#] 2ff#]", "gold_note_token": "4Fn", "gold_piece": "4Fn", "gold_token_id": 1918, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 2, "line_text": "4Fn 4dn\t2f#] 2ff#]", "operation": "replace", "pred_field": "4F# 4dn", "pred_local_replacement_line": "4F# 4dn\t2f#] 2ff#]", "pred_note_token": "4F#", "pred_piece": "4F#", "pred_token_id": 278, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 4.4473512099411385, "LER": 21.476511001586914, "SER": 5.747126436781609}, "nll": 0.2602337598800659, "top1_error_count": 50, "top1_error_rate": 0.0574052812858783}}
+{"dataset_index": 25, "sample_id": "polish_00038", "source_ids": [], "target_token_count": 1471, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 88, "char_start": 83, "component_category": "duration_and_tie_slur", "component_flags": ["duration", "tie_slur"], "component_group": "visual", "field_char_end": 88, "field_char_start": 83, "field_index": 1, "field_text": "16a(L", "gold_field": "16a(L", "gold_line": "8CC 8FF# 8CJ\t16a(L\t8e-", "gold_note_token": "16a(L", "gold_piece": "16a(L", "gold_token_id": 2253, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 92, "line_char_start": 70, "line_index": 5, "line_text": "8CC 8FF# 8CJ\t16a(L\t8e-", "operation": "replace", "pred_field": "32aLLL", "pred_local_replacement_line": "8CC 8FF# 8CJ\t32aLLL\t8e-", "pred_note_token": "32aLLL", "pred_piece": "32aLLL", "pred_token_id": 1732, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 6.594724220623502, "LER": 37.988826751708984, "SER": 7.346938775510205}, "nll": 0.41189873218536377, "top1_error_count": 108, "top1_error_rate": 0.07341944255608429}}
+{"dataset_index": 26, "sample_id": "polish_00039", "source_ids": [], "target_token_count": 1503, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 93, "char_start": 88, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 93, "field_char_start": 88, "field_index": 0, "field_text": "16A-L", "gold_field": "16A-L", "gold_line": "16A-L\t2.AA-[\t16cL\t8ee-", "gold_note_token": "16A-L", "gold_piece": "16A-L", "gold_token_id": 2834, "gold_token_index": 24, "gold_trimmed_index": 24, "group": "visual", "line_char_end": 110, "line_char_start": 88, "line_index": 4, "line_text": "16A-L\t2.AA-[\t16cL\t8ee-", "operation": "replace", "pred_field": "16A-LL", "pred_local_replacement_line": "16A-LL\t2.AA-[\t16cL\t8ee-", "pred_note_token": "16A-LL", "pred_piece": "16A-LL", "pred_token_id": 1109, "pred_trimmed_index": 24, "visually_grounded": true}, "metrics": {"CER": 8.835027365129006, "LER": 37.566139221191406, "SER": 8.721704394141145}, "nll": 0.49164462089538574, "top1_error_count": 131, "top1_error_rate": 0.08715901530272788}}
+{"dataset_index": 27, "sample_id": "polish_00040", "source_ids": [], "target_token_count": 1109, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 9.041501976284586, "LER": 48.29545593261719, "SER": 11.281588447653428}, "nll": 0.5216745734214783, "top1_error_count": 125, "top1_error_rate": 0.1127141568981064}}
+{"dataset_index": 28, "sample_id": "polish_00043", "source_ids": [], "target_token_count": 885, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "replace", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 13.676212741087085, "LER": 56.06936264038086, "SER": 17.081447963800905}, "nll": 0.8767596483230591, "top1_error_count": 151, "top1_error_rate": 0.17062146892655367}}
+{"dataset_index": 29, "sample_id": "polish_00044", "source_ids": [], "target_token_count": 677, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 74, "char_start": 73, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 82, "field_char_start": 70, "field_index": 1, "field_text": "16f( 16dd(Jk", "gold_field": "16f( 16dd(Jk", "gold_line": "16B-(Jk\t16f( 16dd(Jk", "gold_note_token": "16f(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 22, "gold_trimmed_index": 22, "group": "visual", "line_char_end": 82, "line_char_start": 62, "line_index": 4, "line_text": "16B-(Jk\t16f( 16dd(Jk", "operation": "delete", "pred_field": "16f  16dd(Jk", "pred_local_replacement_line": "16B-(Jk\t16f  16dd(Jk", "pred_note_token": "16f", "pred_piece": " ", "pred_token_id": 6, "pred_trimmed_index": 22, "visually_grounded": true}, "metrics": {"CER": 8.844065166795966, "LER": 46.28099060058594, "SER": 10.650887573964498}, "nll": 0.6048157215118408, "top1_error_count": 72, "top1_error_rate": 0.10635155096011817}}
+{"dataset_index": 30, "sample_id": "polish_00046", "source_ids": [], "target_token_count": 1069, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 57, "char_start": 54, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 57, "field_char_start": 54, "field_index": 0, "field_text": "4.F", "gold_field": "4.F", "gold_line": "4.F\t4.f", "gold_note_token": "4.F", "gold_piece": "4.F", "gold_token_id": 614, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": "4.F\t4.f", "operation": "replace", "pred_field": "4F", "pred_local_replacement_line": "4F\t4.f", "pred_note_token": "4F", "pred_piece": "4F", "pred_token_id": 153, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 9.354005167958658, "LER": 39.86928176879883, "SER": 9.269662921348315}, "nll": 0.3939111828804016, "top1_error_count": 99, "top1_error_rate": 0.09260991580916744}}
+{"dataset_index": 31, "sample_id": "polish_00047", "source_ids": [], "target_token_count": 1243, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 84, "char_start": 80, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 84, "field_char_start": 75, "field_index": 1, "field_text": "8ee) 8gg)", "gold_field": "8ee) 8gg)", "gold_line": "8CL\t8ee) 8gg)\t8cc", "gold_note_token": "8gg)", "gold_piece": "8gg)", "gold_token_id": 1085, "gold_token_index": 37, "gold_trimmed_index": 37, "group": "visual", "line_char_end": 88, "line_char_start": 71, "line_index": 5, "line_text": "8CL\t8ee) 8gg)\t8cc", "operation": "replace", "pred_field": "8ee) 8gg", "pred_local_replacement_line": "8CL\t8ee) 8gg\t8cc", "pred_note_token": "8gg", "pred_piece": "8gg", "pred_token_id": 87, "pred_trimmed_index": 37, "visually_grounded": true}, "metrics": {"CER": 5.278310940499041, "LER": 28.04878044128418, "SER": 5.877616747181964}, "nll": 0.312450647354126, "top1_error_count": 67, "top1_error_rate": 0.053901850362027354}}
+{"dataset_index": 32, "sample_id": "polish_00048", "source_ids": [], "target_token_count": 1384, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 51, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 48, "field_index": 0, "field_text": "4c 4f", "gold_field": "4c 4f", "gold_line": "4c 4f\t12aa-(L", "gold_note_token": "4f", "gold_piece": "4f", "gold_token_id": 103, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 48, "line_index": 3, "line_text": "4c 4f\t12aa-(L", "operation": "replace", "pred_field": "4c 4a-", "pred_local_replacement_line": "4c 4a-\t12aa-(L", "pred_note_token": "4a-", "pred_piece": "4a-", "pred_token_id": 341, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 12.807295796986518, "LER": 42.69662857055664, "SER": 9.255242227042661}, "nll": 0.5014135241508484, "top1_error_count": 128, "top1_error_rate": 0.09248554913294797}}
+{"dataset_index": 33, "sample_id": "polish_00049", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 13.113113113113112, "LER": 44.578311920166016, "SER": 12.389380530973451}, "nll": 0.5226848721504211, "top1_error_count": 132, "top1_error_rate": 0.10610932475884244}}
+{"dataset_index": 34, "sample_id": "polish_00050", "source_ids": [], "target_token_count": 1242, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "replace", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 12.215513237343242, "LER": 54.01069641113281, "SER": 11.522965350523771}, "nll": 0.5395593643188477, "top1_error_count": 143, "top1_error_rate": 0.11513687600644122}}
+{"dataset_index": 35, "sample_id": "polish_00052", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 59, "char_start": 58, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 59, "field_char_start": 47, "field_index": 0, "field_text": "2.DD-( 2.D-(", "gold_field": "2.DD-( 2.D-(", "gold_line": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "gold_note_token": "2.D-(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 74, "line_char_start": 47, "line_index": 3, "line_text": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "operation": "replace", "pred_field": "2.DD-( 2.D-", "pred_local_replacement_line": "2.DD-( 2.D-\t\t8ff(L\t2.f 2.a-", "pred_note_token": "2.D-", "pred_piece": "\t", "pred_token_id": 4, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 10.254957507082153, "LER": 45.35519027709961, "SER": 12.962962962962962}, "nll": 0.7004228234291077, "top1_error_count": 126, "top1_error_rate": 0.12949640287769784}}
+{"dataset_index": 36, "sample_id": "polish_00053", "source_ids": [], "target_token_count": 1018, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 48, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 62, "field_char_start": 48, "field_index": 0, "field_text": "4FFn: 4Cn: 4A:", "gold_field": "4FFn: 4Cn: 4A:", "gold_line": "4FFn: 4Cn: 4A:\t8fn", "gold_note_token": "4FFn:", "gold_piece": "4FF", "gold_token_id": 369, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 66, "line_char_start": 48, "line_index": 4, "line_text": "4FFn: 4Cn: 4A:\t8fn", "operation": "replace", "pred_field": "4FF#n: 4Cn: 4A:", "pred_local_replacement_line": "4FF#n: 4Cn: 4A:\t8fn", "pred_note_token": "4FF#n:", "pred_piece": "4FF#", "pred_token_id": 607, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 3.6838978015448602, "LER": 14.917126655578613, "SER": 3.933136676499508}, "nll": 0.23369836807250977, "top1_error_count": 40, "top1_error_rate": 0.03929273084479371}}
+{"dataset_index": 37, "sample_id": "polish_00054", "source_ids": [], "target_token_count": 689, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 46, "field_index": 0, "field_text": "8B-( 8d(L", "gold_field": "8B-( 8d(L", "gold_line": "8B-( 8d(L\t8fL", "gold_note_token": "8d(L", "gold_piece": "8d(L", "gold_token_id": 1103, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 59, "line_char_start": 46, "line_index": 3, "line_text": "8B-( 8d(L\t8fL", "operation": "replace", "pred_field": "8B-( 8dL", "pred_local_replacement_line": "8B-( 8dL\t8fL", "pred_note_token": "8dL", "pred_piece": "8dL", "pred_token_id": 206, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 9.509202453987731, "LER": 41.17647171020508, "SER": 11.627906976744185}, "nll": 0.715951681137085, "top1_error_count": 80, "top1_error_rate": 0.11611030478955008}}
+{"dataset_index": 38, "sample_id": "polish_00056", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 28, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 28, "field_char_start": 16, "field_index": 0, "field_text": "*k[b-e-a-d-]", "gold_field": "*k[b-e-a-d-]", "gold_line": "*k[b-e-a-d-]\t*k[b-e-a-d-]", "gold_note_token": null, "gold_piece": "*k[b-e-a-d-]", "gold_token_id": 637, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 41, "line_char_start": 16, "line_index": 1, "line_text": "*k[b-e-a-d-]\t*k[b-e-a-d-]", "operation": "replace", "pred_field": "*k[b-e-a-d-g-c-]", "pred_local_replacement_line": "*k[b-e-a-d-g-c-]\t*k[b-e-a-d-]", "pred_note_token": null, "pred_piece": "*k[b-e-a-d-g-c-]", "pred_token_id": 827, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 14.093959731543624, "LER": 47.004608154296875, "SER": 14.323843416370108}, "nll": 0.579949676990509, "top1_error_count": 161, "top1_error_rate": 0.1431111111111111}}
+{"dataset_index": 39, "sample_id": "polish_00057", "source_ids": [], "target_token_count": 733, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 52, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 57, "field_char_start": 52, "field_index": 1, "field_text": "8ff#q", "gold_field": "8ff#q", "gold_line": ".\t8ff#q", "gold_note_token": "8ff#q", "gold_piece": "8ff#", "gold_token_id": 92, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": ".\t8ff#q", "operation": "replace", "pred_field": "8ee-q", "pred_local_replacement_line": ".\t8ee-q", "pred_note_token": "8ee-q", "pred_piece": "8ee-", "pred_token_id": 330, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 12.297297297297298, "LER": 46.853145599365234, "SER": 14.89071038251366}, "nll": 0.867639422416687, "top1_error_count": 109, "top1_error_rate": 0.14870395634379263}}
+{"dataset_index": 40, "sample_id": "polish_00058", "source_ids": [], "target_token_count": 900, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 68, "char_start": 64, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 73, "field_char_start": 64, "field_index": 0, "field_text": "8.FF 8.FL", "gold_field": "8.FF 8.FL", "gold_line": "8.FF 8.FL\t8.fL", "gold_note_token": "8.FF", "gold_piece": "8.FF", "gold_token_id": 1569, "gold_token_index": 20, "gold_trimmed_index": 20, "group": "visual", "line_char_end": 78, "line_char_start": 64, "line_index": 5, "line_text": "8.FF 8.FL\t8.fL", "operation": "replace", "pred_field": "8FF 8.FL", "pred_local_replacement_line": "8FF 8.FL\t8.fL", "pred_note_token": "8FF", "pred_piece": "8FF", "pred_token_id": 276, "pred_trimmed_index": 20, "visually_grounded": true}, "metrics": {"CER": 6.869300911854104, "LER": 33.57143020629883, "SER": 7.786429365962181}, "nll": 0.4733186662197113, "top1_error_count": 70, "top1_error_rate": 0.07777777777777778}}
+{"dataset_index": 41, "sample_id": "polish_00059", "source_ids": [], "target_token_count": 1838, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 48, "char_start": 46, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 48, "field_char_start": 46, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 50, "line_char_start": 46, "line_index": 3, "line_text": "*^\t*", "operation": "replace", "pred_field": "4.AA", "pred_local_replacement_line": "4.AA\t*", "pred_note_token": null, "pred_piece": "4.AA", "pred_token_id": 704, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 13.267445342917043, "LER": 66.40926361083984, "SER": 13.500272182906913}, "nll": 0.6855016350746155, "top1_error_count": 248, "top1_error_rate": 0.13492927094668117}}
+{"dataset_index": 42, "sample_id": "polish_00060", "source_ids": [], "target_token_count": 1528, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 62, "char_start": 59, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 66, "field_char_start": 59, "field_index": 0, "field_text": "8FF 8FL", "gold_field": "8FF 8FL", "gold_line": "8FF 8FL\t2.F\t4.f", "gold_note_token": "8FF", "gold_piece": "8FF", "gold_token_id": 276, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 74, "line_char_start": 59, "line_index": 4, "line_text": "8FF 8FL\t2.F\t4.f", "operation": "replace", "pred_field": "8FL 8FL", "pred_local_replacement_line": "8FL 8FL\t2.F\t4.f", "pred_note_token": "8FL", "pred_piece": "8FL", "pred_token_id": 433, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 6.673881673881674, "LER": 45.32710266113281, "SER": 8.775376555337262}, "nll": 0.4827817678451538, "top1_error_count": 134, "top1_error_rate": 0.08769633507853403}}
+{"dataset_index": 43, "sample_id": "polish_00061", "source_ids": [], "target_token_count": 653, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 50, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 59, "field_char_start": 50, "field_index": 0, "field_text": "8e-( 8g(L", "gold_field": "8e-( 8g(L", "gold_line": "8e-( 8g(L\t8gg 8bb-L", "gold_note_token": null, "gold_piece": "8e-", "gold_token_id": 263, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 69, "line_char_start": 50, "line_index": 3, "line_text": "8e-( 8g(L\t8gg 8bb-L", "operation": "replace", "pred_field": "*( 8g(L", "pred_local_replacement_line": "*( 8g(L\t8gg 8bb-L", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 11.607142857142858, "LER": 47.95918273925781, "SER": 15.797546012269938}, "nll": 0.7420379519462585, "top1_error_count": 103, "top1_error_rate": 0.15773353751914243}}
+{"dataset_index": 44, "sample_id": "polish_00062", "source_ids": [], "target_token_count": 781, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 74, "char_start": 71, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 74, "field_char_start": 68, "field_index": 0, "field_text": "8G 8eJ", "gold_field": "8G 8eJ", "gold_line": "8G 8eJ\t.\t.", "gold_note_token": "8eJ", "gold_piece": "8eJ", "gold_token_id": 185, "gold_token_index": 26, "gold_trimmed_index": 26, "group": "visual", "line_char_end": 78, "line_char_start": 68, "line_index": 4, "line_text": "8G 8eJ\t.\t.", "operation": "replace", "pred_field": "8G 8cJ", "pred_local_replacement_line": "8G 8cJ\t.\t.", "pred_note_token": "8cJ", "pred_piece": "8cJ", "pred_token_id": 306, "pred_trimmed_index": 26, "visually_grounded": true}, "metrics": {"CER": 11.76470588235294, "LER": 40.18691635131836, "SER": 9.743589743589745}, "nll": 0.4642954170703888, "top1_error_count": 76, "top1_error_rate": 0.0973111395646607}}
+{"dataset_index": 45, "sample_id": "polish_00063", "source_ids": [], "target_token_count": 963, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 54, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 48, "field_index": 1, "field_text": "16qqddnL", "gold_field": "16qqddnL", "gold_line": ".\t16qqddnL", "gold_note_token": "16qqddnL", "gold_piece": "nL", "gold_token_id": 403, "gold_token_index": 11, "gold_trimmed_index": 11, "group": "visual", "line_char_end": 56, "line_char_start": 46, "line_index": 2, "line_text": ".\t16qqddnL", "operation": "replace", "pred_field": "16qqdd#LL", "pred_local_replacement_line": ".\t16qqdd#LL", "pred_note_token": "16qqdd#LL", "pred_piece": "#LL", "pred_token_id": 269, "pred_trimmed_index": 11, "visually_grounded": true}, "metrics": {"CER": 4.665438919582566, "LER": 30.263158798217773, "SER": 6.652806652806653}, "nll": 0.34757325053215027, "top1_error_count": 64, "top1_error_rate": 0.06645898234683281}}
+{"dataset_index": 46, "sample_id": "polish_00064", "source_ids": [], "target_token_count": 928, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 53, "char_start": 49, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 53, "field_char_start": 49, "field_index": 1, "field_text": "8dd[", "gold_field": "8dd[", "gold_line": "8r\t8dd[", "gold_note_token": "8dd[", "gold_piece": "8dd[", "gold_token_id": 1303, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 53, "line_char_start": 46, "line_index": 3, "line_text": "8r\t8dd[", "operation": "replace", "pred_field": "8dd(", "pred_local_replacement_line": "8r\t8dd(", "pred_note_token": "8dd(", "pred_piece": "8dd(", "pred_token_id": 1124, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 10.005717552887363, "LER": 42.9378547668457, "SER": 12.944983818770226}, "nll": 0.730968713760376, "top1_error_count": 120, "top1_error_rate": 0.12931034482758622}}
+{"dataset_index": 47, "sample_id": "polish_00067", "source_ids": [], "target_token_count": 1138, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 43, "char_start": 39, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 43, "field_char_start": 39, "field_index": 1, "field_text": "16aL", "gold_field": "16aL", "gold_line": "8F(L\t16aL", "gold_note_token": "16aL", "gold_piece": "16aL", "gold_token_id": 602, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 43, "line_char_start": 34, "line_index": 2, "line_text": "8F(L\t16aL", "operation": "replace", "pred_field": "16aLL", "pred_local_replacement_line": "8F(L\t16aLL", "pred_note_token": "16aLL", "pred_piece": "16aLL", "pred_token_id": 338, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 11.73974540311174, "LER": 44.79166793823242, "SER": 14.775725593667547}, "nll": 0.8811405301094055, "top1_error_count": 167, "top1_error_rate": 0.14674868189806678}}
+{"dataset_index": 48, "sample_id": "polish_00068", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 68, "char_start": 66, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 68, "field_char_start": 62, "field_index": 0, "field_text": "16GG(L", "gold_field": "16GG(L", "gold_line": "16GG(L\t8..bb(L", "gold_note_token": "16GG(L", "gold_piece": "(L", "gold_token_id": 171, "gold_token_index": 21, "gold_trimmed_index": 21, "group": "visual", "line_char_end": 76, "line_char_start": 62, "line_index": 5, "line_text": "16GG(L\t8..bb(L", "operation": "replace", "pred_field": "16GG(LL", "pred_local_replacement_line": "16GG(LL\t8..bb(L", "pred_note_token": "16GG(LL", "pred_piece": "(LL", "pred_token_id": 279, "pred_trimmed_index": 21, "visually_grounded": true}, "metrics": {"CER": 23.614609571788414, "LER": 64.08839416503906, "SER": 20.679012345679013}, "nll": 0.9908871650695801, "top1_error_count": 201, "top1_error_rate": 0.2065775950668037}}
+{"dataset_index": 49, "sample_id": "polish_00071", "source_ids": [], "target_token_count": 1433, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 48, "char_start": 46, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 48, "field_char_start": 46, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*^", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 51, "line_char_start": 46, "line_index": 2, "line_text": "*^\t*^", "operation": "replace", "pred_field": ".", "pred_local_replacement_line": ".\t*^", "pred_note_token": null, "pred_piece": ".", "pred_token_id": 14, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 12.215669755686605, "LER": 65.0273208618164, "SER": 14.315642458100559}, "nll": 0.725530207157135, "top1_error_count": 205, "top1_error_rate": 0.14305652477320308}}
+{"dataset_index": 50, "sample_id": "polish_00072", "source_ids": [], "target_token_count": 1123, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 56, "char_start": 55, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 13, "gold_trimmed_index": 13, "group": "structural", "line_char_end": 65, "line_char_start": 52, "line_index": 3, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 13, "visually_grounded": false}, "metrics": {"CER": 4.236006051437216, "LER": 25.53191566467285, "SER": 5.793226381461675}, "nll": 0.27293404936790466, "top1_error_count": 65, "top1_error_rate": 0.0578806767586821}}
+{"dataset_index": 51, "sample_id": "polish_00074", "source_ids": [], "target_token_count": 1355, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 34, "char_start": 30, "component_category": "duration_and_pitch_and_accidental", "component_flags": ["duration", "pitch", "accidental"], "component_group": "visual", "field_char_end": 36, "field_char_start": 30, "field_index": 0, "field_text": "8BBnq(", "gold_field": "8BBnq(", "gold_line": "8BBnq(\t.", "gold_note_token": "8BBnq(", "gold_piece": "8BBn", "gold_token_id": 2042, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 38, "line_char_start": 30, "line_index": 2, "line_text": "8BBnq(\t.", "operation": "replace", "pred_field": "=!|:q(", "pred_local_replacement_line": "=!|:q(\t.", "pred_note_token": "=!|:q(", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 9.098979138925877, "LER": 43.589744567871094, "SER": 9.67503692762186}, "nll": 0.45965346693992615, "top1_error_count": 131, "top1_error_rate": 0.09667896678966789}}
+{"dataset_index": 52, "sample_id": "polish_00078", "source_ids": [], "target_token_count": 1172, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 49, "char_start": 48, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 49, "field_char_start": 48, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*\t*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 56, "line_char_start": 48, "line_index": 3, "line_text": "*\t*\t*^\t*", "operation": "insert", "pred_field": "4Dn", "pred_local_replacement_line": "4Dn\t*\t*^\t*", "pred_note_token": null, "pred_piece": "4Dn", "pred_token_id": 2294, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 16.586409844836812, "LER": 46.75324630737305, "SER": 13.321947053800171}, "nll": 0.5503836870193481, "top1_error_count": 156, "top1_error_rate": 0.13310580204778158}}
+{"dataset_index": 53, "sample_id": "polish_00082", "source_ids": [], "target_token_count": 1027, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "8AA-L", "pred_local_replacement_line": "8AA-L\t*^", "pred_note_token": null, "pred_piece": "8AA-L", "pred_token_id": 901, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 15.060908084163898, "LER": 51.31578826904297, "SER": 10.818713450292398}, "nll": 0.5742161273956299, "top1_error_count": 111, "top1_error_rate": 0.10808179162609542}}
+{"dataset_index": 54, "sample_id": "polish_00084", "source_ids": [], "target_token_count": 860, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 54, "char_start": 50, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 57, "field_char_start": 50, "field_index": 0, "field_text": "32qqCCL", "gold_field": "32qqCCL", "gold_line": "32qqCCL\t32qqe-L", "gold_note_token": "32qqCCL", "gold_piece": "32qq", "gold_token_id": 1235, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 65, "line_char_start": 50, "line_index": 3, "line_text": "32qqCCL\t32qqe-L", "operation": "replace", "pred_field": "32DDCCL", "pred_local_replacement_line": "32DDCCL\t32qqe-L", "pred_note_token": "32DDCCL", "pred_piece": "32DD", "pred_token_id": 1594, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 11.28048780487805, "LER": 45.222930908203125, "SER": 13.62048894062864}, "nll": 0.7655315399169922, "top1_error_count": 117, "top1_error_rate": 0.13604651162790699}}
+{"dataset_index": 55, "sample_id": "polish_00085", "source_ids": [], "target_token_count": 1428, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 54, "char_start": 49, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 54, "field_char_start": 49, "field_index": 2, "field_text": "8aa(L", "gold_field": "8aa(L", "gold_line": "4d\t4D 4A\t8aa(L\t4f# 4dd", "gold_note_token": "8aa(L", "gold_piece": "8aa(L", "gold_token_id": 1057, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 40, "line_index": 3, "line_text": "4d\t4D 4A\t8aa(L\t4f# 4dd", "operation": "replace", "pred_field": "16aa(LL", "pred_local_replacement_line": "4d\t4D 4A\t16aa(LL\t4f# 4dd", "pred_note_token": "16aa(LL", "pred_piece": "16aa(LL", "pred_token_id": 1433, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 4.270152505446624, "LER": 22.360248565673828, "SER": 3.9943938332165385}, "nll": 0.21630828082561493, "top1_error_count": 57, "top1_error_rate": 0.03991596638655462}}
+{"dataset_index": 56, "sample_id": "polish_00086", "source_ids": [], "target_token_count": 1170, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 44, "field_char_start": 42, "field_index": 0, "field_text": "4C", "gold_field": "4C", "gold_line": "4C\t4r", "gold_note_token": null, "gold_piece": "4C", "gold_token_id": 181, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 47, "line_char_start": 42, "line_index": 2, "line_text": "4C\t4r", "operation": "replace", "pred_field": "*M4/4", "pred_local_replacement_line": "*M4/4\t4r", "pred_note_token": null, "pred_piece": "*M4/4", "pred_token_id": 229, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 8.396946564885496, "LER": 38.414634704589844, "SER": 8.126603934987168}, "nll": 0.39756131172180176, "top1_error_count": 95, "top1_error_rate": 0.0811965811965812}}
+{"dataset_index": 57, "sample_id": "polish_00087", "source_ids": [], "target_token_count": 1300, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 47, "char_start": 44, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 56, "field_char_start": 44, "field_index": 1, "field_text": "8ccq( 8ee-q(", "gold_field": "8ccq( 8ee-q(", "gold_line": ".\t8ccq( 8ee-q(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 56, "line_char_start": 42, "line_index": 2, "line_text": ".\t8ccq( 8ee-q(", "operation": "replace", "pred_field": "8ccqq( 8ee-q(", "pred_local_replacement_line": ".\t8ccqq( 8ee-q(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 14.012184508268058, "LER": 56.5217399597168, "SER": 12.317167051578137}, "nll": 0.5792496204376221, "top1_error_count": 160, "top1_error_rate": 0.12307692307692308}}
+{"dataset_index": 58, "sample_id": "polish_00089", "source_ids": [], "target_token_count": 782, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 37, "char_start": 35, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 43, "field_char_start": 35, "field_index": 0, "field_text": "8A( 8c(L", "gold_field": "8A( 8c(L", "gold_line": "8A( 8c(L\t2.D\t16g#(L", "gold_note_token": "8A(", "gold_piece": "8A", "gold_token_id": 109, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 35, "line_index": 3, "line_text": "8A( 8c(L\t2.D\t16g#(L", "operation": "replace", "pred_field": "8A(L( 8c(L", "pred_local_replacement_line": "8A(L( 8c(L\t2.D\t16g#(L", "pred_note_token": "8A(L(", "pred_piece": "8A(L", "pred_token_id": 1848, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 12.879298718813217, "LER": 49.068321228027344, "SER": 15.236875800256081}, "nll": 0.9574787020683289, "top1_error_count": 119, "top1_error_rate": 0.15217391304347827}}
+{"dataset_index": 59, "sample_id": "polish_00090", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 8.609271523178808, "LER": 35.57046890258789, "SER": 7.473309608540925}, "nll": 0.38987988233566284, "top1_error_count": 84, "top1_error_rate": 0.07466666666666667}}
+{"dataset_index": 60, "sample_id": "polish_00092", "source_ids": [], "target_token_count": 1153, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 62, "char_start": 61, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 62, "field_char_start": 52, "field_index": 1, "field_text": "4d-( 4dd-(", "gold_field": "4d-( 4dd-(", "gold_line": "8BB-\t4d-( 4dd-(", "gold_note_token": "4dd-(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 47, "line_index": 3, "line_text": "8BB-\t4d-( 4dd-(", "operation": "replace", "pred_field": "4d-( 4dd-\n", "pred_local_replacement_line": "8BB-\t4d-( 4dd-\n", "pred_note_token": "4dd-\n", "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 5.90513068731849, "LER": 36.87150955200195, "SER": 7.552083333333333}, "nll": 0.5161488652229309, "top1_error_count": 87, "top1_error_rate": 0.07545533391153512}}
+{"dataset_index": 61, "sample_id": "polish_00093", "source_ids": [], "target_token_count": 1047, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 8.693081017149616, "LER": 32.84671401977539, "SER": 7.45697896749522}, "nll": 0.33568888902664185, "top1_error_count": 78, "top1_error_rate": 0.07449856733524356}}
+{"dataset_index": 62, "sample_id": "polish_00094", "source_ids": [], "target_token_count": 773, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 47, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 47, "field_char_start": 40, "field_index": 1, "field_text": "32qqggL", "gold_field": "32qqggL", "gold_line": ".\t32qqggL", "gold_note_token": "32qqggL", "gold_piece": "L", "gold_token_id": 38, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 47, "line_char_start": 38, "line_index": 2, "line_text": ".\t32qqggL", "operation": "replace", "pred_field": "32qqgg(LLL", "pred_local_replacement_line": ".\t32qqgg(LLL", "pred_note_token": "32qqgg(LLL", "pred_piece": "(LLL", "pred_token_id": 1633, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 9.45945945945946, "LER": 43.870967864990234, "SER": 11.658031088082902}, "nll": 0.7603899836540222, "top1_error_count": 90, "top1_error_rate": 0.11642949547218628}}
+{"dataset_index": 63, "sample_id": "polish_00095", "source_ids": [], "target_token_count": 1073, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "4D-", "pred_local_replacement_line": "4D-\t*^", "pred_note_token": null, "pred_piece": "4D-", "pred_token_id": 556, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 16.378859236002093, "LER": 45.87628936767578, "SER": 12.87313432835821}, "nll": 0.6102403402328491, "top1_error_count": 138, "top1_error_rate": 0.12861136999068035}}
+{"dataset_index": 64, "sample_id": "polish_00096", "source_ids": [], "target_token_count": 935, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 52, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "2en", "gold_field": "2en", "gold_line": "2en\t4F#]\t4ggn\t2cc#", "gold_note_token": "2en", "gold_piece": "2en", "gold_token_id": 2374, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 70, "line_char_start": 52, "line_index": 3, "line_text": "2en\t4F#]\t4ggn\t2cc#", "operation": "replace", "pred_field": "2e", "pred_local_replacement_line": "2e\t4F#]\t4ggn\t2cc#", "pred_note_token": "2e", "pred_piece": "2e", "pred_token_id": 296, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 5.798969072164948, "LER": 40.71428680419922, "SER": 6.745182012847965}, "nll": 0.3402958810329437, "top1_error_count": 63, "top1_error_rate": 0.06737967914438503}}
+{"dataset_index": 65, "sample_id": "polish_00097", "source_ids": [], "target_token_count": 1050, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 7.276007215874925, "LER": 32.835819244384766, "SER": 6.48236415633937}, "nll": 0.3443845510482788, "top1_error_count": 68, "top1_error_rate": 0.06476190476190476}}
+{"dataset_index": 66, "sample_id": "polish_00100", "source_ids": [], "target_token_count": 1185, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 36, "char_start": 34, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 36, "field_char_start": 34, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "16BBB", "pred_local_replacement_line": "16BBB\t*", "pred_note_token": null, "pred_piece": "16BBB", "pred_token_id": 944, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 13.732576148683531, "LER": 53.439151763916016, "SER": 13.766891891891891}, "nll": 0.6598047018051147, "top1_error_count": 163, "top1_error_rate": 0.13755274261603376}}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/summary.json

@@ -0,0 +1,465 @@
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_0.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.3116534799517232,
+      "errors": 2324,
+      "errors_per_1k_tokens": 31.188351338656645,
+      "errors_per_page_mean": 34.6865671641791,
+      "errors_per_page_median": 36.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.22475526351079522,
+      "errors": 1676,
+      "errors_per_1k_tokens": 22.49211568140643,
+      "errors_per_page_mean": 25.01492537313433,
+      "errors_per_page_median": 24.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.09762639131017836,
+      "errors": 728,
+      "errors_per_1k_tokens": 9.769844997651479,
+      "errors_per_page_mean": 10.865671641791044,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.04867909346922355,
+      "errors": 363,
+      "errors_per_1k_tokens": 4.871502382070724,
+      "errors_per_page_mean": 5.417910447761194,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.8507462686567164,
+      "pages_with_error": 57
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04787448035402977,
+      "errors": 357,
+      "errors_per_1k_tokens": 4.79098168154063,
+      "errors_per_page_mean": 5.3283582089552235,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.043851414778060885,
+      "errors": 327,
+      "errors_per_1k_tokens": 4.388378178890156,
+      "errors_per_page_mean": 4.880597014925373,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03312323990881051,
+      "errors": 247,
+      "errors_per_1k_tokens": 3.3147688384888947,
+      "errors_per_page_mean": 3.6865671641791047,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.7910447761194029,
+      "pages_with_error": 53
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.030307094005632293,
+      "errors": 226,
+      "errors_per_1k_tokens": 3.0329463866335638,
+      "errors_per_page_mean": 3.373134328358209,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.6865671641791045,
+      "pages_with_error": 46
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.024272495641678958,
+      "errors": 181,
+      "errors_per_1k_tokens": 2.4290411326578543,
+      "errors_per_page_mean": 2.701492537313433,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6417910447761194,
+      "pages_with_error": 43
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.021590451924366368,
+      "errors": 161,
+      "errors_per_1k_tokens": 2.1606387975575387,
+      "errors_per_page_mean": 2.4029850746268657,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5671641791044776,
+      "pages_with_error": 38
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.01622636448974118,
+      "errors": 121,
+      "errors_per_1k_tokens": 1.623834127356908,
+      "errors_per_page_mean": 1.8059701492537314,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6567164179104478,
+      "pages_with_error": 44
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.012337401099637924,
+      "errors": 92,
+      "errors_per_1k_tokens": 1.2346507414614507,
+      "errors_per_page_mean": 1.373134328358209,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5970149253731343,
+      "pages_with_error": 40
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.009789459568190962,
+      "errors": 73,
+      "errors_per_1k_tokens": 0.9796685231161512,
+      "errors_per_page_mean": 1.0895522388059702,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.44776119402985076,
+      "pages_with_error": 30
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.008448437709534665,
+      "errors": 63,
+      "errors_per_1k_tokens": 0.8454673555659934,
+      "errors_per_page_mean": 0.9402985074626866,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.008180233337803405,
+      "errors": 61,
+      "errors_per_1k_tokens": 0.8186271220559619,
+      "errors_per_page_mean": 0.9104477611940298,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.34328358208955223,
+      "pages_with_error": 23
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.0077779267802065176,
+      "errors": 58,
+      "errors_per_1k_tokens": 0.7783667717909146,
+      "errors_per_page_mean": 0.8656716417910447,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.31343283582089554,
+      "pages_with_error": 21
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.006705109293281481,
+      "errors": 50,
+      "errors_per_1k_tokens": 0.6710058377507885,
+      "errors_per_page_mean": 0.746268656716418,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.417910447761194,
+      "pages_with_error": 28
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.005632291806356444,
+      "errors": 42,
+      "errors_per_1k_tokens": 0.5636449037106623,
+      "errors_per_page_mean": 0.6268656716417911,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3283582089552239,
+      "pages_with_error": 22
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005364087434625185,
+      "errors": 40,
+      "errors_per_1k_tokens": 0.5368046702006307,
+      "errors_per_page_mean": 0.5970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.4925373134328358,
+      "pages_with_error": 33
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004961780877028295,
+      "errors": 37,
+      "errors_per_1k_tokens": 0.49654431993558346,
+      "errors_per_page_mean": 0.5522388059701493,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.0046935765052970365,
+      "errors": 35,
+      "errors_per_1k_tokens": 0.4697040864255519,
+      "errors_per_page_mean": 0.5223880597014925,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.19402985074626866,
+      "pages_with_error": 13
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0036207590183719995,
+      "errors": 27,
+      "errors_per_1k_tokens": 0.36234315238542575,
+      "errors_per_page_mean": 0.40298507462686567,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.29850746268656714,
+      "pages_with_error": 20
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0034866568325063697,
+      "errors": 26,
+      "errors_per_1k_tokens": 0.34892303563041,
+      "errors_per_page_mean": 0.3880597014925373,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0029502480890438512,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0029502480890438512,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.0025479415314469626,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.13432835820895522,
+      "pages_with_error": 9
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.0025479415314469626,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "clef",
+      "error_share": 0.0025479415314469626,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0014751240445219256,
+      "errors": 11,
+      "errors_per_1k_tokens": 0.14762128430517346,
+      "errors_per_page_mean": 0.16417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.08955223880597014,
+      "pages_with_error": 6
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0013410218586562962,
+      "errors": 10,
+      "errors_per_1k_tokens": 0.13420116755015768,
+      "errors_per_page_mean": 0.14925373134328357,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1044776119402985,
+      "pages_with_error": 7
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.0013410218586562962,
+      "errors": 10,
+      "errors_per_1k_tokens": 0.13420116755015768,
+      "errors_per_page_mean": 0.14925373134328357,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0006705109293281481,
+      "errors": 5,
+      "errors_per_1k_tokens": 0.06710058377507884,
+      "errors_per_page_mean": 0.07462686567164178,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0004023065575968888,
+      "errors": 3,
+      "errors_per_1k_tokens": 0.040260350265047304,
+      "errors_per_page_mean": 0.04477611940298507,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.04477611940298507,
+      "pages_with_error": 3
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.00013410218586562962,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    },
+    {
+      "category": "duration_and_accidental_and_tie_slur",
+      "error_share": 0.00013410218586562962,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 3,
+    "accidental_and_tie_slur": 50,
+    "accidental_only": 357,
+    "articulation_only": 35,
+    "barline": 27,
+    "chord_structure_and_accidental": 10,
+    "chord_structure_and_accidental_and_articulation": 11,
+    "chord_structure_and_accidental_and_tie_slur": 19,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 61,
+    "chord_structure_and_tie_slur": 92,
+    "chord_structure_and_tie_slur_and_articulation": 10,
+    "chord_structure_only": 42,
+    "clef": 19,
+    "column_or_line_break": 363,
+    "duration_and_accidental": 37,
+    "duration_and_accidental_and_tie_slur": 1,
+    "duration_and_pitch": 121,
+    "duration_and_pitch_and_accidental": 63,
+    "duration_and_pitch_and_accidental_and_tie_slur": 5,
+    "duration_and_pitch_and_tie_slur": 19,
+    "duration_and_tie_slur": 58,
+    "duration_only": 728,
+    "interpretation_or_control": 247,
+    "note_other": 2324,
+    "null_token": 73,
+    "pitch_and_accidental": 226,
+    "pitch_and_accidental_and_tie_slur": 26,
+    "pitch_and_tie_slur": 22,
+    "pitch_only": 327,
+    "rest_note_confusion": 161,
+    "spine_operation": 181,
+    "termination": 40,
+    "tie_slur_and_articulation": 22,
+    "tie_slur_only": 1676
+  },
+  "component_error_groups": {
+    "structural": 904,
+    "visual": 6553
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 12.985633342992514,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3211,
+    "articulation": 39,
+    "barline": 28,
+    "clef": 19,
+    "column_or_line_break": 364,
+    "duration": 999,
+    "interpretation_or_control": 149,
+    "note_other": 204,
+    "null_token": 74,
+    "pitch": 20,
+    "rest": 270,
+    "spine_operation": 216,
+    "termination": 31,
+    "tie_slur": 1833
+  },
+  "gold_error_groups": {
+    "structural": 834,
+    "visual": 6623
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.614958763122559,
+      "LER": 42.33919143676758,
+      "SER": 10.225929260253906,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.10007381064215258,
+      "top1_errors": 7457,
+      "top1_errors_per_page": 111.29850746268657
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 38,
+    "predicted_unk_literal": 77
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/summary.md

@@ -0,0 +1,43 @@
+# Transcoda Dataset Eval
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_0.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Dataset kind: `hf_parquet_split`
+- Mode: `teacher_forced`
+- Pages: `67`
+- Tokens: `74,515`
+- Elapsed: `13.0s`
+
+## Metrics
+
+| mode | CER | SER | LER | exact pages | top-1 error rate | top-1 errors/page |
+|---|---:|---:|---:|---:|---:|---:|
+| `teacher_forced` | 9.614959 | 10.225929 | 42.339191 | 0.000% | 10.0074% | 111.299 |
+
+## Teacher-Forced Error Categories
+
+| category | errors |
+|---|---:|
+| `accidental` | 3,211 |
+| `tie_slur` | 1,833 |
+| `duration` | 999 |
+| `column_or_line_break` | 364 |
+| `rest` | 270 |
+| `spine_operation` | 216 |
+| `note_other` | 204 |
+| `interpretation_or_control` | 149 |
+| `null_token` | 74 |
+| `articulation` | 39 |
+| `termination` | 31 |
+| `barline` | 28 |
+| `pitch` | 20 |
+| `clef` | 19 |
+
+## Outputs
+
+- `summary.json`
+- `per_page_records.jsonl`
+- `teacher_forced_predictions.jsonl` when teacher-forced mode is enabled
+- `free_generation_predictions.jsonl` when free-generation mode is enabled
+- `gold_prefix_errors.jsonl` when teacher-forced mode is enabled
+- `teacher_forced_error_report.md` when teacher-forced mode is enabled

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_0/teacher_forced_error_report.md

@@ -0,0 +1,523 @@
+# Teacher-Forced Error Report
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_0.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Split: `eval`
+- Pages: `67`
+- Tokens: `74,515`
+- Top-1 errors: `7,457`
+- Top-1 error rate: `10.007%`
+
+## Component Buckets
+
+| bucket | errors | error share | mean/page | median/page | pages | errors/1k tok |
+|---|---:|---:|---:|---:|---:|---:|
+| `note_other` | 2,324 | 31.17% | 34.687 | 36.000 | 67 (100.0%) | 31.188 |
+| `tie_slur_only` | 1,676 | 22.48% | 25.015 | 24.000 | 67 (100.0%) | 22.492 |
+| `duration_only` | 728 | 9.76% | 10.866 | 3.000 | 61 (91.0%) | 9.770 |
+| `column_or_line_break` | 363 | 4.87% | 5.418 | 4.000 | 57 (85.1%) | 4.872 |
+| `accidental_only` | 357 | 4.79% | 5.328 | 4.000 | 61 (91.0%) | 4.791 |
+| `pitch_only` | 327 | 4.39% | 4.881 | 3.000 | 61 (91.0%) | 4.388 |
+| `interpretation_or_control` | 247 | 3.31% | 3.687 | 2.000 | 53 (79.1%) | 3.315 |
+| `pitch_and_accidental` | 226 | 3.03% | 3.373 | 2.000 | 46 (68.7%) | 3.033 |
+| `spine_operation` | 181 | 2.43% | 2.701 | 1.000 | 43 (64.2%) | 2.429 |
+| `rest_note_confusion` | 161 | 2.16% | 2.403 | 1.000 | 38 (56.7%) | 2.161 |
+| `duration_and_pitch` | 121 | 1.62% | 1.806 | 1.000 | 44 (65.7%) | 1.624 |
+| `chord_structure_and_tie_slur` | 92 | 1.23% | 1.373 | 1.000 | 40 (59.7%) | 1.235 |
+| `null_token` | 73 | 0.98% | 1.090 | 0.000 | 30 (44.8%) | 0.980 |
+| `duration_and_pitch_and_accidental` | 63 | 0.84% | 0.940 | 0.000 | 29 (43.3%) | 0.845 |
+| `chord_structure_and_articulation` | 61 | 0.82% | 0.910 | 0.000 | 23 (34.3%) | 0.819 |
+| `duration_and_tie_slur` | 58 | 0.78% | 0.866 | 0.000 | 21 (31.3%) | 0.778 |
+| `accidental_and_tie_slur` | 50 | 0.67% | 0.746 | 0.000 | 28 (41.8%) | 0.671 |
+| `chord_structure_only` | 42 | 0.56% | 0.627 | 0.000 | 22 (32.8%) | 0.564 |
+| `termination` | 40 | 0.54% | 0.597 | 0.000 | 33 (49.3%) | 0.537 |
+| `duration_and_accidental` | 37 | 0.50% | 0.552 | 0.000 | 16 (23.9%) | 0.497 |
+| `articulation_only` | 35 | 0.47% | 0.522 | 0.000 | 13 (19.4%) | 0.470 |
+| `barline` | 27 | 0.36% | 0.403 | 0.000 | 20 (29.9%) | 0.362 |
+| `pitch_and_accidental_and_tie_slur` | 26 | 0.35% | 0.388 | 0.000 | 18 (26.9%) | 0.349 |
+| `pitch_and_tie_slur` | 22 | 0.30% | 0.328 | 0.000 | 18 (26.9%) | 0.295 |
+| `tie_slur_and_articulation` | 22 | 0.30% | 0.328 | 0.000 | 16 (23.9%) | 0.295 |
+| `duration_and_pitch_and_tie_slur` | 19 | 0.25% | 0.284 | 0.000 | 9 (13.4%) | 0.255 |
+| `chord_structure_and_accidental_and_tie_slur` | 19 | 0.25% | 0.284 | 0.000 | 12 (17.9%) | 0.255 |
+| `clef` | 19 | 0.25% | 0.284 | 0.000 | 12 (17.9%) | 0.255 |
+| `chord_structure_and_accidental_and_articulation` | 11 | 0.15% | 0.164 | 0.000 | 6 (9.0%) | 0.148 |
+| `chord_structure_and_accidental` | 10 | 0.13% | 0.149 | 0.000 | 7 (10.4%) | 0.134 |
+| `chord_structure_and_tie_slur_and_articulation` | 10 | 0.13% | 0.149 | 0.000 | 8 (11.9%) | 0.134 |
+| `duration_and_pitch_and_accidental_and_tie_slur` | 5 | 0.07% | 0.075 | 0.000 | 4 (6.0%) | 0.067 |
+| `accidental_and_articulation` | 3 | 0.04% | 0.045 | 0.000 | 3 (4.5%) | 0.040 |
+| `chord_structure_and_accidental_and_tie_slur_and_articulation` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+| `duration_and_accidental_and_tie_slur` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+
+## Legacy Regex Buckets
+
+These are retained for compatibility and should not be treated as component diffs.
+
+| bucket | errors |
+|---|---:|
+| `accidental` | 3,211 |
+| `tie_slur` | 1,833 |
+| `duration` | 999 |
+| `column_or_line_break` | 364 |
+| `rest` | 270 |
+| `spine_operation` | 216 |
+| `note_other` | 204 |
+| `interpretation_or_control` | 149 |
+| `null_token` | 74 |
+| `articulation` | 39 |
+| `termination` | 31 |
+| `barline` | 28 |
+| `pitch` | 20 |
+| `clef` | 19 |
+
+## Error Samples
+
+### Sample 1
+
+```text
+8FF##q(	.
+teacher-forced top-1 local replacement:
+8FF#q((	.
+```
+
+### Sample 2
+
+```text
+4Bn:( 4d:( 4f:( 4a-:(	4aa-(
+teacher-forced top-1 local replacement:
+4An:( 4d:( 4f:( 4a-:(	4aa-(
+```
+
+### Sample 3
+
+```text
+4AA#]	4A#]
+teacher-forced top-1 local replacement:
+4AA#]	4A##]
+```
+
+### Sample 4
+
+```text
+4BBB	8.bW( 8.ddW(L
+teacher-forced top-1 local replacement:
+4BBB	8.bT( 8.ddW(L
+```
+
+### Sample 5
+
+```text
+4r	8c#(L
+teacher-forced top-1 local replacement:
+4r	8cn(L
+```
+
+### Sample 6
+
+```text
+8E;)] 8G#;)] 8e;)]J	.	8ee
+teacher-forced top-1 local replacement:
+8E;)] 8G#;)] 8e;)J	.	8ee
+```
+
+### Sample 7
+
+```text
+16f(L	.
+teacher-forced top-1 local replacement:
+16f(LL(L	.
+```
+
+### Sample 8
+
+```text
+8E 8A#	16c#)L
+teacher-forced top-1 local replacement:
+8E 8A#	16c#)LL
+```
+
+### Sample 9
+
+```text
+.	16a)J	.
+teacher-forced top-1 local replacement:
+.	16a)JJ	.
+```
+
+### Sample 10
+
+```text
+4BB	8e( 8g( 8cc#(L
+teacher-forced top-1 local replacement:
+4BB	8e( 8g 8cc#(L
+```
+
+### Sample 11
+
+```text
+.	.	16b-)J
+teacher-forced top-1 local replacement:
+.	.	16b-)JJ
+```
+
+### Sample 12
+
+```text
+16dd)J	.
+teacher-forced top-1 local replacement:
+16dd)JJ	.
+```
+
+### Sample 13
+
+```text
+.	8F 8A- 8B-J	8ffJ
+teacher-forced top-1 local replacement:
+.	8F 8A- 8B-J	8ff)J
+```
+
+### Sample 14
+
+```text
+4g-)	.	4cc( 4een(	4r
+teacher-forced top-1 local replacement:
+4g-)	.	4cc( 4een		4r
+```
+
+### Sample 15
+
+```text
+*v	*v	*
+teacher-forced top-1 local replacement:
+=	*v	*
+```
+
+### Sample 16
+
+```text
+.	.	16dd)J	.
+teacher-forced top-1 local replacement:
+.	.	16dd)JJ)J	.
+```
+
+
+## Sample Metadata
+
+### Sample 1
+
+- Sample: `polish_00071`
+- Row/token: `49` / `1102`
+- Line/field: `148:0`
+- Component bucket: `accidental_and_tie_slur`; flags: `accidental, tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `#q` -> `q(`
+- Gold note -> top-1 note: `8FF##q(` -> `8FF#q((`
+- Gold token likelihood: rank `3`, p `8.309%`
+- Predicted token likelihood: p `50.897%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 2232 | `"q("` | 50.897% |
+| 2 | 8 | `"#"` | 21.217% |
+| 3 | 1833 | `"#q"` | 8.309% |
+| 4 | 61 | `"q"` | 5.535% |
+| 5 | 59 | `"n"` | 2.236% |
+
+### Sample 2
+
+- Sample: `polish_00084`
+- Row/token: `54` / `318`
+- Line/field: `53:0`
+- Component bucket: `pitch_only`; flags: `pitch`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `4Bn` -> `4An`
+- Gold note -> top-1 note: `4Bn:(` -> `4An:(`
+- Gold token likelihood: rank `3`, p `4.993%`
+- Predicted token likelihood: p `50.428%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1360 | `"4An"` | 50.428% |
+| 2 | 127 | `"4A"` | 23.820% |
+| 3 | 1086 | `"4Bn"` | 4.993% |
+| 4 | 713 | `"4A#"` | 4.406% |
+| 5 | 1818 | `"4A("` | 2.017% |
+
+### Sample 3
+
+- Sample: `polish_00024`
+- Row/token: `15` / `1034`
+- Line/field: `168:1`
+- Component bucket: `accidental_only`; flags: `accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `4A` -> `4A#`
+- Gold note -> top-1 note: `4A#]` -> `4A##]`
+- Gold token likelihood: rank `2`, p `5.725%`
+- Predicted token likelihood: p `89.548%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 713 | `"4A#"` | 89.548% |
+| 2 | 127 | `"4A"` | 5.725% |
+| 3 | 1360 | `"4An"` | 0.964% |
+| 4 | 2774 | `"4.A#"` | 0.585% |
+| 5 | 2087 | `"2A#"` | 0.313% |
+
+### Sample 4
+
+- Sample: `polish_00040`
+- Row/token: `27` / `618`
+- Line/field: `97:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `W` -> `T`
+- Gold note -> top-1 note: `8.bW(` -> `8.bT(`
+- Gold token likelihood: rank `6`, p `3.192%`
+- Predicted token likelihood: p `13.439%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 40 | `"T"` | 13.439% |
+| 2 | 39 | `"M"` | 13.025% |
+| 3 | 65 | `"u"` | 8.410% |
+| 4 | 66 | `"v"` | 7.422% |
+| 5 | 58 | `"m"` | 3.617% |
+
+### Sample 5
+
+- Sample: `polish_00017`
+- Row/token: `10` / `912`
+- Line/field: `153:1`
+- Component bucket: `accidental_only`; flags: `accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `8c#` -> `8cn`
+- Gold note -> top-1 note: `8c#(L` -> `8cn(L`
+- Gold token likelihood: rank `2`, p `10.138%`
+- Predicted token likelihood: p `84.886%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1891 | `"8cn"` | 84.886% |
+| 2 | 246 | `"8c#"` | 10.138% |
+| 3 | 149 | `"8c"` | 1.004% |
+| 4 | 2567 | `"8c##"` | 0.326% |
+| 5 | 2825 | `"8c("` | 0.306% |
+
+### Sample 6
+
+- Sample: `polish_00049`
+- Row/token: `33` / `619`
+- Line/field: `84:0`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `]J` -> `J`
+- Gold note -> top-1 note: `8e;)]J` -> `8e;)J`
+- Gold token likelihood: rank `2`, p `1.440%`
+- Predicted token likelihood: p `89.081%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 36 | `"J"` | 89.081% |
+| 2 | 874 | `"]J"` | 1.440% |
+| 3 | 6 | `" "` | 1.353% |
+| 4 | 268 | `"[J"` | 1.194% |
+| 5 | 4 | `"\t"` | 0.959% |
+
+### Sample 7
+
+- Sample: `polish_00094`
+- Row/token: `62` / `534`
+- Line/field: `103:0`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16f` -> `16f(LL`
+- Gold note -> top-1 note: `16f(L` -> `16f(LL(L`
+- Gold token likelihood: rank `2`, p `4.878%`
+- Predicted token likelihood: p `92.043%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 2685 | `"16f(LL"` | 92.043% |
+| 2 | 107 | `"16f"` | 4.878% |
+| 3 | 2034 | `"16d(LL"` | 0.353% |
+| 4 | 502 | `"16fLL"` | 0.353% |
+| 5 | 1903 | `"16f#(LL"` | 0.267% |
+
+### Sample 8
+
+- Sample: `polish_00020`
+- Row/token: `12` / `666`
+- Line/field: `123:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `)L` -> `)LL`
+- Gold note -> top-1 note: `16c#)L` -> `16c#)LL`
+- Gold token likelihood: rank `6`, p `0.489%`
+- Predicted token likelihood: p `93.133%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 2754 | `")LL"` | 93.133% |
+| 2 | 10 | `")"` | 0.972% |
+| 3 | 410 | `"]LL"` | 0.858% |
+| 4 | 279 | `"(LL"` | 0.781% |
+| 5 | 83 | `"LL"` | 0.520% |
+
+### Sample 9
+
+- Sample: `polish_00038`
+- Row/token: `25` / `715`
+- Line/field: `87:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `)J` -> `)JJ`
+- Gold note -> top-1 note: `16a)J` -> `16a)JJ`
+- Gold token likelihood: rank `10`, p `0.661%`
+- Predicted token likelihood: p `50.858%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 305 | `")JJ"` | 50.858% |
+| 2 | 10 | `")"` | 18.709% |
+| 3 | 2262 | `")]"` | 6.466% |
+| 4 | 1570 | `")JJJ"` | 4.175% |
+| 5 | 363 | `"]JJ"` | 1.398% |
+
+### Sample 10
+
+- Sample: `polish_00037`
+- Row/token: `24` / `202`
+- Line/field: `27:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `8g(` -> `8g`
+- Gold note -> top-1 note: `8g(` -> `8g`
+- Gold token likelihood: rank `12`, p `0.009%`
+- Predicted token likelihood: p `99.198%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 77 | `"8g"` | 99.198% |
+| 2 | 357 | `"8g#"` | 0.297% |
+| 3 | 161 | `"8gL"` | 0.116% |
+| 4 | 142 | `"16g"` | 0.022% |
+| 5 | 104 | `"4g"` | 0.021% |
+
+### Sample 11
+
+- Sample: `polish_00052`
+- Row/token: `35` / `257`
+- Line/field: `46:2`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `)J` -> `)JJ`
+- Gold note -> top-1 note: `16b-)J` -> `16b-)JJ`
+- Gold token likelihood: rank `5`, p `0.102%`
+- Predicted token likelihood: p `98.262%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 305 | `")JJ"` | 98.262% |
+| 2 | 1570 | `")JJJ"` | 0.244% |
+| 3 | 10 | `")"` | 0.222% |
+| 4 | 241 | `"-JJ"` | 0.119% |
+| 5 | 186 | `")J"` | 0.102% |
+
+### Sample 12
+
+- Sample: `polish_00089`
+- Row/token: `58` / `474`
+- Line/field: `87:0`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `)J` -> `)JJ`
+- Gold note -> top-1 note: `16dd)J` -> `16dd)JJ`
+- Gold token likelihood: rank `5`, p `0.783%`
+- Predicted token likelihood: p `85.036%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 305 | `")JJ"` | 85.036% |
+| 2 | 363 | `"]JJ"` | 1.821% |
+| 3 | 1570 | `")JJJ"` | 1.213% |
+| 4 | 2387 | `"qJJ"` | 0.808% |
+| 5 | 186 | `")J"` | 0.783% |
+
+### Sample 13
+
+- Sample: `polish_00027`
+- Row/token: `17` / `1217`
+- Line/field: `174:2`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `8ffJ` -> `8ff)J`
+- Gold note -> top-1 note: `8ffJ` -> `8ff)J`
+- Gold token likelihood: rank `4`, p `2.567%`
+- Predicted token likelihood: p `85.010%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1354 | `"8ff)J"` | 85.010% |
+| 2 | 139 | `"8ff"` | 5.435% |
+| 3 | 1662 | `"8ff)"` | 2.909% |
+| 4 | 267 | `"8ffJ"` | 2.567% |
+| 5 | 1827 | `"8ff[J"` | 0.358% |
+
+### Sample 14
+
+- Sample: `polish_00087`
+- Row/token: `57` / `591`
+- Line/field: `78:2`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `(` -> `	`
+- Gold note -> top-1 note: `4een(` -> `4een`
+- Gold token likelihood: rank `2`, p `1.258%`
+- Predicted token likelihood: p `96.838%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 4 | `"\t"` | 96.838% |
+| 2 | 9 | `"("` | 1.258% |
+| 3 | 55 | `"i"` | 0.170% |
+| 4 | 8 | `"#"` | 0.080% |
+| 5 | 5 | `"\n"` | 0.052% |
+
+### Sample 15
+
+- Sample: `polish_00087`
+- Row/token: `57` / `1073`
+- Line/field: `147:0`
+- Component bucket: `spine_operation`; flags: `none`
+- Legacy bucket: `spine_operation`
+- Gold piece -> top-1 piece: `*v` -> `=`
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `4`, p `0.039%`
+- Predicted token likelihood: p `98.511%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 28 | `"="` | 98.511% |
+| 2 | 14 | `"."` | 0.112% |
+| 3 | 11 | `"*"` | 0.102% |
+| 4 | 157 | `"*v"` | 0.039% |
+| 5 | 236 | `"=\|\|"` | 0.026% |
+
+### Sample 16
+
+- Sample: `polish_00013`
+- Row/token: `7` / `1238`
+- Line/field: `147:2`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16dd` -> `16dd)JJ`
+- Gold note -> top-1 note: `16dd)J` -> `16dd)JJ)J`
+- Gold token likelihood: rank `3`, p `0.722%`
+- Predicted token likelihood: p `91.711%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1010 | `"16dd)JJ"` | 91.711% |
+| 2 | 1630 | `"16dd)"` | 3.138% |
+| 3 | 112 | `"16dd"` | 0.722% |
+| 4 | 384 | `"32dd"` | 0.341% |
+| 5 | 1992 | `"16dd("` | 0.227% |

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/command.log

@@ -0,0 +1,491 @@
+Skipping import of cpp extensions due to incompatible torch version 2.11.0+cu129 for torchao version 0.15.0             Please see https://github.com/pytorch/ao/issues/2919 for more info
+[transformers] You are using a model of type `vit_mae` to instantiate a model of type `vit`. This may be expected if you are loading a checkpoint that shares a subset of the architecture (e.g., loading a `sam2_video` checkpoint into `Sam2Model`), but is otherwise not supported and can yield errors. Please verify that the checkpoint is compatible with the model you are instantiating.
+
+[transformers] ViTModel LOAD REPORT from: PRAIG/musvit
+Key                                                                     | Status     | 
+------------------------------------------------------------------------+------------+-
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.weight          | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.weight  | UNEXPECTED | 
+decoder.decoder_embed.weight                                            | UNEXPECTED | 
+decoder.decoder_norm.bias                                               | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.bias            | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.weight          | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.bias    | UNEXPECTED | 
+decoder.decoder_pos_embed                                               | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.bias            | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.weight | UNEXPECTED | 
+decoder.decoder_norm.weight                                             | UNEXPECTED | 
+decoder.decoder_pred.weight                                             | UNEXPECTED | 
+decoder.mask_token                                                      | UNEXPECTED | 
+decoder.decoder_pred.bias                                               | UNEXPECTED | 
+decoder.decoder_embed.bias                                              | UNEXPECTED | 
+pooler.dense.weight                                                     | MISSING    | 
+pooler.dense.bias                                                       | MISSING    | 
+
+Notes:
+- UNEXPECTED:	can be ignored when loading from different task/architecture; not ok if you expect identical arch.
+- MISSING:	those params were newly initialized because missing from the checkpoint. Consider training on your downstream task.
+[eval] pages=32/67 mode=teacher_forced elapsed=9.4s
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_1.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.3132269407627806,
+      "errors": 2316,
+      "errors_per_1k_tokens": 31.08099040461652,
+      "errors_per_page_mean": 34.56716417910448,
+      "errors_per_page_median": 34.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.2258588044360292,
+      "errors": 1670,
+      "errors_per_1k_tokens": 22.411594980876334,
+      "errors_per_page_mean": 24.925373134328357,
+      "errors_per_page_median": 25.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.10035163646199621,
+      "errors": 742,
+      "errors_per_1k_tokens": 9.9577266322217,
+      "errors_per_page_mean": 11.074626865671641,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05355693805788477,
+      "errors": 396,
+      "errors_per_1k_tokens": 5.314366234986244,
+      "errors_per_page_mean": 5.91044776119403,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04395455774952664,
+      "errors": 325,
+      "errors_per_1k_tokens": 4.361537945380125,
+      "errors_per_page_mean": 4.850746268656716,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.04124966188801731,
+      "errors": 305,
+      "errors_per_1k_tokens": 4.093135610279809,
+      "errors_per_page_mean": 4.552238805970149,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.8955223880597015,
+      "pages_with_error": 60
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03381119826886665,
+      "errors": 250,
+      "errors_per_1k_tokens": 3.3550291887539423,
+      "errors_per_page_mean": 3.7313432835820897,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.7910447761194029,
+      "pages_with_error": 53
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.02664322423586692,
+      "errors": 197,
+      "errors_per_1k_tokens": 2.6437630007381063,
+      "errors_per_page_mean": 2.9402985074626864,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6417910447761194,
+      "pages_with_error": 43
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.02393832837435759,
+      "errors": 177,
+      "errors_per_1k_tokens": 2.3753606656377912,
+      "errors_per_page_mean": 2.6417910447761193,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.6268656716417911,
+      "pages_with_error": 42
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.021909656478225587,
+      "errors": 162,
+      "errors_per_1k_tokens": 2.1740589143125546,
+      "errors_per_page_mean": 2.417910447761194,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5223880597014925,
+      "pages_with_error": 35
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.015958885582905057,
+      "errors": 118,
+      "errors_per_1k_tokens": 1.5835737770918608,
+      "errors_per_page_mean": 1.7611940298507462,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6417910447761194,
+      "pages_with_error": 43
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.013253989721395727,
+      "errors": 98,
+      "errors_per_1k_tokens": 1.3151714419915452,
+      "errors_per_page_mean": 1.462686567164179,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.009872869894509062,
+      "errors": 73,
+      "errors_per_1k_tokens": 0.9796685231161512,
+      "errors_per_page_mean": 1.0895522388059702,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3582089552238806,
+      "pages_with_error": 24
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.009737625101433596,
+      "errors": 72,
+      "errors_per_1k_tokens": 0.9662484063611353,
+      "errors_per_page_mean": 1.0746268656716418,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.008249932377603462,
+      "errors": 61,
+      "errors_per_1k_tokens": 0.8186271220559619,
+      "errors_per_page_mean": 0.9104477611940298,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.007167974032999729,
+      "errors": 53,
+      "errors_per_1k_tokens": 0.7112661880158357,
+      "errors_per_page_mean": 0.7910447761194029,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.29850746268656714,
+      "pages_with_error": 20
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.005815526102245064,
+      "errors": 43,
+      "errors_per_1k_tokens": 0.5770650204656781,
+      "errors_per_page_mean": 0.6417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.373134328358209,
+      "pages_with_error": 25
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005139302136867731,
+      "errors": 38,
+      "errors_per_1k_tokens": 0.5099644366905992,
+      "errors_per_page_mean": 0.5671641791044776,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.004733567757641331,
+      "errors": 35,
+      "errors_per_1k_tokens": 0.4697040864255519,
+      "errors_per_page_mean": 0.5223880597014925,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004463078171490398,
+      "errors": 33,
+      "errors_per_1k_tokens": 0.4428638529155204,
+      "errors_per_page_mean": 0.4925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.22388059701492538,
+      "pages_with_error": 15
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.004327833378414931,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.004192588585339464,
+      "errors": 31,
+      "errors_per_1k_tokens": 0.4160236194054888,
+      "errors_per_page_mean": 0.4626865671641791,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.16417910447761194,
+      "pages_with_error": 11
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0032458750338111983,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0031106302407357318,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0028401406545847984,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.002704895861509332,
+      "errors": 20,
+      "errors_per_1k_tokens": 0.26840233510031536,
+      "errors_per_page_mean": 0.29850746268656714,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.14925373134328357,
+      "pages_with_error": 10
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.002704895861509332,
+      "errors": 20,
+      "errors_per_1k_tokens": 0.26840233510031536,
+      "errors_per_page_mean": 0.29850746268656714,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "clef",
+      "error_share": 0.002434406275358399,
+      "errors": 18,
+      "errors_per_1k_tokens": 0.24156210159028382,
+      "errors_per_page_mean": 0.26865671641791045,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.002299161482282932,
+      "errors": 17,
+      "errors_per_1k_tokens": 0.22814198483526807,
+      "errors_per_page_mean": 0.2537313432835821,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.001352447930754666,
+      "errors": 10,
+      "errors_per_1k_tokens": 0.13420116755015768,
+      "errors_per_page_mean": 0.14925373134328357,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.000676223965377333,
+      "errors": 5,
+      "errors_per_1k_tokens": 0.06710058377507884,
+      "errors_per_page_mean": 0.07462686567164178,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.04477611940298507,
+      "pages_with_error": 3
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0005409791723018664,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0005409791723018664,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.0001352447930754666,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 4,
+    "accidental_and_tie_slur": 43,
+    "accidental_only": 325,
+    "articulation_only": 31,
+    "barline": 21,
+    "chord_structure_and_accidental": 5,
+    "chord_structure_and_accidental_and_articulation": 17,
+    "chord_structure_and_accidental_and_tie_slur": 20,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 73,
+    "chord_structure_and_tie_slur": 98,
+    "chord_structure_and_tie_slur_and_articulation": 10,
+    "chord_structure_only": 35,
+    "clef": 18,
+    "column_or_line_break": 396,
+    "duration_and_accidental": 33,
+    "duration_and_pitch": 118,
+    "duration_and_pitch_and_accidental": 61,
+    "duration_and_pitch_and_accidental_and_tie_slur": 4,
+    "duration_and_pitch_and_tie_slur": 20,
+    "duration_and_tie_slur": 53,
+    "duration_only": 742,
+    "interpretation_or_control": 250,
+    "note_other": 2316,
+    "null_token": 72,
+    "pitch_and_accidental": 197,
+    "pitch_and_accidental_and_tie_slur": 32,
+    "pitch_and_tie_slur": 24,
+    "pitch_only": 305,
+    "rest_note_confusion": 162,
+    "spine_operation": 177,
+    "termination": 38,
+    "tie_slur_and_articulation": 23,
+    "tie_slur_only": 1670
+  },
+  "component_error_groups": {
+    "structural": 933,
+    "visual": 6461
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 13.113788761897013,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3159,
+    "articulation": 45,
+    "barline": 20,
+    "clef": 18,
+    "column_or_line_break": 397,
+    "duration": 985,
+    "interpretation_or_control": 152,
+    "note_other": 204,
+    "null_token": 72,
+    "pitch": 24,
+    "rest": 281,
+    "spine_operation": 201,
+    "termination": 28,
+    "tie_slur": 1808
+  },
+  "gold_error_groups": {
+    "structural": 850,
+    "visual": 6544
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.535521507263184,
+      "LER": 42.10032272338867,
+      "SER": 10.14130687713623,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.0992283432865866,
+      "top1_errors": 7394,
+      "top1_errors_per_page": 110.35820895522389
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 41,
+    "predicted_unk_literal": 94
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/per_page_records.jsonl

@@ -0,0 +1,67 @@
+{"dataset_index": 0, "sample_id": "polish_00002", "source_ids": [], "target_token_count": 1725, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 59, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 55, "field_index": 0, "field_text": "2.AA[", "gold_field": "2.AA[", "gold_line": "2.AA[\t4r\t12r", "gold_note_token": "2.AA[", "gold_piece": "[", "gold_token_id": 43, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 67, "line_char_start": 55, "line_index": 4, "line_text": "2.AA[\t4r\t12r", "operation": "replace", "pred_field": "2.AA(", "pred_local_replacement_line": "2.AA(\t4r\t12r", "pred_note_token": "2.AA(", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 13.612950699043413, "LER": 67.10526275634766, "SER": 10.150812064965198}, "nll": 0.4302118420600891, "top1_error_count": 175, "top1_error_rate": 0.10144927536231885}}
+{"dataset_index": 1, "sample_id": "polish_00004", "source_ids": [], "target_token_count": 989, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 35, "field_char_start": 30, "field_index": 0, "field_text": "8AA(L", "gold_field": "8AA(L", "gold_line": "8AA(L\t4.ee(", "gold_note_token": "8AA(L", "gold_piece": "8AA", "gold_token_id": 210, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 41, "line_char_start": 30, "line_index": 2, "line_text": "8AA(L\t4.ee(", "operation": "replace", "pred_field": "=!|:(L", "pred_local_replacement_line": "=!|:(L\t4.ee(", "pred_note_token": "=!|:(L", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 6.607142857142858, "LER": 33.96226501464844, "SER": 7.388663967611336}, "nll": 0.47386789321899414, "top1_error_count": 73, "top1_error_rate": 0.07381193124368049}}
+{"dataset_index": 2, "sample_id": "polish_00005", "source_ids": [], "target_token_count": 742, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t2cc(", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 55, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t2cc(", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t2cc(", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 5.968077723802915, "LER": 39.370079040527344, "SER": 9.31174089068826}, "nll": 0.5656912922859192, "top1_error_count": 69, "top1_error_rate": 0.09299191374663072}}
+{"dataset_index": 3, "sample_id": "polish_00007", "source_ids": [], "target_token_count": 888, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 85, "char_start": 81, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 87, "field_char_start": 81, "field_index": 1, "field_text": "16dd#L", "gold_field": "16dd#L", "gold_line": "8E 8A 8cJ\t16dd#L", "gold_note_token": "16dd#L", "gold_piece": "16dd", "gold_token_id": 112, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 87, "line_char_start": 71, "line_index": 6, "line_text": "8E 8A 8cJ\t16dd#L", "operation": "replace", "pred_field": "16dd#LL#L", "pred_local_replacement_line": "8E 8A 8cJ\t16dd#LL#L", "pred_note_token": "16dd#LL#L", "pred_piece": "16dd#LL", "pred_token_id": 1164, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 7.83558124598587, "LER": 32.66666793823242, "SER": 8.455467869222097}, "nll": 0.48426854610443115, "top1_error_count": 75, "top1_error_rate": 0.08445945945945946}}
+{"dataset_index": 4, "sample_id": "polish_00008", "source_ids": [], "target_token_count": 1236, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 48, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 48, "field_char_start": 45, "field_index": 1, "field_text": "8d[", "gold_field": "8d[", "gold_line": "8r\t8d[", "gold_note_token": "8d[", "gold_piece": "8d[", "gold_token_id": 1247, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 48, "line_char_start": 42, "line_index": 3, "line_text": "8r\t8d[", "operation": "replace", "pred_field": "8d(", "pred_local_replacement_line": "8r\t8d(", "pred_note_token": "8d(", "pred_piece": "8d(", "pred_token_id": 1822, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 10.20603737422137, "LER": 46.596858978271484, "SER": 10.283400809716598}, "nll": 0.5434036254882812, "top1_error_count": 127, "top1_error_rate": 0.10275080906148867}}
+{"dataset_index": 5, "sample_id": "polish_00009", "source_ids": [], "target_token_count": 942, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 111, "char_start": 110, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 110, "field_char_start": 107, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 52, "gold_trimmed_index": 52, "group": "structural", "line_char_end": 120, "line_char_start": 107, "line_index": 11, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 52, "visually_grounded": false}, "metrics": {"CER": 5.310300703774792, "LER": 22.22222137451172, "SER": 5.526036131774708}, "nll": 0.3397572338581085, "top1_error_count": 52, "top1_error_rate": 0.055201698513800426}}
+{"dataset_index": 6, "sample_id": "polish_00011", "source_ids": [], "target_token_count": 1093, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 7.6417004048583, "LER": 41.05263137817383, "SER": 9.24908424908425}, "nll": 0.5004439949989319, "top1_error_count": 101, "top1_error_rate": 0.09240622140896615}}
+{"dataset_index": 7, "sample_id": "polish_00013", "source_ids": [], "target_token_count": 1877, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 47, "char_start": 43, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 49, "field_char_start": 43, "field_index": 1, "field_text": "16.g(L", "gold_field": "16.g(L", "gold_line": "8CC 8GG 8CL\t16.g(L\t4en", "gold_note_token": "16.g(L", "gold_piece": "16.g", "gold_token_id": 1474, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 53, "line_char_start": 31, "line_index": 3, "line_text": "8CC 8GG 8CL\t16.g(L\t4en", "operation": "replace", "pred_field": "16.a(L", "pred_local_replacement_line": "8CC 8GG 8CL\t16.a(L\t4en", "pred_note_token": "16.a(L", "pred_piece": "16.a", "pred_token_id": 1313, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 7.65423952490927, "LER": 48.01762008666992, "SER": 8.049040511727078}, "nll": 0.46373671293258667, "top1_error_count": 151, "top1_error_rate": 0.08044752264251465}}
+{"dataset_index": 8, "sample_id": "polish_00014", "source_ids": [], "target_token_count": 1099, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t16bn(L", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 57, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t16bn(L", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t16bn(L", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 7.338991512730904, "LER": 32.03463363647461, "SER": 9.65391621129326}, "nll": 0.5251281261444092, "top1_error_count": 106, "top1_error_rate": 0.09645131938125569}}
+{"dataset_index": 9, "sample_id": "polish_00015", "source_ids": [], "target_token_count": 1086, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 38, "char_start": 37, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 38, "field_char_start": 37, "field_index": 1, "field_text": "*", "gold_field": "*", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "*^", "pred_local_replacement_line": "*^\t*^", "pred_note_token": null, "pred_piece": "*^", "pred_token_id": 191, "pred_trimmed_index": 10, "visually_grounded": false}, "metrics": {"CER": 8.503767491926803, "LER": 41.71779251098633, "SER": 9.953917050691244}, "nll": 0.4818364977836609, "top1_error_count": 108, "top1_error_rate": 0.09944751381215469}}
+{"dataset_index": 10, "sample_id": "polish_00017", "source_ids": [], "target_token_count": 1053, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 40, "char_start": 39, "component_category": "chord_structure_and_tie_slur", "component_flags": ["tie_slur", "chord_structure"], "component_group": "visual", "field_char_end": 44, "field_char_start": 34, "field_index": 0, "field_text": "4BBB- 4BB-", "gold_field": "4BBB- 4BB-", "gold_line": "4BBB- 4BB-\t4E 4cn", "gold_note_token": "4BBB-", "gold_piece": " ", "gold_token_id": 6, "gold_token_index": 9, "gold_trimmed_index": 9, "group": "visual", "line_char_end": 51, "line_char_start": 34, "line_index": 2, "line_text": "4BBB- 4BB-\t4E 4cn", "operation": "replace", "pred_field": "4BBB-(4BB-", "pred_local_replacement_line": "4BBB-(4BB-\t4E 4cn", "pred_note_token": "4BBB-(4BB-", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 9, "visually_grounded": true}, "metrics": {"CER": 2.3746701846965697, "LER": 15.55555534362793, "SER": 4.3726235741444865}, "nll": 0.2587093114852905, "top1_error_count": 46, "top1_error_rate": 0.04368471035137702}}
+{"dataset_index": 11, "sample_id": "polish_00019", "source_ids": [], "target_token_count": 1006, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 45, "char_start": 39, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 45, "field_char_start": 34, "field_index": 1, "field_text": "8.cc 8.cccL", "gold_field": "8.cc 8.cccL", "gold_line": "8AA 8AL\t8.cc 8.cccL", "gold_note_token": "8.cccL", "gold_piece": "8.cccL", "gold_token_id": 1661, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 45, "line_char_start": 26, "line_index": 2, "line_text": "8AA 8AL\t8.cc 8.cccL", "operation": "replace", "pred_field": "8.cc 8.aaL", "pred_local_replacement_line": "8AA 8AL\t8.cc 8.aaL", "pred_note_token": "8.aaL", "pred_piece": "8.aaL", "pred_token_id": 519, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 5.434782608695652, "LER": 34.117645263671875, "SER": 7.860696517412935}, "nll": 0.5182878375053406, "top1_error_count": 79, "top1_error_rate": 0.07852882703777336}}
+{"dataset_index": 12, "sample_id": "polish_00020", "source_ids": [], "target_token_count": 1183, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 5.6099732858414955, "LER": 34.92823028564453, "SER": 7.698815566835872}, "nll": 0.44988369941711426, "top1_error_count": 91, "top1_error_rate": 0.07692307692307693}}
+{"dataset_index": 13, "sample_id": "polish_00022", "source_ids": [], "target_token_count": 695, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 56, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 52, "field_index": 0, "field_text": "8G( 8B(L", "gold_field": "8G( 8B(L", "gold_line": "8G( 8B(L\t8ddd)L", "gold_note_token": "8B(L", "gold_piece": "8B(L", "gold_token_id": 2168, "gold_token_index": 20, "gold_trimmed_index": 20, "group": "visual", "line_char_end": 67, "line_char_start": 52, "line_index": 4, "line_text": "8G( 8B(L\t8ddd)L", "operation": "replace", "pred_field": "8G( 8BL", "pred_local_replacement_line": "8G( 8BL\t8ddd)L", "pred_note_token": "8BL", "pred_piece": "8BL", "pred_token_id": 396, "pred_trimmed_index": 20, "visually_grounded": true}, "metrics": {"CER": 7.894736842105263, "LER": 41.3533821105957, "SER": 10.230547550432277}, "nll": 0.6544952392578125, "top1_error_count": 71, "top1_error_rate": 0.10215827338129496}}
+{"dataset_index": 14, "sample_id": "polish_00023", "source_ids": [], "target_token_count": 1332, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 51, "field_index": 1, "field_text": "4gg)", "gold_field": "4gg)", "gold_line": "8GG 8GL\t4gg)\t8bnL", "gold_note_token": "4gg)", "gold_piece": "4gg)", "gold_token_id": 2154, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 60, "line_char_start": 43, "line_index": 3, "line_text": "8GG 8GL\t4gg)\t8bnL", "operation": "replace", "pred_field": "4gg]", "pred_local_replacement_line": "8GG 8GL\t4gg]\t8bnL", "pred_note_token": "4gg]", "pred_piece": "4gg]", "pred_token_id": 1531, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 6.977818853974122, "LER": 42.07316970825195, "SER": 7.663410969196092}, "nll": 0.406427800655365, "top1_error_count": 102, "top1_error_rate": 0.07657657657657657}}
+{"dataset_index": 15, "sample_id": "polish_00024", "source_ids": [], "target_token_count": 1075, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 7.683803206191266, "LER": 35.83815002441406, "SER": 8.56610800744879}, "nll": 0.4821993112564087, "top1_error_count": 92, "top1_error_rate": 0.08558139534883721}}
+{"dataset_index": 16, "sample_id": "polish_00025", "source_ids": [], "target_token_count": 879, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 34, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 45, "line_char_start": 34, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*M2/4", "pred_local_replacement_line": "*M2/4\t*M3/4", "pred_note_token": null, "pred_piece": "*M2/4", "pred_token_id": 304, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 6.748466257668712, "LER": 43.35664367675781, "SER": 10.136674259681094}, "nll": 0.5236173868179321, "top1_error_count": 89, "top1_error_rate": 0.1012514220705347}}
+{"dataset_index": 17, "sample_id": "polish_00027", "source_ids": [], "target_token_count": 1825, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 46, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 46, "field_index": 0, "field_text": "8BBB- 8BB-", "gold_field": "8BBB- 8BB-", "gold_line": "8BBB- 8BB-\t8b- 8dd", "gold_note_token": "8BBB-", "gold_piece": "8BBB-", "gold_token_id": 632, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 3, "line_text": "8BBB- 8BB-\t8b- 8dd", "operation": "replace", "pred_field": "8AAA 8BB-", "pred_local_replacement_line": "8AAA 8BB-\t8b- 8dd", "pred_note_token": "8AAA", "pred_piece": "8AAA", "pred_token_id": 490, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 8.563938246431691, "LER": 50.77519226074219, "SER": 9.649122807017543}, "nll": 0.4922695755958557, "top1_error_count": 176, "top1_error_rate": 0.09643835616438357}}
+{"dataset_index": 18, "sample_id": "polish_00029", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 39, "field_char_start": 30, "field_index": 0, "field_text": "8CC( 8C(L", "gold_field": "8CC( 8C(L", "gold_line": "8CC( 8C(L\t4gg([", "gold_note_token": "8CC(", "gold_piece": "8CC", "gold_token_id": 364, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 45, "line_char_start": 30, "line_index": 2, "line_text": "8CC( 8C(L\t4gg([", "operation": "replace", "pred_field": "=!|:( 8C(L", "pred_local_replacement_line": "=!|:( 8C(L\t4gg([", "pred_note_token": "=!|:(", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 7.962328767123288, "LER": 36.17021179199219, "SER": 10.217216411906678}, "nll": 0.5909623503684998, "top1_error_count": 127, "top1_error_rate": 0.10209003215434084}}
+{"dataset_index": 19, "sample_id": "polish_00030", "source_ids": [], "target_token_count": 1263, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "=!|:", "pred_local_replacement_line": "=!|:\t*^", "pred_note_token": null, "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 12.205270457697642, "LER": 52.093021392822266, "SER": 10.142630744849445}, "nll": 0.45661047101020813, "top1_error_count": 128, "top1_error_rate": 0.10134600158353127}}
+{"dataset_index": 20, "sample_id": "polish_00031", "source_ids": [], "target_token_count": 862, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 34, "field_index": 0, "field_text": "*M3/8", "gold_field": "*M3/8", "gold_line": "*M3/8\t*M3/8", "gold_note_token": null, "gold_piece": "*M3/8", "gold_token_id": 592, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 45, "line_char_start": 34, "line_index": 2, "line_text": "*M3/8\t*M3/8", "operation": "replace", "pred_field": "*M3/4", "pred_local_replacement_line": "*M3/4\t*M3/8", "pred_note_token": null, "pred_piece": "*M3/4", "pred_token_id": 283, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 10.50228310502283, "LER": 42.65734100341797, "SER": 10.220673635307783}, "nll": 0.5673506855964661, "top1_error_count": 88, "top1_error_rate": 0.10208816705336426}}
+{"dataset_index": 21, "sample_id": "polish_00032", "source_ids": [], "target_token_count": 1387, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 50, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 52, "field_char_start": 45, "field_index": 2, "field_text": "16ff#]L", "gold_field": "16ff#]L", "gold_line": "4G#(\t8G#L\t16ff#]L", "gold_note_token": "16ff#]L", "gold_piece": "16ff#", "gold_token_id": 180, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 52, "line_char_start": 35, "line_index": 3, "line_text": "4G#(\t8G#L\t16ff#]L", "operation": "replace", "pred_field": "16ff#)]L", "pred_local_replacement_line": "4G#(\t8G#L\t16ff#)]L", "pred_note_token": "16ff#)]L", "pred_piece": "16ff#)", "pred_token_id": 2192, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 11.022964509394571, "LER": 43.697479248046875, "SER": 11.832611832611832}, "nll": 0.6952135562896729, "top1_error_count": 164, "top1_error_rate": 0.11824080749819754}}
+{"dataset_index": 22, "sample_id": "polish_00034", "source_ids": [], "target_token_count": 788, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 45, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8f( 8a(L", "gold_field": "8f( 8a(L", "gold_line": "8f( 8a(L\t16r", "gold_note_token": null, "gold_piece": "8f(", "gold_token_id": 2288, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 42, "line_index": 3, "line_text": "8f( 8a(L\t16r", "operation": "replace", "pred_field": "* 8a(L", "pred_local_replacement_line": "* 8a(L\t16r", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 24.081115335868187, "LER": 58.385093688964844, "SER": 38.75476493011436}, "nll": 1.4015973806381226, "top1_error_count": 176, "top1_error_rate": 0.2233502538071066}}
+{"dataset_index": 23, "sample_id": "polish_00036", "source_ids": [], "target_token_count": 1317, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 451, "char_start": 447, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 451, "field_char_start": 447, "field_index": 2, "field_text": "16gJ", "gold_field": "16gJ", "gold_line": ".\t.\t16gJ", "gold_note_token": "16gJ", "gold_piece": "16gJ", "gold_token_id": 859, "gold_token_index": 280, "gold_trimmed_index": 280, "group": "visual", "line_char_end": 451, "line_char_start": 443, "line_index": 41, "line_text": ".\t.\t16gJ", "operation": "replace", "pred_field": "16gJJ", "pred_local_replacement_line": ".\t.\t16gJJ", "pred_note_token": "16gJJ", "pred_piece": "16gJJ", "pred_token_id": 351, "pred_trimmed_index": 280, "visually_grounded": true}, "metrics": {"CER": 5.83974649162517, "LER": 23.95833396911621, "SER": 5.091185410334346}, "nll": 0.3015836775302887, "top1_error_count": 67, "top1_error_rate": 0.05087319665907365}}
+{"dataset_index": 24, "sample_id": "polish_00037", "source_ids": [], "target_token_count": 871, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 49, "char_start": 46, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 46, "field_index": 0, "field_text": "4Fn 4dn", "gold_field": "4Fn 4dn", "gold_line": "4Fn 4dn\t2f#] 2ff#]", "gold_note_token": "4Fn", "gold_piece": "4Fn", "gold_token_id": 1918, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 2, "line_text": "4Fn 4dn\t2f#] 2ff#]", "operation": "replace", "pred_field": "4F# 4dn", "pred_local_replacement_line": "4F# 4dn\t2f#] 2ff#]", "pred_note_token": "4F#", "pred_piece": "4F#", "pred_token_id": 278, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 4.054937867887508, "LER": 21.476511001586914, "SER": 5.517241379310345}, "nll": 0.2720746397972107, "top1_error_count": 48, "top1_error_rate": 0.05510907003444317}}
+{"dataset_index": 25, "sample_id": "polish_00038", "source_ids": [], "target_token_count": 1471, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 88, "char_start": 83, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 88, "field_char_start": 83, "field_index": 1, "field_text": "16a(L", "gold_field": "16a(L", "gold_line": "8CC 8FF# 8CJ\t16a(L\t8e-", "gold_note_token": "16a(L", "gold_piece": "16a(L", "gold_token_id": 2253, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 92, "line_char_start": 70, "line_index": 5, "line_text": "8CC 8FF# 8CJ\t16a(L\t8e-", "operation": "replace", "pred_field": "16aLL", "pred_local_replacement_line": "8CC 8FF# 8CJ\t16aLL\t8e-", "pred_note_token": "16aLL", "pred_piece": "16aLL", "pred_token_id": 338, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 6.27498001598721, "LER": 37.988826751708984, "SER": 7.482993197278912}, "nll": 0.425575315952301, "top1_error_count": 110, "top1_error_rate": 0.07477906186267845}}
+{"dataset_index": 26, "sample_id": "polish_00039", "source_ids": [], "target_token_count": 1503, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 93, "char_start": 88, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 93, "field_char_start": 88, "field_index": 0, "field_text": "16A-L", "gold_field": "16A-L", "gold_line": "16A-L\t2.AA-[\t16cL\t8ee-", "gold_note_token": "16A-L", "gold_piece": "16A-L", "gold_token_id": 2834, "gold_token_index": 24, "gold_trimmed_index": 24, "group": "visual", "line_char_end": 110, "line_char_start": 88, "line_index": 4, "line_text": "16A-L\t2.AA-[\t16cL\t8ee-", "operation": "replace", "pred_field": "16A-LL", "pred_local_replacement_line": "16A-LL\t2.AA-[\t16cL\t8ee-", "pred_note_token": "16A-LL", "pred_piece": "16A-LL", "pred_token_id": 1109, "pred_trimmed_index": 24, "visually_grounded": true}, "metrics": {"CER": 8.95230648944488, "LER": 39.68254089355469, "SER": 9.121171770972039}, "nll": 0.5090069770812988, "top1_error_count": 137, "top1_error_rate": 0.09115103127079174}}
+{"dataset_index": 27, "sample_id": "polish_00040", "source_ids": [], "target_token_count": 1109, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 8.843873517786562, "LER": 45.45454406738281, "SER": 10.740072202166065}, "nll": 0.5333619117736816, "top1_error_count": 119, "top1_error_rate": 0.1073038773669973}}
+{"dataset_index": 28, "sample_id": "polish_00043", "source_ids": [], "target_token_count": 885, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "replace", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 13.208649912331968, "LER": 55.491329193115234, "SER": 16.855203619909503}, "nll": 0.9032096862792969, "top1_error_count": 149, "top1_error_rate": 0.16836158192090395}}
+{"dataset_index": 29, "sample_id": "polish_00044", "source_ids": [], "target_token_count": 677, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 74, "char_start": 73, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 82, "field_char_start": 70, "field_index": 1, "field_text": "16f( 16dd(Jk", "gold_field": "16f( 16dd(Jk", "gold_line": "16B-(Jk\t16f( 16dd(Jk", "gold_note_token": "16f(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 22, "gold_trimmed_index": 22, "group": "visual", "line_char_end": 82, "line_char_start": 62, "line_index": 4, "line_text": "16B-(Jk\t16f( 16dd(Jk", "operation": "delete", "pred_field": "16f  16dd(Jk", "pred_local_replacement_line": "16B-(Jk\t16f  16dd(Jk", "pred_note_token": "16f", "pred_piece": " ", "pred_token_id": 6, "pred_trimmed_index": 22, "visually_grounded": true}, "metrics": {"CER": 8.533747090768037, "LER": 47.93388366699219, "SER": 10.502958579881657}, "nll": 0.6337476968765259, "top1_error_count": 71, "top1_error_rate": 0.10487444608567208}}
+{"dataset_index": 30, "sample_id": "polish_00046", "source_ids": [], "target_token_count": 1069, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "clef", "char_end": 15, "char_start": 8, "component_category": "clef", "component_flags": [], "component_group": "visual", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2", "gold_note_token": null, "gold_piece": "*clefG2", "gold_token_id": 147, "gold_token_index": 2, "gold_trimmed_index": 2, "group": "visual", "line_char_end": 15, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2", "operation": "replace", "pred_field": "*clefG", "pred_local_replacement_line": "*clefF4\t*clefG", "pred_note_token": null, "pred_piece": "*clefG", "pred_token_id": 145, "pred_trimmed_index": 2, "visually_grounded": true}, "metrics": {"CER": 7.906976744186046, "LER": 37.25490188598633, "SER": 8.707865168539326}, "nll": 0.363349586725235, "top1_error_count": 93, "top1_error_rate": 0.08699719363891488}}
+{"dataset_index": 31, "sample_id": "polish_00047", "source_ids": [], "target_token_count": 1243, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 66, "char_start": 63, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 68, "field_char_start": 63, "field_index": 1, "field_text": "8eeq(", "gold_field": "8eeq(", "gold_line": ".\t8eeq(\t.", "gold_note_token": "8eeq(", "gold_piece": "8ee", "gold_token_id": 84, "gold_token_index": 28, "gold_trimmed_index": 28, "group": "visual", "line_char_end": 70, "line_char_start": 61, "line_index": 4, "line_text": ".\t8eeq(\t.", "operation": "replace", "pred_field": "8ddq(", "pred_local_replacement_line": ".\t8ddq(\t.", "pred_note_token": "8ddq(", "pred_piece": "8dd", "pred_token_id": 79, "pred_trimmed_index": 28, "visually_grounded": true}, "metrics": {"CER": 5.182341650671785, "LER": 27.439023971557617, "SER": 5.233494363929147}, "nll": 0.32566750049591064, "top1_error_count": 65, "top1_error_rate": 0.05229283990345937}}
+{"dataset_index": 32, "sample_id": "polish_00048", "source_ids": [], "target_token_count": 1384, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 51, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 48, "field_index": 0, "field_text": "4c 4f", "gold_field": "4c 4f", "gold_line": "4c 4f\t12aa-(L", "gold_note_token": "4f", "gold_piece": "4f", "gold_token_id": 103, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 48, "line_index": 3, "line_text": "4c 4f\t12aa-(L", "operation": "replace", "pred_field": "4c 4a-", "pred_local_replacement_line": "4c 4a-\t12aa-(L", "pred_note_token": "4a-", "pred_piece": "4a-", "pred_token_id": 341, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 15.186360031720856, "LER": 47.56554412841797, "SER": 10.773680404916847}, "nll": 0.5395410060882568, "top1_error_count": 149, "top1_error_rate": 0.10765895953757225}}
+{"dataset_index": 33, "sample_id": "polish_00049", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 13.263263263263264, "LER": 44.578311920166016, "SER": 12.308930008045053}, "nll": 0.5278508067131042, "top1_error_count": 127, "top1_error_rate": 0.10209003215434084}}
+{"dataset_index": 34, "sample_id": "polish_00050", "source_ids": [], "target_token_count": 1242, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "replace", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 12.494194147700883, "LER": 54.01069641113281, "SER": 11.522965350523771}, "nll": 0.5604510307312012, "top1_error_count": 143, "top1_error_rate": 0.11513687600644122}}
+{"dataset_index": 35, "sample_id": "polish_00052", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 52, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 59, "field_char_start": 47, "field_index": 0, "field_text": "2.DD-( 2.D-(", "gold_field": "2.DD-( 2.D-(", "gold_line": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "gold_note_token": "2.DD-(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 74, "line_char_start": 47, "line_index": 3, "line_text": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "operation": "replace", "pred_field": "2.DD-([ 2.D-(", "pred_local_replacement_line": "2.DD-([ 2.D-(\t8ff(L\t2.f 2.a-", "pred_note_token": "2.DD-([", "pred_piece": "([", "pred_token_id": 2909, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 10.48158640226629, "LER": 45.35519027709961, "SER": 13.477366255144032}, "nll": 0.7258648872375488, "top1_error_count": 131, "top1_error_rate": 0.13463514902363824}}
+{"dataset_index": 36, "sample_id": "polish_00053", "source_ids": [], "target_token_count": 1018, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 48, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 62, "field_char_start": 48, "field_index": 0, "field_text": "4FFn: 4Cn: 4A:", "gold_field": "4FFn: 4Cn: 4A:", "gold_line": "4FFn: 4Cn: 4A:\t8fn", "gold_note_token": "4FFn:", "gold_piece": "4FF", "gold_token_id": 369, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 66, "line_char_start": 48, "line_index": 4, "line_text": "4FFn: 4Cn: 4A:\t8fn", "operation": "replace", "pred_field": "4FF#n: 4Cn: 4A:", "pred_local_replacement_line": "4FF#n: 4Cn: 4A:\t8fn", "pred_note_token": "4FF#n:", "pred_piece": "4FF#", "pred_token_id": 607, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 3.7433155080213902, "LER": 16.5745849609375, "SER": 4.2281219272369714}, "nll": 0.27027153968811035, "top1_error_count": 43, "top1_error_rate": 0.04223968565815324}}
+{"dataset_index": 37, "sample_id": "polish_00054", "source_ids": [], "target_token_count": 689, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 46, "field_index": 0, "field_text": "8B-( 8d(L", "gold_field": "8B-( 8d(L", "gold_line": "8B-( 8d(L\t8fL", "gold_note_token": "8d(L", "gold_piece": "8d(L", "gold_token_id": 1103, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 59, "line_char_start": 46, "line_index": 3, "line_text": "8B-( 8d(L\t8fL", "operation": "replace", "pred_field": "8B-( 8dL", "pred_local_replacement_line": "8B-( 8dL\t8fL", "pred_note_token": "8dL", "pred_piece": "8dL", "pred_token_id": 206, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 9.585889570552148, "LER": 40.44117736816406, "SER": 11.482558139534884}, "nll": 0.7468580603599548, "top1_error_count": 79, "top1_error_rate": 0.11465892597968069}}
+{"dataset_index": 38, "sample_id": "polish_00056", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 28, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 28, "field_char_start": 16, "field_index": 0, "field_text": "*k[b-e-a-d-]", "gold_field": "*k[b-e-a-d-]", "gold_line": "*k[b-e-a-d-]\t*k[b-e-a-d-]", "gold_note_token": null, "gold_piece": "*k[b-e-a-d-]", "gold_token_id": 637, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 41, "line_char_start": 16, "line_index": 1, "line_text": "*k[b-e-a-d-]\t*k[b-e-a-d-]", "operation": "replace", "pred_field": "*k[b-e-a-d-g-c-]", "pred_local_replacement_line": "*k[b-e-a-d-g-c-]\t*k[b-e-a-d-]", "pred_note_token": null, "pred_piece": "*k[b-e-a-d-g-c-]", "pred_token_id": 827, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 13.691275167785236, "LER": 45.62211990356445, "SER": 13.612099644128115}, "nll": 0.5483160614967346, "top1_error_count": 153, "top1_error_rate": 0.136}}
+{"dataset_index": 39, "sample_id": "polish_00057", "source_ids": [], "target_token_count": 733, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 52, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 57, "field_char_start": 52, "field_index": 1, "field_text": "8ff#q", "gold_field": "8ff#q", "gold_line": ".\t8ff#q", "gold_note_token": "8ff#q", "gold_piece": "8ff#", "gold_token_id": 92, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": ".\t8ff#q", "operation": "replace", "pred_field": "8ee-q", "pred_local_replacement_line": ".\t8ee-q", "pred_note_token": "8ee-q", "pred_piece": "8ee-", "pred_token_id": 330, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 12.905405405405403, "LER": 46.853145599365234, "SER": 15.573770491803279}, "nll": 0.8839104771614075, "top1_error_count": 114, "top1_error_rate": 0.15552523874488403}}
+{"dataset_index": 40, "sample_id": "polish_00058", "source_ids": [], "target_token_count": 900, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 68, "char_start": 64, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 73, "field_char_start": 64, "field_index": 0, "field_text": "8.FF 8.FL", "gold_field": "8.FF 8.FL", "gold_line": "8.FF 8.FL\t8.fL", "gold_note_token": "8.FF", "gold_piece": "8.FF", "gold_token_id": 1569, "gold_token_index": 20, "gold_trimmed_index": 20, "group": "visual", "line_char_end": 78, "line_char_start": 64, "line_index": 5, "line_text": "8.FF 8.FL\t8.fL", "operation": "replace", "pred_field": "8FF 8.FL", "pred_local_replacement_line": "8FF 8.FL\t8.fL", "pred_note_token": "8FF", "pred_piece": "8FF", "pred_token_id": 276, "pred_trimmed_index": 20, "visually_grounded": true}, "metrics": {"CER": 7.173252279635259, "LER": 34.28571319580078, "SER": 8.008898776418242}, "nll": 0.4937729239463806, "top1_error_count": 72, "top1_error_rate": 0.08}}
+{"dataset_index": 41, "sample_id": "polish_00059", "source_ids": [], "target_token_count": 1838, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 60, "char_start": 56, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 60, "field_char_start": 56, "field_index": 1, "field_text": "8AAL", "gold_field": "8AAL", "gold_line": "4.AA\t8AAL\t24aa", "gold_note_token": "8AAL", "gold_piece": "8AAL", "gold_token_id": 512, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 65, "line_char_start": 51, "line_index": 4, "line_text": "4.AA\t8AAL\t24aa", "operation": "replace", "pred_field": "8.AAL", "pred_local_replacement_line": "4.AA\t8.AAL\t24aa", "pred_note_token": "8.AAL", "pred_piece": "8.AAL", "pred_token_id": 2335, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 13.566936208445643, "LER": 67.56756591796875, "SER": 13.77245508982036}, "nll": 0.72260981798172, "top1_error_count": 253, "top1_error_rate": 0.13764961915125137}}
+{"dataset_index": 42, "sample_id": "polish_00060", "source_ids": [], "target_token_count": 1528, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 62, "char_start": 59, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 66, "field_char_start": 59, "field_index": 0, "field_text": "8FF 8FL", "gold_field": "8FF 8FL", "gold_line": "8FF 8FL\t2.F\t4.f", "gold_note_token": "8FF", "gold_piece": "8FF", "gold_token_id": 276, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 74, "line_char_start": 59, "line_index": 4, "line_text": "8FF 8FL\t2.F\t4.f", "operation": "replace", "pred_field": "8FL 8FL", "pred_local_replacement_line": "8FL 8FL\t2.F\t4.f", "pred_note_token": "8FL", "pred_piece": "8FL", "pred_token_id": 433, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 6.3852813852813854, "LER": 44.85981369018555, "SER": 8.316961362148003}, "nll": 0.5030782222747803, "top1_error_count": 127, "top1_error_rate": 0.0831151832460733}}
+{"dataset_index": 43, "sample_id": "polish_00061", "source_ids": [], "target_token_count": 653, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 50, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 59, "field_char_start": 50, "field_index": 0, "field_text": "8e-( 8g(L", "gold_field": "8e-( 8g(L", "gold_line": "8e-( 8g(L\t8gg 8bb-L", "gold_note_token": null, "gold_piece": "8e-", "gold_token_id": 263, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 69, "line_char_start": 50, "line_index": 3, "line_text": "8e-( 8g(L\t8gg 8bb-L", "operation": "replace", "pred_field": "*( 8g(L", "pred_local_replacement_line": "*( 8g(L\t8gg 8bb-L", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 11.086309523809524, "LER": 44.8979606628418, "SER": 15.644171779141105}, "nll": 0.7570512294769287, "top1_error_count": 102, "top1_error_rate": 0.1562021439509954}}
+{"dataset_index": 44, "sample_id": "polish_00062", "source_ids": [], "target_token_count": 781, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 10.866013071895425, "LER": 42.99065399169922, "SER": 9.230769230769232}, "nll": 0.46511903405189514, "top1_error_count": 72, "top1_error_rate": 0.09218950064020487}}
+{"dataset_index": 45, "sample_id": "polish_00063", "source_ids": [], "target_token_count": 963, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 54, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 56, "field_char_start": 48, "field_index": 1, "field_text": "16qqddnL", "gold_field": "16qqddnL", "gold_line": ".\t16qqddnL", "gold_note_token": "16qqddnL", "gold_piece": "nL", "gold_token_id": 403, "gold_token_index": 11, "gold_trimmed_index": 11, "group": "visual", "line_char_end": 56, "line_char_start": 46, "line_index": 2, "line_text": ".\t16qqddnL", "operation": "replace", "pred_field": "16qqddnLL", "pred_local_replacement_line": ".\t16qqddnLL", "pred_note_token": "16qqddnLL", "pred_piece": "nLL", "pred_token_id": 822, "pred_trimmed_index": 11, "visually_grounded": true}, "metrics": {"CER": 4.41988950276243, "LER": 28.947368621826172, "SER": 6.548856548856549}, "nll": 0.35238608717918396, "top1_error_count": 63, "top1_error_rate": 0.06542056074766354}}
+{"dataset_index": 46, "sample_id": "polish_00064", "source_ids": [], "target_token_count": 928, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 53, "char_start": 49, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 53, "field_char_start": 49, "field_index": 1, "field_text": "8dd[", "gold_field": "8dd[", "gold_line": "8r\t8dd[", "gold_note_token": "8dd[", "gold_piece": "8dd[", "gold_token_id": 1303, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 53, "line_char_start": 46, "line_index": 3, "line_text": "8r\t8dd[", "operation": "replace", "pred_field": "8dd(", "pred_local_replacement_line": "8r\t8dd(", "pred_note_token": "8dd(", "pred_piece": "8dd(", "pred_token_id": 1124, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 9.376786735277301, "LER": 44.63276672363281, "SER": 13.160733549083064}, "nll": 0.7679115533828735, "top1_error_count": 122, "top1_error_rate": 0.1314655172413793}}
+{"dataset_index": 47, "sample_id": "polish_00067", "source_ids": [], "target_token_count": 1138, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 43, "char_start": 39, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 43, "field_char_start": 39, "field_index": 1, "field_text": "16aL", "gold_field": "16aL", "gold_line": "8F(L\t16aL", "gold_note_token": "16aL", "gold_piece": "16aL", "gold_token_id": 602, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 43, "line_char_start": 34, "line_index": 2, "line_text": "8F(L\t16aL", "operation": "replace", "pred_field": "16aLL", "pred_local_replacement_line": "8F(L\t16aLL", "pred_note_token": "16aLL", "pred_piece": "16aLL", "pred_token_id": 338, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 11.92833569071193, "LER": 45.3125, "SER": 14.8636763412489}, "nll": 0.9122363924980164, "top1_error_count": 168, "top1_error_rate": 0.14762741652021089}}
+{"dataset_index": 48, "sample_id": "polish_00068", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 57, "char_start": 53, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 57, "field_char_start": 53, "field_index": 1, "field_text": "4ff#", "gold_field": "4ff#", "gold_line": "4r\t4ff#", "gold_note_token": "4ff#", "gold_piece": "4ff#", "gold_token_id": 225, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": "4r\t4ff#", "operation": "replace", "pred_field": "4ff#(", "pred_local_replacement_line": "4r\t4ff#(", "pred_note_token": "4ff#(", "pred_piece": "4ff#(", "pred_token_id": 2286, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 22.229219143576824, "LER": 64.64088439941406, "SER": 20.5761316872428}, "nll": 0.9971667528152466, "top1_error_count": 200, "top1_error_rate": 0.20554984583761562}}
+{"dataset_index": 49, "sample_id": "polish_00071", "source_ids": [], "target_token_count": 1433, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 48, "char_start": 46, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 48, "field_char_start": 46, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*^", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 51, "line_char_start": 46, "line_index": 2, "line_text": "*^\t*^", "operation": "replace", "pred_field": ".", "pred_local_replacement_line": ".\t*^", "pred_note_token": null, "pred_piece": ".", "pred_token_id": 14, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 11.668070766638584, "LER": 60.109291076660156, "SER": 13.547486033519554}, "nll": 0.702937662601471, "top1_error_count": 194, "top1_error_rate": 0.13538032100488487}}
+{"dataset_index": 50, "sample_id": "polish_00072", "source_ids": [], "target_token_count": 1123, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 56, "char_start": 55, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 13, "gold_trimmed_index": 13, "group": "structural", "line_char_end": 65, "line_char_start": 52, "line_index": 3, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 13, "visually_grounded": false}, "metrics": {"CER": 4.992435703479576, "LER": 29.787233352661133, "SER": 6.417112299465241}, "nll": 0.29150640964508057, "top1_error_count": 72, "top1_error_rate": 0.06411398040961709}}
+{"dataset_index": 51, "sample_id": "polish_00074", "source_ids": [], "target_token_count": 1355, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 34, "char_start": 30, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 36, "field_char_start": 30, "field_index": 0, "field_text": "8BBnq(", "gold_field": "8BBnq(", "gold_line": "8BBnq(\t.", "gold_note_token": null, "gold_piece": "8BBn", "gold_token_id": 2042, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 38, "line_char_start": 30, "line_index": 2, "line_text": "8BBnq(\t.", "operation": "replace", "pred_field": "*M6/8q(", "pred_local_replacement_line": "*M6/8q(\t.", "pred_note_token": null, "pred_piece": "*M6/8", "pred_token_id": 329, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 8.965823346648913, "LER": 43.07692337036133, "SER": 9.527326440177253}, "nll": 0.4643954634666443, "top1_error_count": 129, "top1_error_rate": 0.0952029520295203}}
+{"dataset_index": 52, "sample_id": "polish_00078", "source_ids": [], "target_token_count": 1172, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 49, "char_start": 48, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 49, "field_char_start": 48, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*\t*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 56, "line_char_start": 48, "line_index": 3, "line_text": "*\t*\t*^\t*", "operation": "insert", "pred_field": "4Dn", "pred_local_replacement_line": "4Dn\t*\t*^\t*", "pred_note_token": null, "pred_piece": "4Dn", "pred_token_id": 2294, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 16.960941680042804, "LER": 44.80519485473633, "SER": 12.980358667805294}, "nll": 0.5585002899169922, "top1_error_count": 152, "top1_error_rate": 0.1296928327645051}}
+{"dataset_index": 53, "sample_id": "polish_00082", "source_ids": [], "target_token_count": 1027, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "8AA-L", "pred_local_replacement_line": "8AA-L\t*^", "pred_note_token": null, "pred_piece": "8AA-L", "pred_token_id": 901, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 13.898117386489481, "LER": 50.657894134521484, "SER": 10.1364522417154}, "nll": 0.5777429938316345, "top1_error_count": 104, "top1_error_rate": 0.10126582278481013}}
+{"dataset_index": 54, "sample_id": "polish_00084", "source_ids": [], "target_token_count": 860, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 54, "char_start": 50, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 57, "field_char_start": 50, "field_index": 0, "field_text": "32qqCCL", "gold_field": "32qqCCL", "gold_line": "32qqCCL\t32qqe-L", "gold_note_token": "32qqCCL", "gold_piece": "32qq", "gold_token_id": 1235, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 65, "line_char_start": 50, "line_index": 3, "line_text": "32qqCCL\t32qqe-L", "operation": "replace", "pred_field": "32CCCCL", "pred_local_replacement_line": "32CCCCL\t32qqe-L", "pred_note_token": "32CCCCL", "pred_piece": "32CC", "pred_token_id": 1559, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 11.951219512195122, "LER": 44.58598709106445, "SER": 13.62048894062864}, "nll": 0.7850427627563477, "top1_error_count": 117, "top1_error_rate": 0.13604651162790699}}
+{"dataset_index": 55, "sample_id": "polish_00085", "source_ids": [], "target_token_count": 1428, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 117, "char_start": 113, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 119, "field_char_start": 113, "field_index": 2, "field_text": "16ee)J", "gold_field": "16ee)J", "gold_line": ".\t.\t16ee)J\t.", "gold_note_token": "16ee)J", "gold_piece": "16ee", "gold_token_id": 126, "gold_token_index": 56, "gold_trimmed_index": 56, "group": "visual", "line_char_end": 121, "line_char_start": 109, "line_index": 7, "line_text": ".\t.\t16ee)J\t.", "operation": "replace", "pred_field": "16ee)JJ)J", "pred_local_replacement_line": ".\t.\t16ee)JJ)J\t.", "pred_note_token": "16ee)JJ)J", "pred_piece": "16ee)JJ", "pred_token_id": 1119, "pred_trimmed_index": 56, "visually_grounded": true}, "metrics": {"CER": 4.618736383442266, "LER": 23.60248374938965, "SER": 4.134548002803083}, "nll": 0.23760364949703217, "top1_error_count": 59, "top1_error_rate": 0.04131652661064426}}
+{"dataset_index": 56, "sample_id": "polish_00086", "source_ids": [], "target_token_count": 1170, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 44, "field_char_start": 42, "field_index": 0, "field_text": "4C", "gold_field": "4C", "gold_line": "4C\t4r", "gold_note_token": null, "gold_piece": "4C", "gold_token_id": 181, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 47, "line_char_start": 42, "line_index": 2, "line_text": "4C\t4r", "operation": "replace", "pred_field": "*M4/4", "pred_local_replacement_line": "*M4/4\t4r", "pred_note_token": null, "pred_piece": "*M4/4", "pred_token_id": 229, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 8.635496183206106, "LER": 40.85365676879883, "SER": 8.468776732249786}, "nll": 0.3928833305835724, "top1_error_count": 99, "top1_error_rate": 0.08461538461538462}}
+{"dataset_index": 57, "sample_id": "polish_00087", "source_ids": [], "target_token_count": 1300, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 47, "char_start": 44, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 56, "field_char_start": 44, "field_index": 1, "field_text": "8ccq( 8ee-q(", "gold_field": "8ccq( 8ee-q(", "gold_line": ".\t8ccq( 8ee-q(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 56, "line_char_start": 42, "line_index": 2, "line_text": ".\t8ccq( 8ee-q(", "operation": "replace", "pred_field": "8ccqq( 8ee-q(", "pred_local_replacement_line": ".\t8ccqq( 8ee-q(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 13.79460400348129, "LER": 55.434783935546875, "SER": 11.932255581216321}, "nll": 0.5800479054450989, "top1_error_count": 155, "top1_error_rate": 0.11923076923076924}}
+{"dataset_index": 58, "sample_id": "polish_00089", "source_ids": [], "target_token_count": 782, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 37, "char_start": 35, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 43, "field_char_start": 35, "field_index": 0, "field_text": "8A( 8c(L", "gold_field": "8A( 8c(L", "gold_line": "8A( 8c(L\t2.D\t16g#(L", "gold_note_token": "8A(", "gold_piece": "8A", "gold_token_id": 109, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 35, "line_index": 3, "line_text": "8A( 8c(L\t2.D\t16g#(L", "operation": "replace", "pred_field": "8A(L( 8c(L", "pred_local_replacement_line": "8A(L( 8c(L\t2.D\t16g#(L", "pred_note_token": "8A(L(", "pred_piece": "8A(L", "pred_token_id": 1848, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 12.744436952124072, "LER": 49.6894416809082, "SER": 15.236875800256081}, "nll": 1.0106003284454346, "top1_error_count": 119, "top1_error_rate": 0.15217391304347827}}
+{"dataset_index": 59, "sample_id": "polish_00090", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 276, "char_start": 272, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 276, "field_char_start": 272, "field_index": 0, "field_text": "8AAL", "gold_field": "8AAL", "gold_line": "8AAL\t8.r\t4ee", "gold_note_token": "8AAL", "gold_piece": "8AAL", "gold_token_id": 512, "gold_token_index": 164, "gold_trimmed_index": 164, "group": "visual", "line_char_end": 284, "line_char_start": 272, "line_index": 22, "line_text": "8AAL\t8.r\t4ee", "operation": "replace", "pred_field": "8AA", "pred_local_replacement_line": "8AA\t8.r\t4ee", "pred_note_token": "8AA", "pred_piece": "8AA", "pred_token_id": 210, "pred_trimmed_index": 164, "visually_grounded": true}, "metrics": {"CER": 8.388520971302428, "LER": 34.228187561035156, "SER": 7.206405693950178}, "nll": 0.4045591950416565, "top1_error_count": 81, "top1_error_rate": 0.072}}
+{"dataset_index": 60, "sample_id": "polish_00092", "source_ids": [], "target_token_count": 1153, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 62, "char_start": 61, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 62, "field_char_start": 52, "field_index": 1, "field_text": "4d-( 4dd-(", "gold_field": "4d-( 4dd-(", "gold_line": "8BB-\t4d-( 4dd-(", "gold_note_token": "4dd-(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 47, "line_index": 3, "line_text": "8BB-\t4d-( 4dd-(", "operation": "replace", "pred_field": "4d-( 4dd-\n", "pred_local_replacement_line": "8BB-\t4d-( 4dd-\n", "pred_note_token": "4dd-\n", "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 5.517909002904163, "LER": 34.07821273803711, "SER": 7.204861111111111}, "nll": 0.5251636505126953, "top1_error_count": 83, "top1_error_rate": 0.07198612315698179}}
+{"dataset_index": 61, "sample_id": "polish_00093", "source_ids": [], "target_token_count": 1047, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 555, "char_start": 552, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 564, "field_char_start": 552, "field_index": 1, "field_text": "8e[ 8b[ 8ee[", "gold_field": "8e[ 8b[ 8ee[", "gold_line": "8r\t8e[ 8b[ 8ee[\t8ee", "gold_note_token": "8e[", "gold_piece": "8e[", "gold_token_id": 1374, "gold_token_index": 340, "gold_trimmed_index": 340, "group": "visual", "line_char_end": 568, "line_char_start": 549, "line_index": 45, "line_text": "8r\t8e[ 8b[ 8ee[\t8ee", "operation": "replace", "pred_field": "8e( 8b[ 8ee[", "pred_local_replacement_line": "8r\t8e( 8b[ 8ee[\t8ee", "pred_note_token": "8e(", "pred_piece": "8e(", "pred_token_id": 1873, "pred_trimmed_index": 340, "visually_grounded": true}, "metrics": {"CER": 8.279124778237728, "LER": 31.38686180114746, "SER": 7.265774378585086}, "nll": 0.33346208930015564, "top1_error_count": 76, "top1_error_rate": 0.0725883476599809}}
+{"dataset_index": 62, "sample_id": "polish_00094", "source_ids": [], "target_token_count": 773, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 47, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 47, "field_char_start": 40, "field_index": 1, "field_text": "32qqggL", "gold_field": "32qqggL", "gold_line": ".\t32qqggL", "gold_note_token": "32qqggL", "gold_piece": "L", "gold_token_id": 38, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 47, "line_char_start": 38, "line_index": 2, "line_text": ".\t32qqggL", "operation": "replace", "pred_field": "32qqgg(LLL", "pred_local_replacement_line": ".\t32qqgg(LLL", "pred_note_token": "32qqgg(LLL", "pred_piece": "(LLL", "pred_token_id": 1633, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 9.864864864864865, "LER": 44.51612854003906, "SER": 12.176165803108809}, "nll": 0.784931480884552, "top1_error_count": 94, "top1_error_rate": 0.12160413971539456}}
+{"dataset_index": 63, "sample_id": "polish_00095", "source_ids": [], "target_token_count": 1073, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "4D-", "pred_local_replacement_line": "4D-\t*^", "pred_note_token": null, "pred_piece": "4D-", "pred_token_id": 556, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 17.373103087388802, "LER": 44.32989501953125, "SER": 12.779850746268655}, "nll": 0.6512929201126099, "top1_error_count": 137, "top1_error_rate": 0.1276794035414725}}
+{"dataset_index": 64, "sample_id": "polish_00096", "source_ids": [], "target_token_count": 935, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 51, "char_start": 49, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 51, "field_char_start": 49, "field_index": 1, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*^", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "structural", "line_char_end": 51, "line_char_start": 46, "line_index": 2, "line_text": "*^\t*^", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*^\t*", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 10, "visually_grounded": false}, "metrics": {"CER": 4.639175257731959, "LER": 37.14285659790039, "SER": 5.567451820128479}, "nll": 0.31392326951026917, "top1_error_count": 52, "top1_error_rate": 0.05561497326203209}}
+{"dataset_index": 65, "sample_id": "polish_00097", "source_ids": [], "target_token_count": 1050, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 7.03547805171377, "LER": 30.597015380859375, "SER": 6.3870352716873215}, "nll": 0.3556171655654907, "top1_error_count": 67, "top1_error_rate": 0.06380952380952382}}
+{"dataset_index": 66, "sample_id": "polish_00100", "source_ids": [], "target_token_count": 1185, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 36, "char_start": 34, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 36, "field_char_start": 34, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "16DD", "pred_local_replacement_line": "16DD\t*", "pred_note_token": null, "pred_piece": "16DD", "pred_token_id": 526, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 13.62932369643779, "LER": 50.79365158081055, "SER": 13.597972972972974}, "nll": 0.6852867603302002, "top1_error_count": 161, "top1_error_rate": 0.1358649789029536}}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/summary.json

@@ -0,0 +1,454 @@
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_1.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.3132269407627806,
+      "errors": 2316,
+      "errors_per_1k_tokens": 31.08099040461652,
+      "errors_per_page_mean": 34.56716417910448,
+      "errors_per_page_median": 34.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.2258588044360292,
+      "errors": 1670,
+      "errors_per_1k_tokens": 22.411594980876334,
+      "errors_per_page_mean": 24.925373134328357,
+      "errors_per_page_median": 25.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.10035163646199621,
+      "errors": 742,
+      "errors_per_1k_tokens": 9.9577266322217,
+      "errors_per_page_mean": 11.074626865671641,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05355693805788477,
+      "errors": 396,
+      "errors_per_1k_tokens": 5.314366234986244,
+      "errors_per_page_mean": 5.91044776119403,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04395455774952664,
+      "errors": 325,
+      "errors_per_1k_tokens": 4.361537945380125,
+      "errors_per_page_mean": 4.850746268656716,
+      "errors_per_page_median": 4.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.04124966188801731,
+      "errors": 305,
+      "errors_per_1k_tokens": 4.093135610279809,
+      "errors_per_page_mean": 4.552238805970149,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.8955223880597015,
+      "pages_with_error": 60
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03381119826886665,
+      "errors": 250,
+      "errors_per_1k_tokens": 3.3550291887539423,
+      "errors_per_page_mean": 3.7313432835820897,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.7910447761194029,
+      "pages_with_error": 53
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.02664322423586692,
+      "errors": 197,
+      "errors_per_1k_tokens": 2.6437630007381063,
+      "errors_per_page_mean": 2.9402985074626864,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6417910447761194,
+      "pages_with_error": 43
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.02393832837435759,
+      "errors": 177,
+      "errors_per_1k_tokens": 2.3753606656377912,
+      "errors_per_page_mean": 2.6417910447761193,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.6268656716417911,
+      "pages_with_error": 42
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.021909656478225587,
+      "errors": 162,
+      "errors_per_1k_tokens": 2.1740589143125546,
+      "errors_per_page_mean": 2.417910447761194,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5223880597014925,
+      "pages_with_error": 35
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.015958885582905057,
+      "errors": 118,
+      "errors_per_1k_tokens": 1.5835737770918608,
+      "errors_per_page_mean": 1.7611940298507462,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6417910447761194,
+      "pages_with_error": 43
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.013253989721395727,
+      "errors": 98,
+      "errors_per_1k_tokens": 1.3151714419915452,
+      "errors_per_page_mean": 1.462686567164179,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.009872869894509062,
+      "errors": 73,
+      "errors_per_1k_tokens": 0.9796685231161512,
+      "errors_per_page_mean": 1.0895522388059702,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3582089552238806,
+      "pages_with_error": 24
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.009737625101433596,
+      "errors": 72,
+      "errors_per_1k_tokens": 0.9662484063611353,
+      "errors_per_page_mean": 1.0746268656716418,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.008249932377603462,
+      "errors": 61,
+      "errors_per_1k_tokens": 0.8186271220559619,
+      "errors_per_page_mean": 0.9104477611940298,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.007167974032999729,
+      "errors": 53,
+      "errors_per_1k_tokens": 0.7112661880158357,
+      "errors_per_page_mean": 0.7910447761194029,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.29850746268656714,
+      "pages_with_error": 20
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.005815526102245064,
+      "errors": 43,
+      "errors_per_1k_tokens": 0.5770650204656781,
+      "errors_per_page_mean": 0.6417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.373134328358209,
+      "pages_with_error": 25
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005139302136867731,
+      "errors": 38,
+      "errors_per_1k_tokens": 0.5099644366905992,
+      "errors_per_page_mean": 0.5671641791044776,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.004733567757641331,
+      "errors": 35,
+      "errors_per_1k_tokens": 0.4697040864255519,
+      "errors_per_page_mean": 0.5223880597014925,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004463078171490398,
+      "errors": 33,
+      "errors_per_1k_tokens": 0.4428638529155204,
+      "errors_per_page_mean": 0.4925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.22388059701492538,
+      "pages_with_error": 15
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.004327833378414931,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.004192588585339464,
+      "errors": 31,
+      "errors_per_1k_tokens": 0.4160236194054888,
+      "errors_per_page_mean": 0.4626865671641791,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.16417910447761194,
+      "pages_with_error": 11
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0032458750338111983,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0031106302407357318,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0028401406545847984,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.002704895861509332,
+      "errors": 20,
+      "errors_per_1k_tokens": 0.26840233510031536,
+      "errors_per_page_mean": 0.29850746268656714,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.14925373134328357,
+      "pages_with_error": 10
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.002704895861509332,
+      "errors": 20,
+      "errors_per_1k_tokens": 0.26840233510031536,
+      "errors_per_page_mean": 0.29850746268656714,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "clef",
+      "error_share": 0.002434406275358399,
+      "errors": 18,
+      "errors_per_1k_tokens": 0.24156210159028382,
+      "errors_per_page_mean": 0.26865671641791045,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.002299161482282932,
+      "errors": 17,
+      "errors_per_1k_tokens": 0.22814198483526807,
+      "errors_per_page_mean": 0.2537313432835821,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.001352447930754666,
+      "errors": 10,
+      "errors_per_1k_tokens": 0.13420116755015768,
+      "errors_per_page_mean": 0.14925373134328357,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.000676223965377333,
+      "errors": 5,
+      "errors_per_1k_tokens": 0.06710058377507884,
+      "errors_per_page_mean": 0.07462686567164178,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.04477611940298507,
+      "pages_with_error": 3
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0005409791723018664,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0005409791723018664,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.0001352447930754666,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 4,
+    "accidental_and_tie_slur": 43,
+    "accidental_only": 325,
+    "articulation_only": 31,
+    "barline": 21,
+    "chord_structure_and_accidental": 5,
+    "chord_structure_and_accidental_and_articulation": 17,
+    "chord_structure_and_accidental_and_tie_slur": 20,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 73,
+    "chord_structure_and_tie_slur": 98,
+    "chord_structure_and_tie_slur_and_articulation": 10,
+    "chord_structure_only": 35,
+    "clef": 18,
+    "column_or_line_break": 396,
+    "duration_and_accidental": 33,
+    "duration_and_pitch": 118,
+    "duration_and_pitch_and_accidental": 61,
+    "duration_and_pitch_and_accidental_and_tie_slur": 4,
+    "duration_and_pitch_and_tie_slur": 20,
+    "duration_and_tie_slur": 53,
+    "duration_only": 742,
+    "interpretation_or_control": 250,
+    "note_other": 2316,
+    "null_token": 72,
+    "pitch_and_accidental": 197,
+    "pitch_and_accidental_and_tie_slur": 32,
+    "pitch_and_tie_slur": 24,
+    "pitch_only": 305,
+    "rest_note_confusion": 162,
+    "spine_operation": 177,
+    "termination": 38,
+    "tie_slur_and_articulation": 23,
+    "tie_slur_only": 1670
+  },
+  "component_error_groups": {
+    "structural": 933,
+    "visual": 6461
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 13.113788761897013,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3159,
+    "articulation": 45,
+    "barline": 20,
+    "clef": 18,
+    "column_or_line_break": 397,
+    "duration": 985,
+    "interpretation_or_control": 152,
+    "note_other": 204,
+    "null_token": 72,
+    "pitch": 24,
+    "rest": 281,
+    "spine_operation": 201,
+    "termination": 28,
+    "tie_slur": 1808
+  },
+  "gold_error_groups": {
+    "structural": 850,
+    "visual": 6544
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.535521507263184,
+      "LER": 42.10032272338867,
+      "SER": 10.14130687713623,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.0992283432865866,
+      "top1_errors": 7394,
+      "top1_errors_per_page": 110.35820895522389
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 41,
+    "predicted_unk_literal": 94
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/summary.md

@@ -0,0 +1,43 @@
+# Transcoda Dataset Eval
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_1.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Dataset kind: `hf_parquet_split`
+- Mode: `teacher_forced`
+- Pages: `67`
+- Tokens: `74,515`
+- Elapsed: `13.1s`
+
+## Metrics
+
+| mode | CER | SER | LER | exact pages | top-1 error rate | top-1 errors/page |
+|---|---:|---:|---:|---:|---:|---:|
+| `teacher_forced` | 9.535522 | 10.141307 | 42.100323 | 0.000% | 9.9228% | 110.358 |
+
+## Teacher-Forced Error Categories
+
+| category | errors |
+|---|---:|
+| `accidental` | 3,159 |
+| `tie_slur` | 1,808 |
+| `duration` | 985 |
+| `column_or_line_break` | 397 |
+| `rest` | 281 |
+| `note_other` | 204 |
+| `spine_operation` | 201 |
+| `interpretation_or_control` | 152 |
+| `null_token` | 72 |
+| `articulation` | 45 |
+| `termination` | 28 |
+| `pitch` | 24 |
+| `barline` | 20 |
+| `clef` | 18 |
+
+## Outputs
+
+- `summary.json`
+- `per_page_records.jsonl`
+- `teacher_forced_predictions.jsonl` when teacher-forced mode is enabled
+- `free_generation_predictions.jsonl` when free-generation mode is enabled
+- `gold_prefix_errors.jsonl` when teacher-forced mode is enabled
+- `teacher_forced_error_report.md` when teacher-forced mode is enabled

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_1/teacher_forced_error_report.md

@@ -0,0 +1,522 @@
+# Teacher-Forced Error Report
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_1.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Split: `eval`
+- Pages: `67`
+- Tokens: `74,515`
+- Top-1 errors: `7,394`
+- Top-1 error rate: `9.923%`
+
+## Component Buckets
+
+| bucket | errors | error share | mean/page | median/page | pages | errors/1k tok |
+|---|---:|---:|---:|---:|---:|---:|
+| `note_other` | 2,316 | 31.32% | 34.567 | 34.000 | 67 (100.0%) | 31.081 |
+| `tie_slur_only` | 1,670 | 22.59% | 24.925 | 25.000 | 67 (100.0%) | 22.412 |
+| `duration_only` | 742 | 10.04% | 11.075 | 3.000 | 61 (91.0%) | 9.958 |
+| `column_or_line_break` | 396 | 5.36% | 5.910 | 4.000 | 62 (92.5%) | 5.314 |
+| `accidental_only` | 325 | 4.40% | 4.851 | 4.000 | 61 (91.0%) | 4.362 |
+| `pitch_only` | 305 | 4.12% | 4.552 | 3.000 | 60 (89.6%) | 4.093 |
+| `interpretation_or_control` | 250 | 3.38% | 3.731 | 2.000 | 53 (79.1%) | 3.355 |
+| `pitch_and_accidental` | 197 | 2.66% | 2.940 | 1.000 | 43 (64.2%) | 2.644 |
+| `spine_operation` | 177 | 2.39% | 2.642 | 2.000 | 42 (62.7%) | 2.375 |
+| `rest_note_confusion` | 162 | 2.19% | 2.418 | 1.000 | 35 (52.2%) | 2.174 |
+| `duration_and_pitch` | 118 | 1.60% | 1.761 | 1.000 | 43 (64.2%) | 1.584 |
+| `chord_structure_and_tie_slur` | 98 | 1.33% | 1.463 | 1.000 | 41 (61.2%) | 1.315 |
+| `chord_structure_and_articulation` | 73 | 0.99% | 1.090 | 0.000 | 24 (35.8%) | 0.980 |
+| `null_token` | 72 | 0.97% | 1.075 | 0.000 | 29 (43.3%) | 0.966 |
+| `duration_and_pitch_and_accidental` | 61 | 0.82% | 0.910 | 0.000 | 29 (43.3%) | 0.819 |
+| `duration_and_tie_slur` | 53 | 0.72% | 0.791 | 0.000 | 20 (29.9%) | 0.711 |
+| `accidental_and_tie_slur` | 43 | 0.58% | 0.642 | 0.000 | 25 (37.3%) | 0.577 |
+| `termination` | 38 | 0.51% | 0.567 | 0.000 | 29 (43.3%) | 0.510 |
+| `chord_structure_only` | 35 | 0.47% | 0.522 | 0.000 | 18 (26.9%) | 0.470 |
+| `duration_and_accidental` | 33 | 0.45% | 0.493 | 0.000 | 15 (22.4%) | 0.443 |
+| `pitch_and_accidental_and_tie_slur` | 32 | 0.43% | 0.478 | 0.000 | 18 (26.9%) | 0.429 |
+| `articulation_only` | 31 | 0.42% | 0.463 | 0.000 | 11 (16.4%) | 0.416 |
+| `pitch_and_tie_slur` | 24 | 0.32% | 0.358 | 0.000 | 18 (26.9%) | 0.322 |
+| `tie_slur_and_articulation` | 23 | 0.31% | 0.343 | 0.000 | 16 (23.9%) | 0.309 |
+| `barline` | 21 | 0.28% | 0.313 | 0.000 | 16 (23.9%) | 0.282 |
+| `duration_and_pitch_and_tie_slur` | 20 | 0.27% | 0.299 | 0.000 | 10 (14.9%) | 0.268 |
+| `chord_structure_and_accidental_and_tie_slur` | 20 | 0.27% | 0.299 | 0.000 | 12 (17.9%) | 0.268 |
+| `clef` | 18 | 0.24% | 0.269 | 0.000 | 12 (17.9%) | 0.242 |
+| `chord_structure_and_accidental_and_articulation` | 17 | 0.23% | 0.254 | 0.000 | 8 (11.9%) | 0.228 |
+| `chord_structure_and_tie_slur_and_articulation` | 10 | 0.14% | 0.149 | 0.000 | 8 (11.9%) | 0.134 |
+| `chord_structure_and_accidental` | 5 | 0.07% | 0.075 | 0.000 | 3 (4.5%) | 0.067 |
+| `duration_and_pitch_and_accidental_and_tie_slur` | 4 | 0.05% | 0.060 | 0.000 | 4 (6.0%) | 0.054 |
+| `accidental_and_articulation` | 4 | 0.05% | 0.060 | 0.000 | 4 (6.0%) | 0.054 |
+| `chord_structure_and_accidental_and_tie_slur_and_articulation` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+
+## Legacy Regex Buckets
+
+These are retained for compatibility and should not be treated as component diffs.
+
+| bucket | errors |
+|---|---:|
+| `accidental` | 3,159 |
+| `tie_slur` | 1,808 |
+| `duration` | 985 |
+| `column_or_line_break` | 397 |
+| `rest` | 281 |
+| `note_other` | 204 |
+| `spine_operation` | 201 |
+| `interpretation_or_control` | 152 |
+| `null_token` | 72 |
+| `articulation` | 45 |
+| `termination` | 28 |
+| `pitch` | 24 |
+| `barline` | 20 |
+| `clef` | 18 |
+
+## Error Samples
+
+### Sample 1
+
+```text
+4d	4D 4A	16a(L	4f#
+teacher-forced top-1 local replacement:
+4d	4D 4A	16a(L	8g
+```
+
+### Sample 2
+
+```text
+8c( 8e-(L	16aa-(L
+teacher-forced top-1 local replacement:
+8c( 8e-L	16aa-(L
+```
+
+### Sample 3
+
+```text
+4BBB	8.ddW( 8.ff#W(L
+teacher-forced top-1 local replacement:
+4BBB	8.ddT( 8.ff#W(L
+```
+
+### Sample 4
+
+```text
+4FF# 4A#	4e) 4cc#)
+teacher-forced top-1 local replacement:
+4FF# 4A#	4e) 4cc#
+```
+
+### Sample 5
+
+```text
+4A	8ee( 8aa( 8ccc#(L
+teacher-forced top-1 local replacement:
+4A	8ee( 8aa( 8ccc#L(L
+```
+
+### Sample 6
+
+```text
+8c 8fJ	16cc[L	.
+teacher-forced top-1 local replacement:
+8c 8fJ	16cc[LL	.
+```
+
+### Sample 7
+
+```text
+4DD-	12d-L	4r
+teacher-forced top-1 local replacement:
+4DD-	8d-L-L	4r
+```
+
+### Sample 8
+
+```text
+8CC#	16e( 16g#( 16cc#(L
+teacher-forced top-1 local replacement:
+8FF#L	16e( 16g#( 16cc#(L
+```
+
+### Sample 9
+
+```text
+8CC 8GG 8BBL	8.dd(L	4.f
+teacher-forced top-1 local replacement:
+8CC 8GG 8BBL	16.ddLL	4.f
+```
+
+### Sample 10
+
+```text
+8CC 8FF# 8CJ	16a(L	8e-
+teacher-forced top-1 local replacement:
+8CC 8FF# 8CJ	16aLL	8e-
+```
+
+### Sample 11
+
+```text
+8F)J	8dd-)J
+teacher-forced top-1 local replacement:
+8A-	8dd-)J
+```
+
+### Sample 12
+
+```text
+4d	16ff#L
+teacher-forced top-1 local replacement:
+4d	16ff#LL
+```
+
+### Sample 13
+
+```text
+*^	*
+teacher-forced top-1 local replacement:
+16BB-LL	*
+```
+
+### Sample 14
+
+```text
+.	.	24gg
+teacher-forced top-1 local replacement:
+.	.	32gg
+```
+
+### Sample 15
+
+```text
+.	.	16bJ
+teacher-forced top-1 local replacement:
+.	.	16bJJ
+```
+
+### Sample 16
+
+```text
+.	.	16qqddJ	.
+teacher-forced top-1 local replacement:
+.	.	16qqddJJ	.
+```
+
+
+## Sample Metadata
+
+### Sample 1
+
+- Sample: `polish_00071`
+- Row/token: `49` / `1162`
+- Line/field: `153:3`
+- Component bucket: `duration_and_pitch_and_accidental`; flags: `duration, pitch, accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `4f#` -> `8g`
+- Gold note -> top-1 note: `4f#` -> `8g`
+- Gold token likelihood: rank `20`, p `0.306%`
+- Predicted token likelihood: p `26.297%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 77 | `"8g"` | 26.297% |
+| 2 | 104 | `"4g"` | 21.130% |
+| 3 | 91 | `"8r"` | 9.088% |
+| 4 | 629 | `"4g-"` | 3.559% |
+| 5 | 357 | `"8g#"` | 3.449% |
+
+### Sample 2
+
+- Sample: `polish_00084`
+- Row/token: `54` / `568`
+- Line/field: `100:0`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `8e-(L` -> `8e-L`
+- Gold note -> top-1 note: `8e-(L` -> `8e-L`
+- Gold token likelihood: rank `8`, p `0.072%`
+- Predicted token likelihood: p `98.275%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 454 | `"8e-L"` | 98.275% |
+| 2 | 199 | `"8eL"` | 0.178% |
+| 3 | 1770 | `"12e-L"` | 0.148% |
+| 4 | 265 | `"8fL"` | 0.115% |
+| 5 | 206 | `"8dL"` | 0.105% |
+
+### Sample 3
+
+- Sample: `polish_00025`
+- Row/token: `16` / `87`
+- Line/field: `15:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `W` -> `T`
+- Gold note -> top-1 note: `8.ddW(` -> `8.ddT(`
+- Gold token likelihood: rank `7`, p `1.979%`
+- Predicted token likelihood: p `27.321%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 40 | `"T"` | 27.321% |
+| 2 | 6 | `" "` | 18.778% |
+| 3 | 39 | `"M"` | 11.751% |
+| 4 | 9 | `"("` | 8.870% |
+| 5 | 67 | `"w"` | 2.971% |
+
+### Sample 4
+
+- Sample: `polish_00040`
+- Row/token: `27` / `846`
+- Line/field: `135:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `4cc#)` -> `4cc#`
+- Gold note -> top-1 note: `4cc#)` -> `4cc#`
+- Gold token likelihood: rank `2`, p `0.341%`
+- Predicted token likelihood: p `97.441%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 248 | `"4cc#"` | 97.441% |
+| 2 | 2699 | `"4cc#)"` | 0.341% |
+| 3 | 2734 | `"4cc#("` | 0.161% |
+| 4 | 731 | `"2cc#"` | 0.107% |
+| 5 | 1739 | `"4cc#["` | 0.076% |
+
+### Sample 5
+
+- Sample: `polish_00017`
+- Row/token: `10` / `594`
+- Line/field: `92:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `8ccc#` -> `8ccc#L`
+- Gold note -> top-1 note: `8ccc#(L` -> `8ccc#L(L`
+- Gold token likelihood: rank `2`, p `0.612%`
+- Predicted token likelihood: p `96.681%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1037 | `"8ccc#L"` | 96.681% |
+| 2 | 503 | `"8ccc#"` | 0.612% |
+| 3 | 543 | `"8c#L"` | 0.308% |
+| 4 | 1769 | `"8aa#L"` | 0.263% |
+| 5 | 2462 | `"8ddd#L"` | 0.218% |
+
+### Sample 6
+
+- Sample: `polish_00049`
+- Row/token: `33` / `898`
+- Line/field: `122:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `[L` -> `[LL`
+- Gold note -> top-1 note: `16cc[L` -> `16cc[LL`
+- Gold token likelihood: rank `7`, p `0.259%`
+- Predicted token likelihood: p `59.557%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 469 | `"[LL"` | 59.557% |
+| 2 | 410 | `"]LL"` | 23.323% |
+| 3 | 1330 | `"_LL"` | 11.017% |
+| 4 | 83 | `"LL"` | 0.343% |
+| 5 | 822 | `"nLL"` | 0.333% |
+
+### Sample 7
+
+- Sample: `polish_00095`
+- Row/token: `63` / `39`
+- Line/field: `6:1`
+- Component bucket: `duration_only`; flags: `duration`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `12d` -> `8d-L`
+- Gold note -> top-1 note: `12d-L` -> `8d-L-L`
+- Gold token likelihood: rank `12`, p `0.038%`
+- Predicted token likelihood: p `98.021%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 828 | `"8d-L"` | 98.021% |
+| 2 | 1293 | `"16d-LL"` | 0.228% |
+| 3 | 2328 | `"8.d-L"` | 0.134% |
+| 4 | 1911 | `"8ddd-L"` | 0.126% |
+| 5 | 448 | `"8d-"` | 0.115% |
+
+### Sample 8
+
+- Sample: `polish_00020`
+- Row/token: `12` / `886`
+- Line/field: `153:0`
+- Component bucket: `pitch_only`; flags: `pitch`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `8CC#` -> `8FF#L`
+- Gold note -> top-1 note: `8CC#` -> `8FF#L`
+- Gold token likelihood: rank `24`, p `0.085%`
+- Predicted token likelihood: p `85.272%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1031 | `"8FF#L"` | 85.272% |
+| 2 | 638 | `"8FF#"` | 1.770% |
+| 3 | 1839 | `"8CC#L"` | 1.336% |
+| 4 | 1298 | `"8GG#L"` | 1.336% |
+| 5 | 806 | `"8EEL"` | 1.255% |
+
+### Sample 9
+
+- Sample: `polish_00038`
+- Row/token: `25` / `1197`
+- Line/field: `145:1`
+- Component bucket: `duration_and_tie_slur`; flags: `duration, tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `8.dd(L` -> `16.ddLL`
+- Gold note -> top-1 note: `8.dd(L` -> `16.ddLL`
+- Gold token likelihood: rank `4`, p `3.241%`
+- Predicted token likelihood: p `57.455%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 2105 | `"16.ddLL"` | 57.455% |
+| 2 | 1062 | `"16.dd"` | 25.495% |
+| 3 | 307 | `"8.ddL"` | 3.673% |
+| 4 | 1549 | `"8.dd(L"` | 3.241% |
+| 5 | 1220 | `"8.."` | 0.820% |
+
+### Sample 10
+
+- Sample: `polish_00038`
+- Row/token: `25` / `38`
+- Line/field: `5:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16a(L` -> `16aLL`
+- Gold note -> top-1 note: `16a(L` -> `16aLL`
+- Gold token likelihood: rank `132`, p `0.005%`
+- Predicted token likelihood: p `47.218%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 338 | `"16aLL"` | 47.218% |
+| 2 | 1732 | `"32aLLL"` | 34.545% |
+| 3 | 455 | `"32a"` | 6.802% |
+| 4 | 98 | `"16a"` | 1.182% |
+| 5 | 1610 | `"24aLL"` | 1.011% |
+
+### Sample 11
+
+- Sample: `polish_00052`
+- Row/token: `35` / `447`
+- Line/field: `80:0`
+- Component bucket: `pitch_and_accidental_and_tie_slur`; flags: `pitch, accidental, tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `8F)J` -> `8A-`
+- Gold note -> top-1 note: `8F)J` -> `8A-`
+- Gold token likelihood: rank `2`, p `7.100%`
+- Predicted token likelihood: p `71.704%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 353 | `"8A-"` | 71.704% |
+| 2 | 2830 | `"8F)J"` | 7.100% |
+| 3 | 619 | `"8G-"` | 4.880% |
+| 4 | 137 | `"8F"` | 4.584% |
+| 5 | 659 | `"8A-J"` | 1.488% |
+
+### Sample 12
+
+- Sample: `polish_00089`
+- Row/token: `58` / `730`
+- Line/field: `148:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `16ff#L` -> `16ff#LL`
+- Gold note -> top-1 note: `16ff#L` -> `16ff#LL`
+- Gold token likelihood: rank `29`, p `0.025%`
+- Predicted token likelihood: p `78.394%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 473 | `"16ff#LL"` | 78.394% |
+| 2 | 1092 | `"16ff#(LL"` | 16.432% |
+| 3 | 180 | `"16ff#"` | 0.927% |
+| 4 | 173 | `"8ff#L"` | 0.311% |
+| 5 | 2635 | `"32ff#LLL"` | 0.183% |
+
+### Sample 13
+
+- Sample: `polish_00027`
+- Row/token: `17` / `1600`
+- Line/field: `230:0`
+- Component bucket: `spine_operation`; flags: `none`
+- Legacy bucket: `spine_operation`
+- Gold piece -> top-1 piece: `*^` -> `16BB-LL`
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `3`, p `14.138%`
+- Predicted token likelihood: p `23.309%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1034 | `"16BB-LL"` | 23.309% |
+| 2 | 596 | `"8BB-L"` | 19.324% |
+| 3 | 191 | `"*^"` | 14.138% |
+| 4 | 444 | `"16BB-"` | 11.011% |
+| 5 | 11 | `"*"` | 8.575% |
+
+### Sample 14
+
+- Sample: `polish_00087`
+- Row/token: `57` / `868`
+- Line/field: `117:2`
+- Component bucket: `duration_only`; flags: `duration`
+- Legacy bucket: `duration`
+- Gold piece -> top-1 piece: `24gg` -> `32gg`
+- Gold note -> top-1 note: `24gg` -> `32gg`
+- Gold token likelihood: rank `3`, p `3.473%`
+- Predicted token likelihood: p `84.133%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 477 | `"32gg"` | 84.133% |
+| 2 | 146 | `"16gg"` | 5.725% |
+| 3 | 830 | `"24gg"` | 3.473% |
+| 4 | 1767 | `"16qqgg"` | 0.825% |
+| 5 | 1178 | `"16.gg"` | 0.428% |
+
+### Sample 15
+
+- Sample: `polish_00089`
+- Row/token: `58` / `265`
+- Line/field: `40:2`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `duration`
+- Gold piece -> top-1 piece: `16bJ` -> `16bJJ`
+- Gold note -> top-1 note: `16bJ` -> `16bJJ`
+- Gold token likelihood: rank `17`, p `0.082%`
+- Predicted token likelihood: p `52.573%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 342 | `"16bJJ"` | 52.573% |
+| 2 | 125 | `"16b"` | 31.887% |
+| 3 | 1274 | `"16b)JJ"` | 3.361% |
+| 4 | 2393 | `"16b("` | 2.459% |
+| 5 | 2085 | `"16bnJJ"` | 1.491% |
+
+### Sample 16
+
+- Sample: `polish_00013`
+- Row/token: `7` / `1196`
+- Line/field: `142:2`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `note_other`
+- Gold piece -> top-1 piece: `J` -> `JJ`
+- Gold note -> top-1 note: `16qqddJ` -> `16qqddJJ`
+- Gold token likelihood: rank `5`, p `0.109%`
+- Predicted token likelihood: p `96.261%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 80 | `"JJ"` | 96.261% |
+| 2 | 1487 | `"JJJJ"` | 0.591% |
+| 3 | 2387 | `"qJJ"` | 0.116% |
+| 4 | 321 | `"JJJ"` | 0.116% |
+| 5 | 36 | `"J"` | 0.109% |

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/command.log

@@ -0,0 +1,491 @@
+Skipping import of cpp extensions due to incompatible torch version 2.11.0+cu129 for torchao version 0.15.0             Please see https://github.com/pytorch/ao/issues/2919 for more info
+[transformers] You are using a model of type `vit_mae` to instantiate a model of type `vit`. This may be expected if you are loading a checkpoint that shares a subset of the architecture (e.g., loading a `sam2_video` checkpoint into `Sam2Model`), but is otherwise not supported and can yield errors. Please verify that the checkpoint is compatible with the model you are instantiating.
+
+[transformers] ViTModel LOAD REPORT from: PRAIG/musvit
+Key                                                                     | Status     | 
+------------------------------------------------------------------------+------------+-
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.weight  | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.bias            | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.weight          | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.bias    | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.weight | UNEXPECTED | 
+decoder.decoder_pos_embed                                               | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.bias   | UNEXPECTED | 
+decoder.decoder_norm.weight                                             | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.weight          | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.weight | UNEXPECTED | 
+decoder.decoder_pred.bias                                               | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.bias            | UNEXPECTED | 
+decoder.mask_token                                                      | UNEXPECTED | 
+decoder.decoder_embed.bias                                              | UNEXPECTED | 
+decoder.decoder_norm.bias                                               | UNEXPECTED | 
+decoder.decoder_embed.weight                                            | UNEXPECTED | 
+decoder.decoder_pred.weight                                             | UNEXPECTED | 
+pooler.dense.bias                                                       | MISSING    | 
+pooler.dense.weight                                                     | MISSING    | 
+
+Notes:
+- UNEXPECTED:	can be ignored when loading from different task/architecture; not ok if you expect identical arch.
+- MISSING:	those params were newly initialized because missing from the checkpoint. Consider training on your downstream task.
+[eval] pages=32/67 mode=teacher_forced elapsed=9.4s
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_2.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.315440778799351,
+      "errors": 2333,
+      "errors_per_1k_tokens": 31.30913238945179,
+      "errors_per_page_mean": 34.82089552238806,
+      "errors_per_page_median": 34.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.22214710654407788,
+      "errors": 1643,
+      "errors_per_1k_tokens": 22.04925182849091,
+      "errors_per_page_mean": 24.52238805970149,
+      "errors_per_page_median": 24.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.10032449972958356,
+      "errors": 742,
+      "errors_per_1k_tokens": 9.9577266322217,
+      "errors_per_page_mean": 11.074626865671641,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05543537047052461,
+      "errors": 410,
+      "errors_per_1k_tokens": 5.502247869556465,
+      "errors_per_page_mean": 6.119402985074627,
+      "errors_per_page_median": 5.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04488912925905895,
+      "errors": 332,
+      "errors_per_1k_tokens": 4.455478762665235,
+      "errors_per_page_mean": 4.955223880597015,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9402985074626866,
+      "pages_with_error": 63
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.03880475932936722,
+      "errors": 287,
+      "errors_per_1k_tokens": 3.8515735086895257,
+      "errors_per_page_mean": 4.2835820895522385,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.8805970149253731,
+      "pages_with_error": 59
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03339643050297458,
+      "errors": 247,
+      "errors_per_1k_tokens": 3.3147688384888947,
+      "errors_per_page_mean": 3.6865671641791047,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.8059701492537313,
+      "pages_with_error": 54
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.02609518658734451,
+      "errors": 193,
+      "errors_per_1k_tokens": 2.5900825337180433,
+      "errors_per_page_mean": 2.8805970149253732,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5671641791044776,
+      "pages_with_error": 38
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.022850189291508925,
+      "errors": 169,
+      "errors_per_1k_tokens": 2.2679997315976648,
+      "errors_per_page_mean": 2.5223880597014925,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5373134328358209,
+      "pages_with_error": 36
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.022714981070849107,
+      "errors": 168,
+      "errors_per_1k_tokens": 2.2545796148426493,
+      "errors_per_page_mean": 2.5074626865671643,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.014872904272579772,
+      "errors": 110,
+      "errors_per_1k_tokens": 1.4762128430517345,
+      "errors_per_page_mean": 1.6417910447761195,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5970149253731343,
+      "pages_with_error": 40
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.012709572742022715,
+      "errors": 94,
+      "errors_per_1k_tokens": 1.2614909749714822,
+      "errors_per_page_mean": 1.4029850746268657,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5970149253731343,
+      "pages_with_error": 40
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.011357490535424553,
+      "errors": 84,
+      "errors_per_1k_tokens": 1.1272898074213247,
+      "errors_per_page_mean": 1.2537313432835822,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.40298507462686567,
+      "pages_with_error": 27
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.010005408328826392,
+      "errors": 74,
+      "errors_per_1k_tokens": 0.9930886398711669,
+      "errors_per_page_mean": 1.1044776119402986,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.417910447761194,
+      "pages_with_error": 28
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.008247701460248783,
+      "errors": 61,
+      "errors_per_1k_tokens": 0.8186271220559619,
+      "errors_per_page_mean": 0.9104477611940298,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.417910447761194,
+      "pages_with_error": 28
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.007842076798269334,
+      "errors": 58,
+      "errors_per_1k_tokens": 0.7783667717909146,
+      "errors_per_page_mean": 0.8656716417910447,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.34328358208955223,
+      "pages_with_error": 23
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.005813953488372093,
+      "errors": 43,
+      "errors_per_1k_tokens": 0.5770650204656781,
+      "errors_per_page_mean": 0.6417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3880597014925373,
+      "pages_with_error": 26
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005543537047052461,
+      "errors": 41,
+      "errors_per_1k_tokens": 0.5502247869556465,
+      "errors_per_page_mean": 0.6119402985074627,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.0048674959437533805,
+      "errors": 36,
+      "errors_per_1k_tokens": 0.48312420318056765,
+      "errors_per_page_mean": 0.5373134328358209,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.004597079502433748,
+      "errors": 34,
+      "errors_per_1k_tokens": 0.45628396967053614,
+      "errors_per_page_mean": 0.5074626865671642,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.19402985074626866,
+      "pages_with_error": 13
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004326663061114116,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.208955223880597,
+      "pages_with_error": 14
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.0036506219578150353,
+      "errors": 27,
+      "errors_per_1k_tokens": 0.36234315238542575,
+      "errors_per_page_mean": 0.40298507462686567,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.003515413737155219,
+      "errors": 26,
+      "errors_per_1k_tokens": 0.34892303563041,
+      "errors_per_page_mean": 0.3880597014925373,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.0029745808545159546,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0029745808545159546,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.22388059701492538,
+      "pages_with_error": 15
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0029745808545159546,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0028393726338561383,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1044776119402985,
+      "pages_with_error": 7
+    },
+    {
+      "category": "clef",
+      "error_share": 0.002568956192536506,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0024337479718766902,
+      "errors": 18,
+      "errors_per_1k_tokens": 0.24156210159028382,
+      "errors_per_page_mean": 0.26865671641791045,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.208955223880597,
+      "pages_with_error": 14
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.0014872904272579773,
+      "errors": 11,
+      "errors_per_1k_tokens": 0.14762128430517346,
+      "errors_per_page_mean": 0.16417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.13432835820895522,
+      "pages_with_error": 9
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0008112493239588967,
+      "errors": 6,
+      "errors_per_1k_tokens": 0.08052070053009461,
+      "errors_per_page_mean": 0.08955223880597014,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.08955223880597014,
+      "pages_with_error": 6
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0008112493239588967,
+      "errors": 6,
+      "errors_per_1k_tokens": 0.08052070053009461,
+      "errors_per_page_mean": 0.08955223880597014,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.07462686567164178,
+      "pages_with_error": 5
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0005408328826392645,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.00013520822065981613,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 4,
+    "accidental_and_tie_slur": 43,
+    "accidental_only": 332,
+    "articulation_only": 34,
+    "barline": 22,
+    "chord_structure_and_accidental": 6,
+    "chord_structure_and_accidental_and_articulation": 21,
+    "chord_structure_and_accidental_and_tie_slur": 22,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 84,
+    "chord_structure_and_tie_slur": 94,
+    "chord_structure_and_tie_slur_and_articulation": 11,
+    "chord_structure_only": 36,
+    "clef": 19,
+    "column_or_line_break": 410,
+    "duration_and_accidental": 32,
+    "duration_and_pitch": 110,
+    "duration_and_pitch_and_accidental": 61,
+    "duration_and_pitch_and_accidental_and_tie_slur": 6,
+    "duration_and_pitch_and_tie_slur": 27,
+    "duration_and_tie_slur": 58,
+    "duration_only": 742,
+    "interpretation_or_control": 247,
+    "note_other": 2333,
+    "null_token": 74,
+    "pitch_and_accidental": 193,
+    "pitch_and_accidental_and_tie_slur": 26,
+    "pitch_and_tie_slur": 18,
+    "pitch_only": 287,
+    "rest_note_confusion": 169,
+    "spine_operation": 168,
+    "termination": 41,
+    "tie_slur_and_articulation": 22,
+    "tie_slur_only": 1643
+  },
+  "component_error_groups": {
+    "structural": 940,
+    "visual": 6456
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 13.163790344959125,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3144,
+    "articulation": 51,
+    "barline": 21,
+    "clef": 19,
+    "column_or_line_break": 411,
+    "duration": 989,
+    "interpretation_or_control": 144,
+    "note_other": 212,
+    "null_token": 75,
+    "pitch": 20,
+    "rest": 289,
+    "spine_operation": 199,
+    "termination": 29,
+    "tie_slur": 1793
+  },
+  "gold_error_groups": {
+    "structural": 858,
+    "visual": 6538
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.481289863586426,
+      "LER": 42.202693939208984,
+      "SER": 10.14264965057373,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.09925518352009663,
+      "top1_errors": 7396,
+      "top1_errors_per_page": 110.38805970149254
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 41,
+    "predicted_unk_literal": 96
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/per_page_records.jsonl

@@ -0,0 +1,67 @@
+{"dataset_index": 0, "sample_id": "polish_00002", "source_ids": [], "target_token_count": 1725, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 59, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 55, "field_index": 0, "field_text": "2.AA[", "gold_field": "2.AA[", "gold_line": "2.AA[\t4r\t12r", "gold_note_token": "2.AA[", "gold_piece": "[", "gold_token_id": 43, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 67, "line_char_start": 55, "line_index": 4, "line_text": "2.AA[\t4r\t12r", "operation": "replace", "pred_field": "2.AA(", "pred_local_replacement_line": "2.AA(\t4r\t12r", "pred_note_token": "2.AA(", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 13.686534216335541, "LER": 67.54386138916016, "SER": 10.092807424593968}, "nll": 0.4322904646396637, "top1_error_count": 174, "top1_error_rate": 0.10086956521739131}}
+{"dataset_index": 1, "sample_id": "polish_00004", "source_ids": [], "target_token_count": 989, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 35, "field_char_start": 30, "field_index": 0, "field_text": "8AA(L", "gold_field": "8AA(L", "gold_line": "8AA(L\t4.ee(", "gold_note_token": "8AA(L", "gold_piece": "8AA", "gold_token_id": 210, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 41, "line_char_start": 30, "line_index": 2, "line_text": "8AA(L\t4.ee(", "operation": "replace", "pred_field": "=!|:(L", "pred_local_replacement_line": "=!|:(L\t4.ee(", "pred_note_token": "=!|:(L", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 6.726190476190476, "LER": 33.96226501464844, "SER": 7.388663967611336}, "nll": 0.46826496720314026, "top1_error_count": 73, "top1_error_rate": 0.07381193124368049}}
+{"dataset_index": 2, "sample_id": "polish_00005", "source_ids": [], "target_token_count": 742, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t2cc(", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 55, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t2cc(", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t2cc(", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 5.968077723802915, "LER": 39.370079040527344, "SER": 9.31174089068826}, "nll": 0.5656948089599609, "top1_error_count": 69, "top1_error_rate": 0.09299191374663072}}
+{"dataset_index": 3, "sample_id": "polish_00007", "source_ids": [], "target_token_count": 888, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 85, "char_start": 81, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 87, "field_char_start": 81, "field_index": 1, "field_text": "16dd#L", "gold_field": "16dd#L", "gold_line": "8E 8A 8cJ\t16dd#L", "gold_note_token": "16dd#L", "gold_piece": "16dd", "gold_token_id": 112, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 87, "line_char_start": 71, "line_index": 6, "line_text": "8E 8A 8cJ\t16dd#L", "operation": "replace", "pred_field": "16dd#LL#L", "pred_local_replacement_line": "8E 8A 8cJ\t16dd#LL#L", "pred_note_token": "16dd#LL#L", "pred_piece": "16dd#LL", "pred_token_id": 1164, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 7.771355170199101, "LER": 33.33333206176758, "SER": 8.455467869222097}, "nll": 0.48283064365386963, "top1_error_count": 75, "top1_error_rate": 0.08445945945945946}}
+{"dataset_index": 4, "sample_id": "polish_00008", "source_ids": [], "target_token_count": 1236, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 48, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 48, "field_char_start": 45, "field_index": 1, "field_text": "8d[", "gold_field": "8d[", "gold_line": "8r\t8d[", "gold_note_token": "8d[", "gold_piece": "8d[", "gold_token_id": 1247, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 48, "line_char_start": 42, "line_index": 3, "line_text": "8r\t8d[", "operation": "replace", "pred_field": "8d(", "pred_local_replacement_line": "8r\t8d(", "pred_note_token": "8d(", "pred_piece": "8d(", "pred_token_id": 1822, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 10.349784379492094, "LER": 48.167537689208984, "SER": 10.364372469635628}, "nll": 0.5453541278839111, "top1_error_count": 128, "top1_error_rate": 0.10355987055016182}}
+{"dataset_index": 5, "sample_id": "polish_00009", "source_ids": [], "target_token_count": 942, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 111, "char_start": 110, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 110, "field_char_start": 107, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 52, "gold_trimmed_index": 52, "group": "structural", "line_char_end": 120, "line_char_start": 107, "line_index": 11, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 52, "visually_grounded": false}, "metrics": {"CER": 5.438259756877799, "LER": 23.456790924072266, "SER": 5.738575982996812}, "nll": 0.33515357971191406, "top1_error_count": 54, "top1_error_rate": 0.05732484076433121}}
+{"dataset_index": 6, "sample_id": "polish_00011", "source_ids": [], "target_token_count": 1093, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 7.540485829959515, "LER": 41.578948974609375, "SER": 9.157509157509157}, "nll": 0.5035285353660583, "top1_error_count": 100, "top1_error_rate": 0.09149130832570906}}
+{"dataset_index": 7, "sample_id": "polish_00013", "source_ids": [], "target_token_count": 1877, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 47, "char_start": 43, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 49, "field_char_start": 43, "field_index": 1, "field_text": "16.g(L", "gold_field": "16.g(L", "gold_line": "8CC 8GG 8CL\t16.g(L\t4en", "gold_note_token": "16.g(L", "gold_piece": "16.g", "gold_token_id": 1474, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 53, "line_char_start": 31, "line_index": 3, "line_text": "8CC 8GG 8CL\t16.g(L\t4en", "operation": "replace", "pred_field": "16.a(L", "pred_local_replacement_line": "8CC 8GG 8CL\t16.a(L\t4en", "pred_note_token": "16.a(L", "pred_piece": "16.a", "pred_token_id": 1313, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 7.7532167601451665, "LER": 50.66079330444336, "SER": 8.208955223880597}, "nll": 0.4604509770870209, "top1_error_count": 154, "top1_error_rate": 0.08204581779435269}}
+{"dataset_index": 8, "sample_id": "polish_00014", "source_ids": [], "target_token_count": 1099, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t16bn(L", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 57, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t16bn(L", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t16bn(L", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 7.488766849725412, "LER": 33.7662353515625, "SER": 9.927140255009109}, "nll": 0.5350803732872009, "top1_error_count": 109, "top1_error_rate": 0.09918107370336669}}
+{"dataset_index": 9, "sample_id": "polish_00015", "source_ids": [], "target_token_count": 1086, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 38, "char_start": 37, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 38, "field_char_start": 37, "field_index": 1, "field_text": "*", "gold_field": "*", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "*^", "pred_local_replacement_line": "*^\t*^", "pred_note_token": null, "pred_piece": "*^", "pred_token_id": 191, "pred_trimmed_index": 10, "visually_grounded": false}, "metrics": {"CER": 8.503767491926803, "LER": 41.71779251098633, "SER": 10.046082949308756}, "nll": 0.47078028321266174, "top1_error_count": 109, "top1_error_rate": 0.1003683241252302}}
+{"dataset_index": 10, "sample_id": "polish_00017", "source_ids": [], "target_token_count": 1053, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 40, "char_start": 39, "component_category": "chord_structure_and_tie_slur", "component_flags": ["tie_slur", "chord_structure"], "component_group": "visual", "field_char_end": 44, "field_char_start": 34, "field_index": 0, "field_text": "4BBB- 4BB-", "gold_field": "4BBB- 4BB-", "gold_line": "4BBB- 4BB-\t4E 4cn", "gold_note_token": "4BBB-", "gold_piece": " ", "gold_token_id": 6, "gold_token_index": 9, "gold_trimmed_index": 9, "group": "visual", "line_char_end": 51, "line_char_start": 34, "line_index": 2, "line_text": "4BBB- 4BB-\t4E 4cn", "operation": "replace", "pred_field": "4BBB-(4BB-", "pred_local_replacement_line": "4BBB-(4BB-\t4E 4cn", "pred_note_token": "4BBB-(4BB-", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 9, "visually_grounded": true}, "metrics": {"CER": 2.480211081794195, "LER": 16.11111068725586, "SER": 4.562737642585551}, "nll": 0.2669675946235657, "top1_error_count": 48, "top1_error_rate": 0.045584045584045586}}
+{"dataset_index": 11, "sample_id": "polish_00019", "source_ids": [], "target_token_count": 1006, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 177, "char_start": 175, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 177, "field_char_start": 170, "field_index": 1, "field_text": "16eee)J", "gold_field": "16eee)J", "gold_line": ".\t16eee)J", "gold_note_token": "16eee)J", "gold_piece": ")J", "gold_token_id": 186, "gold_token_index": 90, "gold_trimmed_index": 90, "group": "visual", "line_char_end": 177, "line_char_start": 168, "line_index": 12, "line_text": ".\t16eee)J", "operation": "replace", "pred_field": "16eee)JJ", "pred_local_replacement_line": ".\t16eee)JJ", "pred_note_token": "16eee)JJ", "pred_piece": ")JJ", "pred_token_id": 305, "pred_trimmed_index": 90, "visually_grounded": true}, "metrics": {"CER": 5.020703933747412, "LER": 34.117645263671875, "SER": 7.761194029850746}, "nll": 0.5168823003768921, "top1_error_count": 78, "top1_error_rate": 0.0775347912524851}}
+{"dataset_index": 12, "sample_id": "polish_00020", "source_ids": [], "target_token_count": 1183, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 5.2983081032947466, "LER": 33.971290588378906, "SER": 7.4450084602368864}, "nll": 0.44750088453292847, "top1_error_count": 88, "top1_error_rate": 0.07438715131022823}}
+{"dataset_index": 13, "sample_id": "polish_00022", "source_ids": [], "target_token_count": 695, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 56, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 52, "field_index": 0, "field_text": "8G( 8B(L", "gold_field": "8G( 8B(L", "gold_line": "8G( 8B(L\t8ddd)L", "gold_note_token": "8B(L", "gold_piece": "8B(L", "gold_token_id": 2168, "gold_token_index": 20, "gold_trimmed_index": 20, "group": "visual", "line_char_end": 67, "line_char_start": 52, "line_index": 4, "line_text": "8G( 8B(L\t8ddd)L", "operation": "replace", "pred_field": "8G( 8BL", "pred_local_replacement_line": "8G( 8BL\t8ddd)L", "pred_note_token": "8BL", "pred_piece": "8BL", "pred_token_id": 396, "pred_trimmed_index": 20, "visually_grounded": true}, "metrics": {"CER": 8.04953560371517, "LER": 42.105262756347656, "SER": 10.37463976945245}, "nll": 0.6721609234809875, "top1_error_count": 72, "top1_error_rate": 0.10359712230215827}}
+{"dataset_index": 14, "sample_id": "polish_00023", "source_ids": [], "target_token_count": 1332, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 51, "field_index": 1, "field_text": "4gg)", "gold_field": "4gg)", "gold_line": "8GG 8GL\t4gg)\t8bnL", "gold_note_token": "4gg)", "gold_piece": "4gg)", "gold_token_id": 2154, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 60, "line_char_start": 43, "line_index": 3, "line_text": "8GG 8GL\t4gg)\t8bnL", "operation": "replace", "pred_field": "4gg]", "pred_local_replacement_line": "8GG 8GL\t4gg]\t8bnL", "pred_note_token": "4gg]", "pred_piece": "4gg]", "pred_token_id": 1531, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 6.977818853974122, "LER": 42.682926177978516, "SER": 7.738542449286251}, "nll": 0.406036376953125, "top1_error_count": 103, "top1_error_rate": 0.07732732732732733}}
+{"dataset_index": 15, "sample_id": "polish_00024", "source_ids": [], "target_token_count": 1075, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 7.628524046434494, "LER": 36.41618347167969, "SER": 8.752327746741155}, "nll": 0.4849332869052887, "top1_error_count": 94, "top1_error_rate": 0.08744186046511628}}
+{"dataset_index": 16, "sample_id": "polish_00025", "source_ids": [], "target_token_count": 879, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 34, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 45, "line_char_start": 34, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*M2/4", "pred_local_replacement_line": "*M2/4\t*M3/4", "pred_note_token": null, "pred_piece": "*M2/4", "pred_token_id": 304, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 6.687116564417178, "LER": 43.35664367675781, "SER": 9.79498861047836}, "nll": 0.5217281579971313, "top1_error_count": 86, "top1_error_rate": 0.09783845278725825}}
+{"dataset_index": 17, "sample_id": "polish_00027", "source_ids": [], "target_token_count": 1825, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 46, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 46, "field_index": 0, "field_text": "8BBB- 8BB-", "gold_field": "8BBB- 8BB-", "gold_line": "8BBB- 8BB-\t8b- 8dd", "gold_note_token": "8BBB-", "gold_piece": "8BBB-", "gold_token_id": 632, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 3, "line_text": "8BBB- 8BB-\t8b- 8dd", "operation": "replace", "pred_field": "8AAA 8BB-", "pred_local_replacement_line": "8AAA 8BB-\t8b- 8dd", "pred_note_token": "8AAA", "pred_piece": "8AAA", "pred_token_id": 490, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 8.360034954849986, "LER": 49.612403869628906, "SER": 9.429824561403509}, "nll": 0.4881656765937805, "top1_error_count": 172, "top1_error_rate": 0.09424657534246575}}
+{"dataset_index": 18, "sample_id": "polish_00029", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 39, "field_char_start": 30, "field_index": 0, "field_text": "8CC( 8C(L", "gold_field": "8CC( 8C(L", "gold_line": "8CC( 8C(L\t4gg([", "gold_note_token": "8CC(", "gold_piece": "8CC", "gold_token_id": 364, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 45, "line_char_start": 30, "line_index": 2, "line_text": "8CC( 8C(L\t4gg([", "operation": "replace", "pred_field": "=!|:( 8C(L", "pred_local_replacement_line": "=!|:( 8C(L\t4gg([", "pred_note_token": "=!|:(", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 8.00513698630137, "LER": 36.17021179199219, "SER": 10.136765888978278}, "nll": 0.5880449414253235, "top1_error_count": 126, "top1_error_rate": 0.10128617363344052}}
+{"dataset_index": 19, "sample_id": "polish_00030", "source_ids": [], "target_token_count": 1263, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "=!|:", "pred_local_replacement_line": "=!|:\t*^", "pred_note_token": null, "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 12.159038372630606, "LER": 54.41860580444336, "SER": 10.45958795562599}, "nll": 0.45838662981987, "top1_error_count": 132, "top1_error_rate": 0.10451306413301663}}
+{"dataset_index": 20, "sample_id": "polish_00031", "source_ids": [], "target_token_count": 862, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 34, "field_index": 0, "field_text": "*M3/8", "gold_field": "*M3/8", "gold_line": "*M3/8\t*M3/8", "gold_note_token": null, "gold_piece": "*M3/8", "gold_token_id": 592, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 45, "line_char_start": 34, "line_index": 2, "line_text": "*M3/8\t*M3/8", "operation": "replace", "pred_field": "*M3/4", "pred_local_replacement_line": "*M3/4\t*M3/8", "pred_note_token": null, "pred_piece": "*M3/4", "pred_token_id": 283, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 9.58904109589041, "LER": 41.25873947143555, "SER": 9.639953542392567}, "nll": 0.5470021963119507, "top1_error_count": 83, "top1_error_rate": 0.0962877030162413}}
+{"dataset_index": 21, "sample_id": "polish_00032", "source_ids": [], "target_token_count": 1387, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 50, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 52, "field_char_start": 45, "field_index": 2, "field_text": "16ff#]L", "gold_field": "16ff#]L", "gold_line": "4G#(\t8G#L\t16ff#]L", "gold_note_token": "16ff#]L", "gold_piece": "16ff#", "gold_token_id": 180, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 52, "line_char_start": 35, "line_index": 3, "line_text": "4G#(\t8G#L\t16ff#]L", "operation": "replace", "pred_field": "16ff#)]L", "pred_local_replacement_line": "4G#(\t8G#L\t16ff#)]L", "pred_note_token": "16ff#)]L", "pred_piece": "16ff#)", "pred_token_id": 2192, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 10.981210855949895, "LER": 44.53781509399414, "SER": 11.976911976911978}, "nll": 0.6978175044059753, "top1_error_count": 166, "top1_error_rate": 0.11968276856524873}}
+{"dataset_index": 22, "sample_id": "polish_00034", "source_ids": [], "target_token_count": 788, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 45, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8f( 8a(L", "gold_field": "8f( 8a(L", "gold_line": "8f( 8a(L\t16r", "gold_note_token": null, "gold_piece": "8f(", "gold_token_id": 2288, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 42, "line_index": 3, "line_text": "8f( 8a(L\t16r", "operation": "replace", "pred_field": "* 8a(L", "pred_local_replacement_line": "* 8a(L\t16r", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 23.70088719898606, "LER": 57.76397705078125, "SER": 38.75476493011436}, "nll": 1.4377975463867188, "top1_error_count": 177, "top1_error_rate": 0.22461928934010153}}
+{"dataset_index": 23, "sample_id": "polish_00036", "source_ids": [], "target_token_count": 1317, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 451, "char_start": 447, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 451, "field_char_start": 447, "field_index": 2, "field_text": "16gJ", "gold_field": "16gJ", "gold_line": ".\t.\t16gJ", "gold_note_token": "16gJ", "gold_piece": "16gJ", "gold_token_id": 859, "gold_token_index": 280, "gold_trimmed_index": 280, "group": "visual", "line_char_end": 451, "line_char_start": 443, "line_index": 41, "line_text": ".\t.\t16gJ", "operation": "replace", "pred_field": "16gJJ", "pred_local_replacement_line": ".\t.\t16gJJ", "pred_note_token": "16gJJ", "pred_piece": "16gJJ", "pred_token_id": 351, "pred_trimmed_index": 280, "visually_grounded": true}, "metrics": {"CER": 5.658669081032141, "LER": 22.91666603088379, "SER": 4.939209726443769}, "nll": 0.299640029668808, "top1_error_count": 65, "top1_error_rate": 0.04935459377372817}}
+{"dataset_index": 24, "sample_id": "polish_00037", "source_ids": [], "target_token_count": 871, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 49, "char_start": 46, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 46, "field_index": 0, "field_text": "4Fn 4dn", "gold_field": "4Fn 4dn", "gold_line": "4Fn 4dn\t2f#] 2ff#]", "gold_note_token": "4Fn", "gold_piece": "4Fn", "gold_token_id": 1918, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 2, "line_text": "4Fn 4dn\t2f#] 2ff#]", "operation": "replace", "pred_field": "4F# 4dn", "pred_local_replacement_line": "4F# 4dn\t2f#] 2ff#]", "pred_note_token": "4F#", "pred_piece": "4F#", "pred_token_id": 278, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 4.381948986265534, "LER": 24.161073684692383, "SER": 6.206896551724138}, "nll": 0.27900344133377075, "top1_error_count": 54, "top1_error_rate": 0.06199770378874857}}
+{"dataset_index": 25, "sample_id": "polish_00038", "source_ids": [], "target_token_count": 1471, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 88, "char_start": 83, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 88, "field_char_start": 83, "field_index": 1, "field_text": "16a(L", "gold_field": "16a(L", "gold_line": "8CC 8FF# 8CJ\t16a(L\t8e-", "gold_note_token": "16a(L", "gold_piece": "16a(L", "gold_token_id": 2253, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 92, "line_char_start": 70, "line_index": 5, "line_text": "8CC 8FF# 8CJ\t16a(L\t8e-", "operation": "replace", "pred_field": "16aLL", "pred_local_replacement_line": "8CC 8FF# 8CJ\t16aLL\t8e-", "pred_note_token": "16aLL", "pred_piece": "16aLL", "pred_token_id": 338, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 6.514788169464429, "LER": 38.5474853515625, "SER": 7.414965986394558}, "nll": 0.417838454246521, "top1_error_count": 109, "top1_error_rate": 0.07409925220938138}}
+{"dataset_index": 26, "sample_id": "polish_00039", "source_ids": [], "target_token_count": 1503, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 93, "char_start": 88, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 93, "field_char_start": 88, "field_index": 0, "field_text": "16A-L", "gold_field": "16A-L", "gold_line": "16A-L\t2.AA-[\t16cL\t8ee-", "gold_note_token": "16A-L", "gold_piece": "16A-L", "gold_token_id": 2834, "gold_token_index": 24, "gold_trimmed_index": 24, "group": "visual", "line_char_end": 110, "line_char_start": 88, "line_index": 4, "line_text": "16A-L\t2.AA-[\t16cL\t8ee-", "operation": "replace", "pred_field": "16A-LL", "pred_local_replacement_line": "16A-LL\t2.AA-[\t16cL\t8ee-", "pred_note_token": "16A-LL", "pred_piece": "16A-LL", "pred_token_id": 1109, "pred_trimmed_index": 24, "visually_grounded": true}, "metrics": {"CER": 8.639562157935888, "LER": 39.153438568115234, "SER": 8.921438082556591}, "nll": 0.5169796347618103, "top1_error_count": 134, "top1_error_rate": 0.08915502328675981}}
+{"dataset_index": 27, "sample_id": "polish_00040", "source_ids": [], "target_token_count": 1109, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 8.448616600790514, "LER": 43.75, "SER": 10.379061371841155}, "nll": 0.5251513123512268, "top1_error_count": 115, "top1_error_rate": 0.1036970243462579}}
+{"dataset_index": 28, "sample_id": "polish_00043", "source_ids": [], "target_token_count": 885, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "replace", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 13.676212741087085, "LER": 55.491329193115234, "SER": 17.420814479638008}, "nll": 0.9326702952384949, "top1_error_count": 154, "top1_error_rate": 0.17401129943502824}}
+{"dataset_index": 29, "sample_id": "polish_00044", "source_ids": [], "target_token_count": 677, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 74, "char_start": 73, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 82, "field_char_start": 70, "field_index": 1, "field_text": "16f( 16dd(Jk", "gold_field": "16f( 16dd(Jk", "gold_line": "16B-(Jk\t16f( 16dd(Jk", "gold_note_token": "16f(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 22, "gold_trimmed_index": 22, "group": "visual", "line_char_end": 82, "line_char_start": 62, "line_index": 4, "line_text": "16B-(Jk\t16f( 16dd(Jk", "operation": "delete", "pred_field": "16f  16dd(Jk", "pred_local_replacement_line": "16B-(Jk\t16f  16dd(Jk", "pred_note_token": "16f", "pred_piece": " ", "pred_token_id": 6, "pred_trimmed_index": 22, "visually_grounded": true}, "metrics": {"CER": 8.533747090768037, "LER": 47.93388366699219, "SER": 10.502958579881657}, "nll": 0.6429754495620728, "top1_error_count": 71, "top1_error_rate": 0.10487444608567208}}
+{"dataset_index": 30, "sample_id": "polish_00046", "source_ids": [], "target_token_count": 1069, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "clef", "char_end": 15, "char_start": 8, "component_category": "clef", "component_flags": [], "component_group": "visual", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2", "gold_note_token": null, "gold_piece": "*clefG2", "gold_token_id": 147, "gold_token_index": 2, "gold_trimmed_index": 2, "group": "visual", "line_char_end": 15, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2", "operation": "replace", "pred_field": "*clefG", "pred_local_replacement_line": "*clefF4\t*clefG", "pred_note_token": null, "pred_piece": "*clefG", "pred_token_id": 145, "pred_trimmed_index": 2, "visually_grounded": true}, "metrics": {"CER": 8.423772609819121, "LER": 38.56209182739258, "SER": 8.98876404494382}, "nll": 0.361417293548584, "top1_error_count": 96, "top1_error_rate": 0.08980355472404115}}
+{"dataset_index": 31, "sample_id": "polish_00047", "source_ids": [], "target_token_count": 1243, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 66, "char_start": 63, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 68, "field_char_start": 63, "field_index": 1, "field_text": "8eeq(", "gold_field": "8eeq(", "gold_line": ".\t8eeq(\t.", "gold_note_token": "8eeq(", "gold_piece": "8ee", "gold_token_id": 84, "gold_token_index": 28, "gold_trimmed_index": 28, "group": "visual", "line_char_end": 70, "line_char_start": 61, "line_index": 4, "line_text": ".\t8eeq(\t.", "operation": "replace", "pred_field": "8ddq(", "pred_local_replacement_line": ".\t8ddq(\t.", "pred_note_token": "8ddq(", "pred_piece": "8dd", "pred_token_id": 79, "pred_trimmed_index": 28, "visually_grounded": true}, "metrics": {"CER": 5.038387715930901, "LER": 26.829267501831055, "SER": 5.072463768115942}, "nll": 0.3277817666530609, "top1_error_count": 63, "top1_error_rate": 0.050683829444891394}}
+{"dataset_index": 32, "sample_id": "polish_00048", "source_ids": [], "target_token_count": 1384, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 51, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 48, "field_index": 0, "field_text": "4c 4f", "gold_field": "4c 4f", "gold_line": "4c 4f\t12aa-(L", "gold_note_token": "4f", "gold_piece": "4f", "gold_token_id": 103, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 48, "line_index": 3, "line_text": "4c 4f\t12aa-(L", "operation": "replace", "pred_field": "4c 4a-", "pred_local_replacement_line": "4c 4a-\t12aa-(L", "pred_note_token": "4a-", "pred_piece": "4a-", "pred_token_id": 341, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 14.789849325931801, "LER": 45.6928825378418, "SER": 10.484454085321765}, "nll": 0.5281092524528503, "top1_error_count": 145, "top1_error_rate": 0.10476878612716763}}
+{"dataset_index": 33, "sample_id": "polish_00049", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 66, "char_start": 65, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 70, "field_char_start": 65, "field_index": 1, "field_text": "48bbL", "gold_field": "48bbL", "gold_line": ".\t48bbL\t.", "gold_note_token": "48bbL", "gold_piece": "4", "gold_token_id": 20, "gold_token_index": 28, "gold_trimmed_index": 28, "group": "visual", "line_char_end": 72, "line_char_start": 63, "line_index": 4, "line_text": ".\t48bbL\t.", "operation": "replace", "pred_field": "32bbLLL8bbL", "pred_local_replacement_line": ".\t32bbLLL8bbL\t.", "pred_note_token": "32bbLLL8bbL", "pred_piece": "32bbLLL", "pred_token_id": 2581, "pred_trimmed_index": 28, "visually_grounded": true}, "metrics": {"CER": 13.113113113113112, "LER": 42.77108383178711, "SER": 12.148028962188254}, "nll": 0.5234226584434509, "top1_error_count": 125, "top1_error_rate": 0.10048231511254019}}
+{"dataset_index": 34, "sample_id": "polish_00050", "source_ids": [], "target_token_count": 1242, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "replace", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 11.9368323269856, "LER": 54.01069641113281, "SER": 11.522965350523771}, "nll": 0.5523250102996826, "top1_error_count": 143, "top1_error_rate": 0.11513687600644122}}
+{"dataset_index": 35, "sample_id": "polish_00052", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 52, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 59, "field_char_start": 47, "field_index": 0, "field_text": "2.DD-( 2.D-(", "gold_field": "2.DD-( 2.D-(", "gold_line": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "gold_note_token": "2.DD-(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 74, "line_char_start": 47, "line_index": 3, "line_text": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "operation": "replace", "pred_field": "2.DD-([ 2.D-(", "pred_local_replacement_line": "2.DD-([ 2.D-(\t8ff(L\t2.f 2.a-", "pred_note_token": "2.DD-([", "pred_piece": "([", "pred_token_id": 2909, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 10.48158640226629, "LER": 45.35519027709961, "SER": 13.580246913580247}, "nll": 0.7448434829711914, "top1_error_count": 132, "top1_error_rate": 0.13566289825282632}}
+{"dataset_index": 36, "sample_id": "polish_00053", "source_ids": [], "target_token_count": 1018, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 48, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 62, "field_char_start": 48, "field_index": 0, "field_text": "4FFn: 4Cn: 4A:", "gold_field": "4FFn: 4Cn: 4A:", "gold_line": "4FFn: 4Cn: 4A:\t8fn", "gold_note_token": "4FFn:", "gold_piece": "4FF", "gold_token_id": 369, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 66, "line_char_start": 48, "line_index": 4, "line_text": "4FFn: 4Cn: 4A:\t8fn", "operation": "replace", "pred_field": "4FF#n: 4Cn: 4A:", "pred_local_replacement_line": "4FF#n: 4Cn: 4A:\t8fn", "pred_note_token": "4FF#n:", "pred_piece": "4FF#", "pred_token_id": 607, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 3.9809863339275102, "LER": 17.127071380615234, "SER": 4.523107177974435}, "nll": 0.27179908752441406, "top1_error_count": 46, "top1_error_rate": 0.04518664047151277}}
+{"dataset_index": 37, "sample_id": "polish_00054", "source_ids": [], "target_token_count": 689, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 46, "field_index": 0, "field_text": "8B-( 8d(L", "gold_field": "8B-( 8d(L", "gold_line": "8B-( 8d(L\t8fL", "gold_note_token": "8d(L", "gold_piece": "8d(L", "gold_token_id": 1103, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 59, "line_char_start": 46, "line_index": 3, "line_text": "8B-( 8d(L\t8fL", "operation": "replace", "pred_field": "8B-( 8dL", "pred_local_replacement_line": "8B-( 8dL\t8fL", "pred_note_token": "8dL", "pred_piece": "8dL", "pred_token_id": 206, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 9.432515337423313, "LER": 40.44117736816406, "SER": 11.337209302325581}, "nll": 0.7641922831535339, "top1_error_count": 78, "top1_error_rate": 0.11320754716981132}}
+{"dataset_index": 38, "sample_id": "polish_00056", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 28, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 28, "field_char_start": 16, "field_index": 0, "field_text": "*k[b-e-a-d-]", "gold_field": "*k[b-e-a-d-]", "gold_line": "*k[b-e-a-d-]\t*k[b-e-a-d-]", "gold_note_token": null, "gold_piece": "*k[b-e-a-d-]", "gold_token_id": 637, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 41, "line_char_start": 16, "line_index": 1, "line_text": "*k[b-e-a-d-]\t*k[b-e-a-d-]", "operation": "replace", "pred_field": "*k[b-e-a-d-g-c-]", "pred_local_replacement_line": "*k[b-e-a-d-g-c-]\t*k[b-e-a-d-]", "pred_note_token": null, "pred_piece": "*k[b-e-a-d-g-c-]", "pred_token_id": 827, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 14.13870246085011, "LER": 47.004608154296875, "SER": 13.967971530249109}, "nll": 0.5589922070503235, "top1_error_count": 157, "top1_error_rate": 0.13955555555555554}}
+{"dataset_index": 39, "sample_id": "polish_00057", "source_ids": [], "target_token_count": 733, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 52, "component_category": "duration_and_pitch_and_accidental", "component_flags": ["duration", "pitch", "accidental"], "component_group": "visual", "field_char_end": 57, "field_char_start": 52, "field_index": 1, "field_text": "8ff#q", "gold_field": "8ff#q", "gold_line": ".\t8ff#q", "gold_note_token": "8ff#q", "gold_piece": "8ff#", "gold_token_id": 92, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": ".\t8ff#q", "operation": "replace", "pred_field": "ee-q", "pred_local_replacement_line": ".\tee-q", "pred_note_token": "ee-q", "pred_piece": "ee-", "pred_token_id": 1672, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 12.972972972972974, "LER": 45.45454406738281, "SER": 15.710382513661203}, "nll": 0.8955805897712708, "top1_error_count": 115, "top1_error_rate": 0.15688949522510232}}
+{"dataset_index": 40, "sample_id": "polish_00058", "source_ids": [], "target_token_count": 900, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "rest", "char_end": 178, "char_start": 175, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 178, "field_char_start": 175, "field_index": 1, "field_text": "16r", "gold_field": "16r", "gold_line": "8D 8F 8A 8dL\t16r", "gold_note_token": "16r", "gold_piece": "16r", "gold_token_id": 168, "gold_token_index": 72, "gold_trimmed_index": 72, "group": "visual", "line_char_end": 178, "line_char_start": 162, "line_index": 12, "line_text": "8D 8F 8A 8dL\t16r", "operation": "replace", "pred_field": "8r", "pred_local_replacement_line": "8D 8F 8A 8dL\t8r", "pred_note_token": "8r", "pred_piece": "8r", "pred_token_id": 91, "pred_trimmed_index": 72, "visually_grounded": true}, "metrics": {"CER": 7.477203647416414, "LER": 33.57143020629883, "SER": 8.231368186874304}, "nll": 0.4938296973705292, "top1_error_count": 74, "top1_error_rate": 0.08222222222222222}}
+{"dataset_index": 41, "sample_id": "polish_00059", "source_ids": [], "target_token_count": 1838, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 60, "char_start": 56, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 60, "field_char_start": 56, "field_index": 1, "field_text": "8AAL", "gold_field": "8AAL", "gold_line": "4.AA\t8AAL\t24aa", "gold_note_token": "8AAL", "gold_piece": "8AAL", "gold_token_id": 512, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 65, "line_char_start": 51, "line_index": 4, "line_text": "4.AA\t8AAL\t24aa", "operation": "replace", "pred_field": "8.AAL", "pred_local_replacement_line": "4.AA\t8.AAL\t24aa", "pred_note_token": "8.AAL", "pred_piece": "8.AAL", "pred_token_id": 2335, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 12.428870919436957, "LER": 65.25096893310547, "SER": 12.847033206314643}, "nll": 0.7022934556007385, "top1_error_count": 236, "top1_error_rate": 0.12840043525571274}}
+{"dataset_index": 42, "sample_id": "polish_00060", "source_ids": [], "target_token_count": 1528, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 62, "char_start": 59, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 66, "field_char_start": 59, "field_index": 0, "field_text": "8FF 8FL", "gold_field": "8FF 8FL", "gold_line": "8FF 8FL\t2.F\t4.f", "gold_note_token": "8FF", "gold_piece": "8FF", "gold_token_id": 276, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 74, "line_char_start": 59, "line_index": 4, "line_text": "8FF 8FL\t2.F\t4.f", "operation": "replace", "pred_field": "8FL 8FL", "pred_local_replacement_line": "8FL 8FL\t2.F\t4.f", "pred_note_token": "8FL", "pred_piece": "8FL", "pred_token_id": 433, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 6.240981240981241, "LER": 43.92523193359375, "SER": 8.055009823182711}, "nll": 0.5008909106254578, "top1_error_count": 123, "top1_error_rate": 0.08049738219895287}}
+{"dataset_index": 43, "sample_id": "polish_00061", "source_ids": [], "target_token_count": 653, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 50, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 59, "field_char_start": 50, "field_index": 0, "field_text": "8e-( 8g(L", "gold_field": "8e-( 8g(L", "gold_line": "8e-( 8g(L\t8gg 8bb-L", "gold_note_token": null, "gold_piece": "8e-", "gold_token_id": 263, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 69, "line_char_start": 50, "line_index": 3, "line_text": "8e-( 8g(L\t8gg 8bb-L", "operation": "replace", "pred_field": "*( 8g(L", "pred_local_replacement_line": "*( 8g(L\t8gg 8bb-L", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 11.383928571428571, "LER": 46.93877410888672, "SER": 15.950920245398773}, "nll": 0.7654175758361816, "top1_error_count": 104, "top1_error_rate": 0.15926493108728942}}
+{"dataset_index": 44, "sample_id": "polish_00062", "source_ids": [], "target_token_count": 781, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 87, "char_start": 83, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 87, "field_char_start": 83, "field_index": 1, "field_text": "8eeL", "gold_field": "8eeL", "gold_line": "8CL\t8eeL\t.", "gold_note_token": "8eeL", "gold_piece": "8eeL", "gold_token_id": 135, "gold_token_index": 34, "gold_trimmed_index": 34, "group": "visual", "line_char_end": 89, "line_char_start": 79, "line_index": 5, "line_text": "8CL\t8eeL\t.", "operation": "replace", "pred_field": "8ee", "pred_local_replacement_line": "8CL\t8ee\t.", "pred_note_token": "8ee", "pred_piece": "8ee", "pred_token_id": 84, "pred_trimmed_index": 34, "visually_grounded": true}, "metrics": {"CER": 10.130718954248366, "LER": 37.3831787109375, "SER": 8.717948717948717}, "nll": 0.45523568987846375, "top1_error_count": 68, "top1_error_rate": 0.08706786171574904}}
+{"dataset_index": 45, "sample_id": "polish_00063", "source_ids": [], "target_token_count": 963, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 54, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 48, "field_index": 1, "field_text": "16qqddnL", "gold_field": "16qqddnL", "gold_line": ".\t16qqddnL", "gold_note_token": "16qqddnL", "gold_piece": "nL", "gold_token_id": 403, "gold_token_index": 11, "gold_trimmed_index": 11, "group": "visual", "line_char_end": 56, "line_char_start": 46, "line_index": 2, "line_text": ".\t16qqddnL", "operation": "replace", "pred_field": "16qqdd#LL", "pred_local_replacement_line": ".\t16qqdd#LL", "pred_note_token": "16qqdd#LL", "pred_piece": "#LL", "pred_token_id": 269, "pred_trimmed_index": 11, "visually_grounded": true}, "metrics": {"CER": 4.174340085942296, "LER": 28.289474487304688, "SER": 6.237006237006238}, "nll": 0.3493999242782593, "top1_error_count": 60, "top1_error_rate": 0.06230529595015576}}
+{"dataset_index": 46, "sample_id": "polish_00064", "source_ids": [], "target_token_count": 928, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 53, "char_start": 49, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 53, "field_char_start": 49, "field_index": 1, "field_text": "8dd[", "gold_field": "8dd[", "gold_line": "8r\t8dd[", "gold_note_token": "8dd[", "gold_piece": "8dd[", "gold_token_id": 1303, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 53, "line_char_start": 46, "line_index": 3, "line_text": "8r\t8dd[", "operation": "replace", "pred_field": "8dd(", "pred_local_replacement_line": "8r\t8dd(", "pred_note_token": "8dd(", "pred_piece": "8dd(", "pred_token_id": 1124, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 9.433962264150944, "LER": 45.19773864746094, "SER": 13.160733549083064}, "nll": 0.7758469581604004, "top1_error_count": 122, "top1_error_rate": 0.1314655172413793}}
+{"dataset_index": 47, "sample_id": "polish_00067", "source_ids": [], "target_token_count": 1138, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 43, "char_start": 39, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 43, "field_char_start": 39, "field_index": 1, "field_text": "16aL", "gold_field": "16aL", "gold_line": "8F(L\t16aL", "gold_note_token": "16aL", "gold_piece": "16aL", "gold_token_id": 602, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 43, "line_char_start": 34, "line_index": 2, "line_text": "8F(L\t16aL", "operation": "replace", "pred_field": "16aLL", "pred_local_replacement_line": "8F(L\t16aLL", "pred_note_token": "16aLL", "pred_piece": "16aLL", "pred_token_id": 338, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 12.116925978312116, "LER": 46.35416793823242, "SER": 14.951627088830255}, "nll": 0.9147043824195862, "top1_error_count": 169, "top1_error_rate": 0.14850615114235502}}
+{"dataset_index": 48, "sample_id": "polish_00068", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 57, "char_start": 53, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 57, "field_char_start": 53, "field_index": 1, "field_text": "4ff#", "gold_field": "4ff#", "gold_line": "4r\t4ff#", "gold_note_token": "4ff#", "gold_piece": "4ff#", "gold_token_id": 225, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": "4r\t4ff#", "operation": "replace", "pred_field": "4ff#(", "pred_local_replacement_line": "4r\t4ff#(", "pred_note_token": "4ff#(", "pred_piece": "4ff#(", "pred_token_id": 2286, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 21.91435768261965, "LER": 64.64088439941406, "SER": 20.061728395061728}, "nll": 1.007832407951355, "top1_error_count": 195, "top1_error_rate": 0.20041109969167523}}
+{"dataset_index": 49, "sample_id": "polish_00071", "source_ids": [], "target_token_count": 1433, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 48, "char_start": 46, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 48, "field_char_start": 46, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*^", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 51, "line_char_start": 46, "line_index": 2, "line_text": "*^\t*^", "operation": "replace", "pred_field": ".", "pred_local_replacement_line": ".\t*^", "pred_note_token": null, "pred_piece": ".", "pred_token_id": 14, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 11.878685762426285, "LER": 62.295082092285156, "SER": 14.03631284916201}, "nll": 0.7100946307182312, "top1_error_count": 201, "top1_error_rate": 0.14026517794836008}}
+{"dataset_index": 50, "sample_id": "polish_00072", "source_ids": [], "target_token_count": 1123, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 56, "char_start": 55, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 13, "gold_trimmed_index": 13, "group": "structural", "line_char_end": 65, "line_char_start": 52, "line_index": 3, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 13, "visually_grounded": false}, "metrics": {"CER": 4.43772062531518, "LER": 29.787233352661133, "SER": 6.327985739750446}, "nll": 0.29230329394340515, "top1_error_count": 71, "top1_error_rate": 0.06322350845948353}}
+{"dataset_index": 51, "sample_id": "polish_00074", "source_ids": [], "target_token_count": 1355, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 34, "char_start": 30, "component_category": "duration_and_pitch_and_accidental", "component_flags": ["duration", "pitch", "accidental"], "component_group": "visual", "field_char_end": 36, "field_char_start": 30, "field_index": 0, "field_text": "8BBnq(", "gold_field": "8BBnq(", "gold_line": "8BBnq(\t.", "gold_note_token": "8BBnq(", "gold_piece": "8BBn", "gold_token_id": 2042, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 38, "line_char_start": 30, "line_index": 2, "line_text": "8BBnq(\t.", "operation": "replace", "pred_field": "=!|:q(", "pred_local_replacement_line": "=!|:q(\t.", "pred_note_token": "=!|:q(", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 8.788282290279627, "LER": 42.56410217285156, "SER": 9.67503692762186}, "nll": 0.464192658662796, "top1_error_count": 131, "top1_error_rate": 0.09667896678966789}}
+{"dataset_index": 52, "sample_id": "polish_00078", "source_ids": [], "target_token_count": 1172, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 49, "char_start": 48, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 49, "field_char_start": 48, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*\t*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 56, "line_char_start": 48, "line_index": 3, "line_text": "*\t*\t*^\t*", "operation": "insert", "pred_field": "4Dn", "pred_local_replacement_line": "4Dn\t*\t*^\t*", "pred_note_token": null, "pred_piece": "4Dn", "pred_token_id": 2294, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 16.31888710540396, "LER": 45.45454406738281, "SER": 12.724167378309136}, "nll": 0.5516060590744019, "top1_error_count": 149, "top1_error_rate": 0.12713310580204779}}
+{"dataset_index": 53, "sample_id": "polish_00082", "source_ids": [], "target_token_count": 1027, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "8AA-L", "pred_local_replacement_line": "8AA-L\t*^", "pred_note_token": null, "pred_piece": "8AA-L", "pred_token_id": 901, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 14.34108527131783, "LER": 51.31578826904297, "SER": 10.721247563352826}, "nll": 0.566083550453186, "top1_error_count": 110, "top1_error_rate": 0.10710808179162609}}
+{"dataset_index": 54, "sample_id": "polish_00084", "source_ids": [], "target_token_count": 860, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 54, "char_start": 50, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 57, "field_char_start": 50, "field_index": 0, "field_text": "32qqCCL", "gold_field": "32qqCCL", "gold_line": "32qqCCL\t32qqe-L", "gold_note_token": "32qqCCL", "gold_piece": "32qq", "gold_token_id": 1235, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 65, "line_char_start": 50, "line_index": 3, "line_text": "32qqCCL\t32qqe-L", "operation": "replace", "pred_field": "32CCCCL", "pred_local_replacement_line": "32CCCCL\t32qqe-L", "pred_note_token": "32CCCCL", "pred_piece": "32CC", "pred_token_id": 1559, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 11.097560975609756, "LER": 43.312103271484375, "SER": 13.271245634458673}, "nll": 0.7630321383476257, "top1_error_count": 114, "top1_error_rate": 0.1325581395348837}}
+{"dataset_index": 55, "sample_id": "polish_00085", "source_ids": [], "target_token_count": 1428, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 54, "char_start": 49, "component_category": "duration_and_tie_slur", "component_flags": ["duration", "tie_slur"], "component_group": "visual", "field_char_end": 54, "field_char_start": 49, "field_index": 2, "field_text": "8aa(L", "gold_field": "8aa(L", "gold_line": "4d\t4D 4A\t8aa(L\t4f# 4dd", "gold_note_token": "8aa(L", "gold_piece": "8aa(L", "gold_token_id": 1057, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 40, "line_index": 3, "line_text": "4d\t4D 4A\t8aa(L\t4f# 4dd", "operation": "replace", "pred_field": "8.aa", "pred_local_replacement_line": "4d\t4D 4A\t8.aa\t4f# 4dd", "pred_note_token": "8.aa", "pred_piece": "8.aa", "pred_token_id": 386, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 5.010893246187364, "LER": 25.465839385986328, "SER": 4.344779257182902}, "nll": 0.23825819790363312, "top1_error_count": 62, "top1_error_rate": 0.04341736694677871}}
+{"dataset_index": 56, "sample_id": "polish_00086", "source_ids": [], "target_token_count": 1170, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 44, "field_char_start": 42, "field_index": 0, "field_text": "4C", "gold_field": "4C", "gold_line": "4C\t4r", "gold_note_token": null, "gold_piece": "4C", "gold_token_id": 181, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 47, "line_char_start": 42, "line_index": 2, "line_text": "4C\t4r", "operation": "replace", "pred_field": "*M4/4", "pred_local_replacement_line": "*M4/4\t4r", "pred_note_token": null, "pred_piece": "*M4/4", "pred_token_id": 229, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 8.444656488549619, "LER": 40.24390411376953, "SER": 8.383233532934131}, "nll": 0.3874559700489044, "top1_error_count": 98, "top1_error_rate": 0.08376068376068375}}
+{"dataset_index": 57, "sample_id": "polish_00087", "source_ids": [], "target_token_count": 1300, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 47, "char_start": 44, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 56, "field_char_start": 44, "field_index": 1, "field_text": "8ccq( 8ee-q(", "gold_field": "8ccq( 8ee-q(", "gold_line": ".\t8ccq( 8ee-q(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 56, "line_char_start": 42, "line_index": 2, "line_text": ".\t8ccq( 8ee-q(", "operation": "replace", "pred_field": "8ccqq( 8ee-q(", "pred_local_replacement_line": ".\t8ccqq( 8ee-q(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 14.403829416884248, "LER": 56.5217399597168, "SER": 12.240184757505773}, "nll": 0.5775654315948486, "top1_error_count": 159, "top1_error_rate": 0.12230769230769231}}
+{"dataset_index": 58, "sample_id": "polish_00089", "source_ids": [], "target_token_count": 782, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 37, "char_start": 35, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 43, "field_char_start": 35, "field_index": 0, "field_text": "8A( 8c(L", "gold_field": "8A( 8c(L", "gold_line": "8A( 8c(L\t2.D\t16g#(L", "gold_note_token": "8A(", "gold_piece": "8A", "gold_token_id": 109, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 35, "line_index": 3, "line_text": "8A( 8c(L\t2.D\t16g#(L", "operation": "replace", "pred_field": "8A(L( 8c(L", "pred_local_replacement_line": "8A(L( 8c(L\t2.D\t16g#(L", "pred_note_token": "8A(L(", "pred_piece": "8A(L", "pred_token_id": 1848, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 12.407282535401214, "LER": 50.931678771972656, "SER": 15.236875800256081}, "nll": 1.0140860080718994, "top1_error_count": 119, "top1_error_rate": 0.15217391304347827}}
+{"dataset_index": 59, "sample_id": "polish_00090", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 276, "char_start": 272, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 276, "field_char_start": 272, "field_index": 0, "field_text": "8AAL", "gold_field": "8AAL", "gold_line": "8AAL\t8.r\t4ee", "gold_note_token": "8AAL", "gold_piece": "8AAL", "gold_token_id": 512, "gold_token_index": 164, "gold_trimmed_index": 164, "group": "visual", "line_char_end": 284, "line_char_start": 272, "line_index": 22, "line_text": "8AAL\t8.r\t4ee", "operation": "replace", "pred_field": "8AA", "pred_local_replacement_line": "8AA\t8.r\t4ee", "pred_note_token": "8AA", "pred_piece": "8AA", "pred_token_id": 210, "pred_trimmed_index": 164, "visually_grounded": true}, "metrics": {"CER": 8.554083885209714, "LER": 34.899330139160156, "SER": 7.384341637010676}, "nll": 0.4016970992088318, "top1_error_count": 83, "top1_error_rate": 0.07377777777777778}}
+{"dataset_index": 60, "sample_id": "polish_00092", "source_ids": [], "target_token_count": 1153, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 62, "char_start": 61, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 62, "field_char_start": 52, "field_index": 1, "field_text": "4d-( 4dd-(", "gold_field": "4d-( 4dd-(", "gold_line": "8BB-\t4d-( 4dd-(", "gold_note_token": "4dd-(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 47, "line_index": 3, "line_text": "8BB-\t4d-( 4dd-(", "operation": "replace", "pred_field": "4d-( 4dd-\n", "pred_local_replacement_line": "8BB-\t4d-( 4dd-\n", "pred_note_token": "4dd-\n", "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 5.856727976766699, "LER": 35.754188537597656, "SER": 7.552083333333333}, "nll": 0.530324399471283, "top1_error_count": 87, "top1_error_rate": 0.07545533391153512}}
+{"dataset_index": 61, "sample_id": "polish_00093", "source_ids": [], "target_token_count": 1047, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 555, "char_start": 552, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 564, "field_char_start": 552, "field_index": 1, "field_text": "8e[ 8b[ 8ee[", "gold_field": "8e[ 8b[ 8ee[", "gold_line": "8r\t8e[ 8b[ 8ee[\t8ee", "gold_note_token": "8e[", "gold_piece": "8e[", "gold_token_id": 1374, "gold_token_index": 340, "gold_trimmed_index": 340, "group": "visual", "line_char_end": 568, "line_char_start": 549, "line_index": 45, "line_text": "8r\t8e[ 8b[ 8ee[\t8ee", "operation": "replace", "pred_field": "8e 8b[ 8ee[", "pred_local_replacement_line": "8r\t8e 8b[ 8ee[\t8ee", "pred_note_token": "8e", "pred_piece": "8e", "pred_token_id": 90, "pred_trimmed_index": 340, "visually_grounded": true}, "metrics": {"CER": 8.219988172678889, "LER": 32.11678695678711, "SER": 7.170172084130019}, "nll": 0.34149670600891113, "top1_error_count": 75, "top1_error_rate": 0.07163323782234957}}
+{"dataset_index": 62, "sample_id": "polish_00094", "source_ids": [], "target_token_count": 773, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 40, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 47, "field_char_start": 40, "field_index": 1, "field_text": "32qqggL", "gold_field": "32qqggL", "gold_line": ".\t32qqggL", "gold_note_token": "32qqggL", "gold_piece": "32qq", "gold_token_id": 1235, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 47, "line_char_start": 38, "line_index": 2, "line_text": ".\t32qqggL", "operation": "replace", "pred_field": "32ggggL", "pred_local_replacement_line": ".\t32ggggL", "pred_note_token": "32ggggL", "pred_piece": "32gg", "pred_token_id": 477, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 10.202702702702702, "LER": 45.16128921508789, "SER": 12.435233160621761}, "nll": 0.7796458601951599, "top1_error_count": 96, "top1_error_rate": 0.12419146183699871}}
+{"dataset_index": 63, "sample_id": "polish_00095", "source_ids": [], "target_token_count": 1073, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "4D-", "pred_local_replacement_line": "4D-\t*^", "pred_note_token": null, "pred_piece": "4D-", "pred_token_id": 556, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 17.687074829931973, "LER": 43.81443405151367, "SER": 13.152985074626866}, "nll": 0.6600613594055176, "top1_error_count": 141, "top1_error_rate": 0.13140726933830382}}
+{"dataset_index": 64, "sample_id": "polish_00096", "source_ids": [], "target_token_count": 935, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 52, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "2en", "gold_field": "2en", "gold_line": "2en\t4F#]\t4ggn\t2cc#", "gold_note_token": "2en", "gold_piece": "2en", "gold_token_id": 2374, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 70, "line_char_start": 52, "line_index": 3, "line_text": "2en\t4F#]\t4ggn\t2cc#", "operation": "replace", "pred_field": "2e", "pred_local_replacement_line": "2e\t4F#]\t4ggn\t2cc#", "pred_note_token": "2e", "pred_piece": "2e", "pred_token_id": 296, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 4.123711340206185, "LER": 33.57143020629883, "SER": 4.925053533190578}, "nll": 0.2970961034297943, "top1_error_count": 46, "top1_error_rate": 0.04919786096256684}}
+{"dataset_index": 65, "sample_id": "polish_00097", "source_ids": [], "target_token_count": 1050, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 61, "char_start": 57, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 63, "field_char_start": 57, "field_index": 1, "field_text": "16gg[L", "gold_field": "16gg[L", "gold_line": "8c 8eJ\t16gg[L\t16dd(L", "gold_note_token": "16gg[L", "gold_piece": "16gg", "gold_token_id": 146, "gold_token_index": 22, "gold_trimmed_index": 22, "group": "visual", "line_char_end": 70, "line_char_start": 50, "line_index": 3, "line_text": "8c 8eJ\t16gg[L\t16dd(L", "operation": "replace", "pred_field": "16ggLL[L", "pred_local_replacement_line": "8c 8eJ\t16ggLL[L\t16dd(L", "pred_note_token": "16ggLL[L", "pred_piece": "16ggLL", "pred_token_id": 382, "pred_trimmed_index": 22, "visually_grounded": true}, "metrics": {"CER": 7.03547805171377, "LER": 29.850746154785156, "SER": 6.291706387035272}, "nll": 0.3562990427017212, "top1_error_count": 66, "top1_error_rate": 0.06285714285714286}}
+{"dataset_index": 66, "sample_id": "polish_00100", "source_ids": [], "target_token_count": 1185, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 36, "char_start": 34, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 36, "field_char_start": 34, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "16DD", "pred_local_replacement_line": "16DD\t*", "pred_note_token": null, "pred_piece": "16DD", "pred_token_id": 526, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 14.248838409912235, "LER": 50.79365158081055, "SER": 13.93581081081081}, "nll": 0.6828967928886414, "top1_error_count": 165, "top1_error_rate": 0.13924050632911392}}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/summary.json

@@ -0,0 +1,454 @@
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_2.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.315440778799351,
+      "errors": 2333,
+      "errors_per_1k_tokens": 31.30913238945179,
+      "errors_per_page_mean": 34.82089552238806,
+      "errors_per_page_median": 34.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.22214710654407788,
+      "errors": 1643,
+      "errors_per_1k_tokens": 22.04925182849091,
+      "errors_per_page_mean": 24.52238805970149,
+      "errors_per_page_median": 24.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.10032449972958356,
+      "errors": 742,
+      "errors_per_1k_tokens": 9.9577266322217,
+      "errors_per_page_mean": 11.074626865671641,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05543537047052461,
+      "errors": 410,
+      "errors_per_1k_tokens": 5.502247869556465,
+      "errors_per_page_mean": 6.119402985074627,
+      "errors_per_page_median": 5.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04488912925905895,
+      "errors": 332,
+      "errors_per_1k_tokens": 4.455478762665235,
+      "errors_per_page_mean": 4.955223880597015,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9402985074626866,
+      "pages_with_error": 63
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.03880475932936722,
+      "errors": 287,
+      "errors_per_1k_tokens": 3.8515735086895257,
+      "errors_per_page_mean": 4.2835820895522385,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.8805970149253731,
+      "pages_with_error": 59
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03339643050297458,
+      "errors": 247,
+      "errors_per_1k_tokens": 3.3147688384888947,
+      "errors_per_page_mean": 3.6865671641791047,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.8059701492537313,
+      "pages_with_error": 54
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.02609518658734451,
+      "errors": 193,
+      "errors_per_1k_tokens": 2.5900825337180433,
+      "errors_per_page_mean": 2.8805970149253732,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5671641791044776,
+      "pages_with_error": 38
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.022850189291508925,
+      "errors": 169,
+      "errors_per_1k_tokens": 2.2679997315976648,
+      "errors_per_page_mean": 2.5223880597014925,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5373134328358209,
+      "pages_with_error": 36
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.022714981070849107,
+      "errors": 168,
+      "errors_per_1k_tokens": 2.2545796148426493,
+      "errors_per_page_mean": 2.5074626865671643,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.014872904272579772,
+      "errors": 110,
+      "errors_per_1k_tokens": 1.4762128430517345,
+      "errors_per_page_mean": 1.6417910447761195,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5970149253731343,
+      "pages_with_error": 40
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.012709572742022715,
+      "errors": 94,
+      "errors_per_1k_tokens": 1.2614909749714822,
+      "errors_per_page_mean": 1.4029850746268657,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5970149253731343,
+      "pages_with_error": 40
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.011357490535424553,
+      "errors": 84,
+      "errors_per_1k_tokens": 1.1272898074213247,
+      "errors_per_page_mean": 1.2537313432835822,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.40298507462686567,
+      "pages_with_error": 27
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.010005408328826392,
+      "errors": 74,
+      "errors_per_1k_tokens": 0.9930886398711669,
+      "errors_per_page_mean": 1.1044776119402986,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.417910447761194,
+      "pages_with_error": 28
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.008247701460248783,
+      "errors": 61,
+      "errors_per_1k_tokens": 0.8186271220559619,
+      "errors_per_page_mean": 0.9104477611940298,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.417910447761194,
+      "pages_with_error": 28
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.007842076798269334,
+      "errors": 58,
+      "errors_per_1k_tokens": 0.7783667717909146,
+      "errors_per_page_mean": 0.8656716417910447,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.34328358208955223,
+      "pages_with_error": 23
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.005813953488372093,
+      "errors": 43,
+      "errors_per_1k_tokens": 0.5770650204656781,
+      "errors_per_page_mean": 0.6417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3880597014925373,
+      "pages_with_error": 26
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005543537047052461,
+      "errors": 41,
+      "errors_per_1k_tokens": 0.5502247869556465,
+      "errors_per_page_mean": 0.6119402985074627,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.0048674959437533805,
+      "errors": 36,
+      "errors_per_1k_tokens": 0.48312420318056765,
+      "errors_per_page_mean": 0.5373134328358209,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.004597079502433748,
+      "errors": 34,
+      "errors_per_1k_tokens": 0.45628396967053614,
+      "errors_per_page_mean": 0.5074626865671642,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.19402985074626866,
+      "pages_with_error": 13
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004326663061114116,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.208955223880597,
+      "pages_with_error": 14
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.0036506219578150353,
+      "errors": 27,
+      "errors_per_1k_tokens": 0.36234315238542575,
+      "errors_per_page_mean": 0.40298507462686567,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.003515413737155219,
+      "errors": 26,
+      "errors_per_1k_tokens": 0.34892303563041,
+      "errors_per_page_mean": 0.3880597014925373,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.0029745808545159546,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0029745808545159546,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.22388059701492538,
+      "pages_with_error": 15
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0029745808545159546,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0028393726338561383,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1044776119402985,
+      "pages_with_error": 7
+    },
+    {
+      "category": "clef",
+      "error_share": 0.002568956192536506,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0024337479718766902,
+      "errors": 18,
+      "errors_per_1k_tokens": 0.24156210159028382,
+      "errors_per_page_mean": 0.26865671641791045,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.208955223880597,
+      "pages_with_error": 14
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.0014872904272579773,
+      "errors": 11,
+      "errors_per_1k_tokens": 0.14762128430517346,
+      "errors_per_page_mean": 0.16417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.13432835820895522,
+      "pages_with_error": 9
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0008112493239588967,
+      "errors": 6,
+      "errors_per_1k_tokens": 0.08052070053009461,
+      "errors_per_page_mean": 0.08955223880597014,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.08955223880597014,
+      "pages_with_error": 6
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0008112493239588967,
+      "errors": 6,
+      "errors_per_1k_tokens": 0.08052070053009461,
+      "errors_per_page_mean": 0.08955223880597014,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.07462686567164178,
+      "pages_with_error": 5
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0005408328826392645,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.00013520822065981613,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 4,
+    "accidental_and_tie_slur": 43,
+    "accidental_only": 332,
+    "articulation_only": 34,
+    "barline": 22,
+    "chord_structure_and_accidental": 6,
+    "chord_structure_and_accidental_and_articulation": 21,
+    "chord_structure_and_accidental_and_tie_slur": 22,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 84,
+    "chord_structure_and_tie_slur": 94,
+    "chord_structure_and_tie_slur_and_articulation": 11,
+    "chord_structure_only": 36,
+    "clef": 19,
+    "column_or_line_break": 410,
+    "duration_and_accidental": 32,
+    "duration_and_pitch": 110,
+    "duration_and_pitch_and_accidental": 61,
+    "duration_and_pitch_and_accidental_and_tie_slur": 6,
+    "duration_and_pitch_and_tie_slur": 27,
+    "duration_and_tie_slur": 58,
+    "duration_only": 742,
+    "interpretation_or_control": 247,
+    "note_other": 2333,
+    "null_token": 74,
+    "pitch_and_accidental": 193,
+    "pitch_and_accidental_and_tie_slur": 26,
+    "pitch_and_tie_slur": 18,
+    "pitch_only": 287,
+    "rest_note_confusion": 169,
+    "spine_operation": 168,
+    "termination": 41,
+    "tie_slur_and_articulation": 22,
+    "tie_slur_only": 1643
+  },
+  "component_error_groups": {
+    "structural": 940,
+    "visual": 6456
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 13.163790344959125,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3144,
+    "articulation": 51,
+    "barline": 21,
+    "clef": 19,
+    "column_or_line_break": 411,
+    "duration": 989,
+    "interpretation_or_control": 144,
+    "note_other": 212,
+    "null_token": 75,
+    "pitch": 20,
+    "rest": 289,
+    "spine_operation": 199,
+    "termination": 29,
+    "tie_slur": 1793
+  },
+  "gold_error_groups": {
+    "structural": 858,
+    "visual": 6538
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.481289863586426,
+      "LER": 42.202693939208984,
+      "SER": 10.14264965057373,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.09925518352009663,
+      "top1_errors": 7396,
+      "top1_errors_per_page": 110.38805970149254
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 41,
+    "predicted_unk_literal": 96
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/summary.md

@@ -0,0 +1,43 @@
+# Transcoda Dataset Eval
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_2.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Dataset kind: `hf_parquet_split`
+- Mode: `teacher_forced`
+- Pages: `67`
+- Tokens: `74,515`
+- Elapsed: `13.2s`
+
+## Metrics
+
+| mode | CER | SER | LER | exact pages | top-1 error rate | top-1 errors/page |
+|---|---:|---:|---:|---:|---:|---:|
+| `teacher_forced` | 9.481290 | 10.142650 | 42.202694 | 0.000% | 9.9255% | 110.388 |
+
+## Teacher-Forced Error Categories
+
+| category | errors |
+|---|---:|
+| `accidental` | 3,144 |
+| `tie_slur` | 1,793 |
+| `duration` | 989 |
+| `column_or_line_break` | 411 |
+| `rest` | 289 |
+| `note_other` | 212 |
+| `spine_operation` | 199 |
+| `interpretation_or_control` | 144 |
+| `null_token` | 75 |
+| `articulation` | 51 |
+| `termination` | 29 |
+| `barline` | 21 |
+| `pitch` | 20 |
+| `clef` | 19 |
+
+## Outputs
+
+- `summary.json`
+- `per_page_records.jsonl`
+- `teacher_forced_predictions.jsonl` when teacher-forced mode is enabled
+- `free_generation_predictions.jsonl` when free-generation mode is enabled
+- `gold_prefix_errors.jsonl` when teacher-forced mode is enabled
+- `teacher_forced_error_report.md` when teacher-forced mode is enabled

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_2/teacher_forced_error_report.md

@@ -0,0 +1,524 @@
+# Teacher-Forced Error Report
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_2.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Split: `eval`
+- Pages: `67`
+- Tokens: `74,515`
+- Top-1 errors: `7,396`
+- Top-1 error rate: `9.926%`
+
+## Component Buckets
+
+| bucket | errors | error share | mean/page | median/page | pages | errors/1k tok |
+|---|---:|---:|---:|---:|---:|---:|
+| `note_other` | 2,333 | 31.54% | 34.821 | 34.000 | 67 (100.0%) | 31.309 |
+| `tie_slur_only` | 1,643 | 22.21% | 24.522 | 24.000 | 67 (100.0%) | 22.049 |
+| `duration_only` | 742 | 10.03% | 11.075 | 3.000 | 62 (92.5%) | 9.958 |
+| `column_or_line_break` | 410 | 5.54% | 6.119 | 5.000 | 61 (91.0%) | 5.502 |
+| `accidental_only` | 332 | 4.49% | 4.955 | 3.000 | 63 (94.0%) | 4.455 |
+| `pitch_only` | 287 | 3.88% | 4.284 | 3.000 | 59 (88.1%) | 3.852 |
+| `interpretation_or_control` | 247 | 3.34% | 3.687 | 2.000 | 54 (80.6%) | 3.315 |
+| `pitch_and_accidental` | 193 | 2.61% | 2.881 | 1.000 | 38 (56.7%) | 2.590 |
+| `rest_note_confusion` | 169 | 2.29% | 2.522 | 1.000 | 36 (53.7%) | 2.268 |
+| `spine_operation` | 168 | 2.27% | 2.507 | 1.000 | 41 (61.2%) | 2.255 |
+| `duration_and_pitch` | 110 | 1.49% | 1.642 | 1.000 | 40 (59.7%) | 1.476 |
+| `chord_structure_and_tie_slur` | 94 | 1.27% | 1.403 | 1.000 | 40 (59.7%) | 1.261 |
+| `chord_structure_and_articulation` | 84 | 1.14% | 1.254 | 0.000 | 27 (40.3%) | 1.127 |
+| `null_token` | 74 | 1.00% | 1.104 | 0.000 | 28 (41.8%) | 0.993 |
+| `duration_and_pitch_and_accidental` | 61 | 0.82% | 0.910 | 0.000 | 28 (41.8%) | 0.819 |
+| `duration_and_tie_slur` | 58 | 0.78% | 0.866 | 0.000 | 23 (34.3%) | 0.778 |
+| `accidental_and_tie_slur` | 43 | 0.58% | 0.642 | 0.000 | 26 (38.8%) | 0.577 |
+| `termination` | 41 | 0.55% | 0.612 | 0.000 | 29 (43.3%) | 0.550 |
+| `chord_structure_only` | 36 | 0.49% | 0.537 | 0.000 | 18 (26.9%) | 0.483 |
+| `articulation_only` | 34 | 0.46% | 0.507 | 0.000 | 13 (19.4%) | 0.456 |
+| `duration_and_accidental` | 32 | 0.43% | 0.478 | 0.000 | 14 (20.9%) | 0.429 |
+| `duration_and_pitch_and_tie_slur` | 27 | 0.37% | 0.403 | 0.000 | 12 (17.9%) | 0.362 |
+| `pitch_and_accidental_and_tie_slur` | 26 | 0.35% | 0.388 | 0.000 | 16 (23.9%) | 0.349 |
+| `chord_structure_and_accidental_and_tie_slur` | 22 | 0.30% | 0.328 | 0.000 | 12 (17.9%) | 0.295 |
+| `tie_slur_and_articulation` | 22 | 0.30% | 0.328 | 0.000 | 15 (22.4%) | 0.295 |
+| `barline` | 22 | 0.30% | 0.328 | 0.000 | 16 (23.9%) | 0.295 |
+| `chord_structure_and_accidental_and_articulation` | 21 | 0.28% | 0.313 | 0.000 | 7 (10.4%) | 0.282 |
+| `clef` | 19 | 0.26% | 0.284 | 0.000 | 12 (17.9%) | 0.255 |
+| `pitch_and_tie_slur` | 18 | 0.24% | 0.269 | 0.000 | 14 (20.9%) | 0.242 |
+| `chord_structure_and_tie_slur_and_articulation` | 11 | 0.15% | 0.164 | 0.000 | 9 (13.4%) | 0.148 |
+| `duration_and_pitch_and_accidental_and_tie_slur` | 6 | 0.08% | 0.090 | 0.000 | 6 (9.0%) | 0.081 |
+| `chord_structure_and_accidental` | 6 | 0.08% | 0.090 | 0.000 | 5 (7.5%) | 0.081 |
+| `accidental_and_articulation` | 4 | 0.05% | 0.060 | 0.000 | 4 (6.0%) | 0.054 |
+| `chord_structure_and_accidental_and_tie_slur_and_articulation` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+
+## Legacy Regex Buckets
+
+These are retained for compatibility and should not be treated as component diffs.
+
+| bucket | errors |
+|---|---:|
+| `accidental` | 3,144 |
+| `tie_slur` | 1,793 |
+| `duration` | 989 |
+| `column_or_line_break` | 411 |
+| `rest` | 289 |
+| `note_other` | 212 |
+| `spine_operation` | 199 |
+| `interpretation_or_control` | 144 |
+| `null_token` | 75 |
+| `articulation` | 51 |
+| `termination` | 29 |
+| `barline` | 21 |
+| `pitch` | 20 |
+| `clef` | 19 |
+
+## Error Samples
+
+### Sample 1
+
+```text
+4d	4D 4A	8ff#(L	4f# 4a
+teacher-forced top-1 local replacement:
+4d	4D 4A-	8ff#(L	4f# 4a
+```
+
+### Sample 2
+
+```text
+8Bn( 8d( 8f(L	8.A-(L
+teacher-forced top-1 local replacement:
+8Bn( 8d 8f(L	8.A-(L
+```
+
+### Sample 3
+
+```text
+8r	24f#(L
+teacher-forced top-1 local replacement:
+8r	24f#(LL
+```
+
+### Sample 4
+
+```text
+4FF# 4A#	4e) 4cc#)
+teacher-forced top-1 local replacement:
+4FF# 4A#	4e) 4cc#
+```
+
+### Sample 5
+
+```text
+4Cn	8g( 8ccn( 8ee(L
+teacher-forced top-1 local replacement:
+4Cn	8g( 8ccn  8ee(L
+```
+
+### Sample 6
+
+```text
+8c 8fJ	16cc[L	.
+teacher-forced top-1 local replacement:
+8c 8fJ	16ccLL[L	.
+```
+
+### Sample 7
+
+```text
+.	12g-J	.
+teacher-forced top-1 local replacement:
+.	.-J	.
+```
+
+### Sample 8
+
+```text
+8Gn 8e	16cc#L
+teacher-forced top-1 local replacement:
+8Gn 8e	16cc#LL
+```
+
+### Sample 9
+
+```text
+.	32qqbJ	.
+teacher-forced top-1 local replacement:
+.	32qqbJ)JJJ.
+```
+
+### Sample 10
+
+```text
+4A	8a 8cc#L
+teacher-forced top-1 local replacement:
+4A	8a`8cc#L
+```
+
+### Sample 11
+
+```text
+4B-( 4d-(	4g( 4een(
+teacher-forced top-1 local replacement:
+4B-([4d-(	4g( 4een(
+```
+
+### Sample 12
+
+```text
+*clefF4	*clefG2
+teacher-forced top-1 local replacement:
+
+clefF4	*clefG2
+```
+
+### Sample 13
+
+```text
+8EE- 8E-	8ee- 8gg 8eee-
+teacher-forced top-1 local replacement:
+8EE- 8E-	8ee- 8aa- 8eee-
+```
+
+### Sample 14
+
+```text
+4en( 4g(	2c	4cc)[
+teacher-forced top-1 local replacement:
+4e( 4g(	2c	4cc)[
+```
+
+### Sample 15
+
+```text
+8C]L	16ddn(L
+teacher-forced top-1 local replacement:
+8C]L	16ddn(LL
+```
+
+### Sample 16
+
+```text
+.	.	16dd)J	.
+teacher-forced top-1 local replacement:
+.	.	16dd)JJ	.
+```
+
+
+## Sample Metadata
+
+### Sample 1
+
+- Sample: `polish_00071`
+- Row/token: `49` / `1327`
+- Line/field: `171:1`
+- Component bucket: `accidental_only`; flags: `accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `4A` -> `4A-`
+- Gold note -> top-1 note: `4A` -> `4A-`
+- Gold token likelihood: rank `2`, p `42.594%`
+- Predicted token likelihood: p `54.692%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 314 | `"4A-"` | 54.692% |
+| 2 | 127 | `"4A"` | 42.594% |
+| 3 | 208 | `"4B-"` | 0.459% |
+| 4 | 134 | `"4G"` | 0.217% |
+| 5 | 713 | `"4A#"` | 0.197% |
+
+### Sample 2
+
+- Sample: `polish_00084`
+- Row/token: `54` / `680`
+- Line/field: `121:0`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `8d(` -> `8d`
+- Gold note -> top-1 note: `8d(` -> `8d`
+- Gold token likelihood: rank `2`, p `7.778%`
+- Predicted token likelihood: p `83.625%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 96 | `"8d"` | 83.625% |
+| 2 | 1822 | `"8d("` | 7.778% |
+| 3 | 1247 | `"8d["` | 1.086% |
+| 4 | 1428 | `"8d)"` | 0.820% |
+| 5 | 206 | `"8dL"` | 0.513% |
+
+### Sample 3
+
+- Sample: `polish_00025`
+- Row/token: `16` / `15`
+- Line/field: `3:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `(L` -> `(LL`
+- Gold note -> top-1 note: `24f#(L` -> `24f#(LL`
+- Gold token likelihood: rank `18`, p `0.022%`
+- Predicted token likelihood: p `87.281%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 279 | `"(LL"` | 87.281% |
+| 2 | 1633 | `"(LLL"` | 9.793% |
+| 3 | 83 | `"LL"` | 0.417% |
+| 4 | 308 | `"LLL"` | 0.305% |
+| 5 | 269 | `"#LL"` | 0.127% |
+
+### Sample 4
+
+- Sample: `polish_00040`
+- Row/token: `27` / `846`
+- Line/field: `135:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `4cc#)` -> `4cc#`
+- Gold note -> top-1 note: `4cc#)` -> `4cc#`
+- Gold token likelihood: rank `2`, p `0.195%`
+- Predicted token likelihood: p `98.029%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 248 | `"4cc#"` | 98.029% |
+| 2 | 2699 | `"4cc#)"` | 0.195% |
+| 3 | 2734 | `"4cc#("` | 0.115% |
+| 4 | 731 | `"2cc#"` | 0.089% |
+| 5 | 1739 | `"4cc#["` | 0.065% |
+
+### Sample 5
+
+- Sample: `polish_00017`
+- Row/token: `10` / `477`
+- Line/field: `71:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `(` -> ` `
+- Gold note -> top-1 note: `8ccn(` -> `8ccn`
+- Gold token likelihood: rank `2`, p `0.102%`
+- Predicted token likelihood: p `99.092%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 6 | `" "` | 99.092% |
+| 2 | 55 | `"i"` | 0.102% |
+| 3 | 9 | `"("` | 0.102% |
+| 4 | 26 | `":"` | 0.033% |
+| 5 | 59 | `"n"` | 0.024% |
+
+### Sample 6
+
+- Sample: `polish_00049`
+- Row/token: `33` / `897`
+- Line/field: `122:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16cc` -> `16ccLL`
+- Gold note -> top-1 note: `16cc[L` -> `16ccLL[L`
+- Gold token likelihood: rank `3`, p `2.080%`
+- Predicted token likelihood: p `83.105%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 375 | `"16ccLL"` | 83.105% |
+| 2 | 1889 | `"16cc]LL"` | 7.730% |
+| 3 | 106 | `"16cc"` | 2.080% |
+| 4 | 2813 | `"16cc[LL"` | 1.079% |
+| 5 | 1309 | `"16cc(LL"` | 0.436% |
+
+### Sample 7
+
+- Sample: `polish_00095`
+- Row/token: `63` / `32`
+- Line/field: `5:1`
+- Component bucket: `duration_and_pitch_and_accidental`; flags: `duration, pitch, accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `12g` -> `.`
+- Gold note -> top-1 note: `12g-J` -> `.-J`
+- Gold token likelihood: rank `6`, p `1.729%`
+- Predicted token likelihood: p `46.010%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 14 | `"."` | 46.010% |
+| 2 | 1461 | `"16g-JJ"` | 9.348% |
+| 3 | 968 | `"8g-J"` | 8.511% |
+| 4 | 2325 | `"16g-Jk"` | 4.415% |
+| 5 | 142 | `"16g"` | 3.230% |
+
+### Sample 8
+
+- Sample: `polish_00020`
+- Row/token: `12` / `588`
+- Line/field: `108:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `16cc#L` -> `16cc#LL`
+- Gold note -> top-1 note: `16cc#L` -> `16cc#LL`
+- Gold token likelihood: rank `15`, p `0.066%`
+- Predicted token likelihood: p `95.548%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 545 | `"16cc#LL"` | 95.548% |
+| 2 | 2448 | `"32cc#LLL"` | 0.685% |
+| 3 | 207 | `"16cc#"` | 0.551% |
+| 4 | 228 | `"8cc#L"` | 0.168% |
+| 5 | 366 | `"32cc"` | 0.153% |
+
+### Sample 9
+
+- Sample: `polish_00038`
+- Row/token: `25` / `1118`
+- Line/field: `135:1`
+- Component bucket: `column_or_line_break`; flags: `none`
+- Legacy bucket: `column_or_line_break`
+- Gold piece -> top-1 piece: `	` -> `)JJJ`
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `2`, p `13.968%`
+- Predicted token likelihood: p `30.508%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1570 | `")JJJ"` | 30.508% |
+| 2 | 4 | `"\t"` | 13.968% |
+| 3 | 305 | `")JJ"` | 9.018% |
+| 4 | 321 | `"JJJ"` | 6.395% |
+| 5 | 2968 | `")JJk"` | 2.280% |
+
+### Sample 10
+
+- Sample: `polish_00037`
+- Row/token: `24` / `686`
+- Line/field: `112:1`
+- Component bucket: `chord_structure_and_accidental_and_articulation`; flags: `accidental, articulation, chord_structure`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: ` ` -> ```
+- Gold note -> top-1 note: `8a` -> `8a`8cc#L`
+- Gold token likelihood: rank `4`, p `4.355%`
+- Predicted token likelihood: p `31.188%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 47 | `"\`"` | 31.188% |
+| 2 | 66 | `"v"` | 18.334% |
+| 3 | 2620 | `"vL"` | 5.091% |
+| 4 | 6 | `" "` | 4.355% |
+| 5 | 12 | `","` | 3.186% |
+
+### Sample 11
+
+- Sample: `polish_00052`
+- Row/token: `35` / `457`
+- Line/field: `82:0`
+- Component bucket: `chord_structure_and_tie_slur`; flags: `tie_slur, chord_structure`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: ` ` -> `[`
+- Gold note -> top-1 note: `4B-(` -> `4B-([4d-(`
+- Gold token likelihood: rank `3`, p `17.667%`
+- Predicted token likelihood: p `39.813%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 43 | `"["` | 39.813% |
+| 2 | 9 | `"("` | 21.311% |
+| 3 | 6 | `" "` | 17.667% |
+| 4 | 4 | `"\t"` | 11.056% |
+| 5 | 2909 | `"(["` | 0.664% |
+
+### Sample 12
+
+- Sample: `polish_00089`
+- Row/token: `58` / `781`
+- Line/field: `0:0`
+- Component bucket: `termination`; flags: `none`
+- Legacy bucket: `termination`
+- Gold piece -> top-1 piece: `<eos>` -> `
+`
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `2`, p `0.039%`
+- Predicted token likelihood: p `99.477%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 5 | `"\n"` | 99.477% |
+| 2 | 2 | `"<eos>"` | 0.039% |
+| 3 | 6 | `" "` | 0.009% |
+| 4 | 26 | `":"` | 0.003% |
+| 5 | 4 | `"\t"` | 0.002% |
+
+### Sample 13
+
+- Sample: `polish_00027`
+- Row/token: `17` / `1453`
+- Line/field: `208:1`
+- Component bucket: `pitch_and_accidental`; flags: `pitch, accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `8gg` -> `8aa-`
+- Gold note -> top-1 note: `8gg` -> `8aa-`
+- Gold token likelihood: rank `5`, p `0.375%`
+- Predicted token likelihood: p `94.666%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 683 | `"8aa-"` | 94.666% |
+| 2 | 293 | `"8bb-"` | 1.846% |
+| 3 | 105 | `"8aa"` | 1.119% |
+| 4 | 909 | `"8gg-"` | 0.638% |
+| 5 | 87 | `"8gg"` | 0.375% |
+
+### Sample 14
+
+- Sample: `polish_00087`
+- Row/token: `57` / `922`
+- Line/field: `126:0`
+- Component bucket: `accidental_only`; flags: `accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `4en` -> `4e`
+- Gold note -> top-1 note: `4en(` -> `4e(`
+- Gold token likelihood: rank `2`, p `44.043%`
+- Predicted token likelihood: p `46.884%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 114 | `"4e"` | 46.884% |
+| 2 | 1198 | `"4en"` | 44.043% |
+| 3 | 1426 | `"4e("` | 1.507% |
+| 4 | 191 | `"*^"` | 0.886% |
+| 5 | 1135 | `"4e["` | 0.832% |
+
+### Sample 15
+
+- Sample: `polish_00089`
+- Row/token: `58` / `297`
+- Line/field: `47:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `(L` -> `(LL`
+- Gold note -> top-1 note: `16ddn(L` -> `16ddn(LL`
+- Gold token likelihood: rank `6`, p `0.106%`
+- Predicted token likelihood: p `96.261%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 279 | `"(LL"` | 96.261% |
+| 2 | 55 | `"i"` | 1.069% |
+| 3 | 9 | `"("` | 0.306% |
+| 4 | 1633 | `"(LLL"` | 0.231% |
+| 5 | 305 | `")JJ"` | 0.128% |
+
+### Sample 16
+
+- Sample: `polish_00013`
+- Row/token: `7` / `1170`
+- Line/field: `139:2`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `)J` -> `)JJ`
+- Gold note -> top-1 note: `16dd)J` -> `16dd)JJ`
+- Gold token likelihood: rank `4`, p `0.566%`
+- Predicted token likelihood: p `92.238%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 305 | `")JJ"` | 92.238% |
+| 2 | 1570 | `")JJJ"` | 1.357% |
+| 3 | 10 | `")"` | 0.662% |
+| 4 | 186 | `")J"` | 0.566% |
+| 5 | 363 | `"]JJ"` | 0.222% |

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/command.log

@@ -0,0 +1,491 @@
+Skipping import of cpp extensions due to incompatible torch version 2.11.0+cu129 for torchao version 0.15.0             Please see https://github.com/pytorch/ao/issues/2919 for more info
+[transformers] You are using a model of type `vit_mae` to instantiate a model of type `vit`. This may be expected if you are loading a checkpoint that shares a subset of the architecture (e.g., loading a `sam2_video` checkpoint into `Sam2Model`), but is otherwise not supported and can yield errors. Please verify that the checkpoint is compatible with the model you are instantiating.
+
+[transformers] ViTModel LOAD REPORT from: PRAIG/musvit
+Key                                                                     | Status     | 
+------------------------------------------------------------------------+------------+-
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.bias    | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.bias            | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.weight  | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.weight          | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.bias            | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.weight          | UNEXPECTED | 
+decoder.decoder_norm.weight                                             | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.bias   | UNEXPECTED | 
+decoder.decoder_pos_embed                                               | UNEXPECTED | 
+decoder.decoder_pred.bias                                               | UNEXPECTED | 
+decoder.decoder_norm.bias                                               | UNEXPECTED | 
+decoder.decoder_embed.bias                                              | UNEXPECTED | 
+decoder.mask_token                                                      | UNEXPECTED | 
+decoder.decoder_embed.weight                                            | UNEXPECTED | 
+decoder.decoder_pred.weight                                             | UNEXPECTED | 
+pooler.dense.weight                                                     | MISSING    | 
+pooler.dense.bias                                                       | MISSING    | 
+
+Notes:
+- UNEXPECTED:	can be ignored when loading from different task/architecture; not ok if you expect identical arch.
+- MISSING:	those params were newly initialized because missing from the checkpoint. Consider training on your downstream task.
+[eval] pages=32/67 mode=teacher_forced elapsed=9.1s
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_3.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.3157254796001081,
+      "errors": 2337,
+      "errors_per_1k_tokens": 31.36281285647185,
+      "errors_per_page_mean": 34.88059701492537,
+      "errors_per_page_median": 35.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.22466900837611456,
+      "errors": 1663,
+      "errors_per_1k_tokens": 22.31765416359122,
+      "errors_per_page_mean": 24.82089552238806,
+      "errors_per_page_median": 25.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.10497162928938125,
+      "errors": 777,
+      "errors_per_1k_tokens": 10.427430718647253,
+      "errors_per_page_mean": 11.597014925373134,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05498513915158065,
+      "errors": 407,
+      "errors_per_1k_tokens": 5.461987519291418,
+      "errors_per_page_mean": 6.074626865671642,
+      "errors_per_page_median": 5.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04228586868413942,
+      "errors": 313,
+      "errors_per_1k_tokens": 4.200496544319936,
+      "errors_per_page_mean": 4.6716417910447765,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03485544447446636,
+      "errors": 258,
+      "errors_per_1k_tokens": 3.4623901227940683,
+      "errors_per_page_mean": 3.8507462686567164,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.7611940298507462,
+      "pages_with_error": 51
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.03390975412050797,
+      "errors": 251,
+      "errors_per_1k_tokens": 3.3684493055089577,
+      "errors_per_page_mean": 3.746268656716418,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.835820895522388,
+      "pages_with_error": 56
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.026074034044852742,
+      "errors": 193,
+      "errors_per_1k_tokens": 2.5900825337180433,
+      "errors_per_page_mean": 2.8805970149253732,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.582089552238806,
+      "pages_with_error": 39
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.023372061604971628,
+      "errors": 173,
+      "errors_per_1k_tokens": 2.321680198617728,
+      "errors_per_page_mean": 2.582089552238806,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5074626865671642,
+      "pages_with_error": 34
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.023236962982977574,
+      "errors": 172,
+      "errors_per_1k_tokens": 2.3082600818627124,
+      "errors_per_page_mean": 2.5671641791044775,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.014860848419346123,
+      "errors": 110,
+      "errors_per_1k_tokens": 1.4762128430517345,
+      "errors_per_page_mean": 1.6417910447761195,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.012293974601459066,
+      "errors": 91,
+      "errors_per_1k_tokens": 1.2212306247064348,
+      "errors_per_page_mean": 1.3582089552238805,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.582089552238806,
+      "pages_with_error": 39
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.012158875979465009,
+      "errors": 90,
+      "errors_per_1k_tokens": 1.2078105079514192,
+      "errors_per_page_mean": 1.3432835820895523,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.373134328358209,
+      "pages_with_error": 25
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.008511213185625506,
+      "errors": 63,
+      "errors_per_1k_tokens": 0.8454673555659934,
+      "errors_per_page_mean": 0.9402985074626866,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.31343283582089554,
+      "pages_with_error": 21
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.008241015941637396,
+      "errors": 61,
+      "errors_per_1k_tokens": 0.8186271220559619,
+      "errors_per_page_mean": 0.9104477611940298,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.44776119402985076,
+      "pages_with_error": 30
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.007565522831667117,
+      "errors": 56,
+      "errors_per_1k_tokens": 0.751526538280883,
+      "errors_per_page_mean": 0.835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.373134328358209,
+      "pages_with_error": 25
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005944339367738449,
+      "errors": 44,
+      "errors_per_1k_tokens": 0.5904851372206938,
+      "errors_per_page_mean": 0.6567164179104478,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.47761194029850745,
+      "pages_with_error": 32
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.005539043501756282,
+      "errors": 41,
+      "errors_per_1k_tokens": 0.5502247869556465,
+      "errors_per_page_mean": 0.6119402985074627,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3880597014925373,
+      "pages_with_error": 26
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.0045933531477978925,
+      "errors": 34,
+      "errors_per_1k_tokens": 0.45628396967053614,
+      "errors_per_page_mean": 0.5074626865671642,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.004323155903809781,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.004323155903809781,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.004323155903809781,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0032423669278573357,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0032423669278573357,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.0031072683058632803,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.14925373134328357,
+      "pages_with_error": 10
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0031072683058632803,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.0029721696838692245,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0025668738178870575,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1044776119402985,
+      "pages_with_error": 7
+    },
+    {
+      "category": "clef",
+      "error_share": 0.0022966765738989463,
+      "errors": 17,
+      "errors_per_1k_tokens": 0.22814198483526807,
+      "errors_per_page_mean": 0.2537313432835821,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.16417910447761194,
+      "pages_with_error": 11
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.0013509862199405566,
+      "errors": 10,
+      "errors_per_1k_tokens": 0.13420116755015768,
+      "errors_per_page_mean": 0.14925373134328357,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0005403944879762226,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.00040529586598216696,
+      "errors": 3,
+      "errors_per_1k_tokens": 0.040260350265047304,
+      "errors_per_page_mean": 0.04477611940298507,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.04477611940298507,
+      "pages_with_error": 3
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0002701972439881113,
+      "errors": 2,
+      "errors_per_1k_tokens": 0.026840233510031536,
+      "errors_per_page_mean": 0.029850746268656716,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.029850746268656716,
+      "pages_with_error": 2
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.00013509862199405565,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 4,
+    "accidental_and_tie_slur": 41,
+    "accidental_only": 313,
+    "articulation_only": 32,
+    "barline": 24,
+    "chord_structure_and_accidental": 2,
+    "chord_structure_and_accidental_and_articulation": 19,
+    "chord_structure_and_accidental_and_tie_slur": 22,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 90,
+    "chord_structure_and_tie_slur": 91,
+    "chord_structure_and_tie_slur_and_articulation": 10,
+    "chord_structure_only": 32,
+    "clef": 17,
+    "column_or_line_break": 407,
+    "duration_and_accidental": 34,
+    "duration_and_pitch": 110,
+    "duration_and_pitch_and_accidental": 56,
+    "duration_and_pitch_and_accidental_and_tie_slur": 3,
+    "duration_and_pitch_and_tie_slur": 23,
+    "duration_and_tie_slur": 63,
+    "duration_only": 777,
+    "interpretation_or_control": 258,
+    "note_other": 2337,
+    "null_token": 61,
+    "pitch_and_accidental": 193,
+    "pitch_and_accidental_and_tie_slur": 32,
+    "pitch_and_tie_slur": 23,
+    "pitch_only": 251,
+    "rest_note_confusion": 173,
+    "spine_operation": 172,
+    "termination": 44,
+    "tie_slur_and_articulation": 24,
+    "tie_slur_only": 1663
+  },
+  "component_error_groups": {
+    "structural": 942,
+    "visual": 6460
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 12.819884218974039,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3151,
+    "articulation": 54,
+    "barline": 23,
+    "clef": 17,
+    "column_or_line_break": 408,
+    "duration": 994,
+    "interpretation_or_control": 146,
+    "note_other": 209,
+    "null_token": 62,
+    "pitch": 21,
+    "rest": 281,
+    "spine_operation": 205,
+    "termination": 32,
+    "tie_slur": 1799
+  },
+  "gold_error_groups": {
+    "structural": 853,
+    "visual": 6549
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.608084678649902,
+      "LER": 42.074729919433594,
+      "SER": 10.15070915222168,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.09933570422062672,
+      "top1_errors": 7402,
+      "top1_errors_per_page": 110.4776119402985
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 38,
+    "predicted_unk_literal": 104
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/per_page_records.jsonl

@@ -0,0 +1,67 @@
+{"dataset_index": 0, "sample_id": "polish_00002", "source_ids": [], "target_token_count": 1725, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 59, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 55, "field_index": 0, "field_text": "2.AA[", "gold_field": "2.AA[", "gold_line": "2.AA[\t4r\t12r", "gold_note_token": "2.AA[", "gold_piece": "[", "gold_token_id": 43, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 67, "line_char_start": 55, "line_index": 4, "line_text": "2.AA[\t4r\t12r", "operation": "replace", "pred_field": "2.AA([", "pred_local_replacement_line": "2.AA([\t4r\t12r", "pred_note_token": "2.AA([", "pred_piece": "([", "pred_token_id": 2909, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 13.944076526857984, "LER": 69.29824829101562, "SER": 10.324825986078887}, "nll": 0.45150211453437805, "top1_error_count": 178, "top1_error_rate": 0.10318840579710145}}
+{"dataset_index": 1, "sample_id": "polish_00004", "source_ids": [], "target_token_count": 989, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 35, "field_char_start": 30, "field_index": 0, "field_text": "8AA(L", "gold_field": "8AA(L", "gold_line": "8AA(L\t4.ee(", "gold_note_token": "8AA(L", "gold_piece": "8AA", "gold_token_id": 210, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 41, "line_char_start": 30, "line_index": 2, "line_text": "8AA(L\t4.ee(", "operation": "replace", "pred_field": "=!|:(L", "pred_local_replacement_line": "=!|:(L\t4.ee(", "pred_note_token": "=!|:(L", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 6.428571428571428, "LER": 33.33333206176758, "SER": 7.28744939271255}, "nll": 0.4819011390209198, "top1_error_count": 72, "top1_error_rate": 0.07280080889787664}}
+{"dataset_index": 2, "sample_id": "polish_00005", "source_ids": [], "target_token_count": 742, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t2cc(", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 55, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t2cc(", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t2cc(", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 5.968077723802915, "LER": 40.944881439208984, "SER": 9.176788124156547}, "nll": 0.5751280784606934, "top1_error_count": 68, "top1_error_rate": 0.09164420485175202}}
+{"dataset_index": 3, "sample_id": "polish_00007", "source_ids": [], "target_token_count": 888, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 85, "char_start": 81, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 87, "field_char_start": 81, "field_index": 1, "field_text": "16dd#L", "gold_field": "16dd#L", "gold_line": "8E 8A 8cJ\t16dd#L", "gold_note_token": "16dd#L", "gold_piece": "16dd", "gold_token_id": 112, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 87, "line_char_start": 71, "line_index": 6, "line_text": "8E 8A 8cJ\t16dd#L", "operation": "replace", "pred_field": "16dd#LL#L", "pred_local_replacement_line": "8E 8A 8cJ\t16dd#LL#L", "pred_note_token": "16dd#LL#L", "pred_piece": "16dd#LL", "pred_token_id": 1164, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 7.193320488118175, "LER": 33.33333206176758, "SER": 8.342728297632469}, "nll": 0.5038120746612549, "top1_error_count": 74, "top1_error_rate": 0.08333333333333333}}
+{"dataset_index": 4, "sample_id": "polish_00008", "source_ids": [], "target_token_count": 1236, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 48, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 48, "field_char_start": 45, "field_index": 1, "field_text": "8d[", "gold_field": "8d[", "gold_line": "8r\t8d[", "gold_note_token": "8d[", "gold_piece": "8d[", "gold_token_id": 1247, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 48, "line_char_start": 42, "line_index": 3, "line_text": "8r\t8d[", "operation": "replace", "pred_field": "8d(", "pred_local_replacement_line": "8r\t8d(", "pred_note_token": "8d(", "pred_piece": "8d(", "pred_token_id": 1822, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 10.20603737422137, "LER": 46.073299407958984, "SER": 10.121457489878543}, "nll": 0.5564486384391785, "top1_error_count": 125, "top1_error_rate": 0.1011326860841424}}
+{"dataset_index": 5, "sample_id": "polish_00009", "source_ids": [], "target_token_count": 942, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 111, "char_start": 110, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 110, "field_char_start": 107, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 52, "gold_trimmed_index": 52, "group": "structural", "line_char_end": 120, "line_char_start": 107, "line_index": 11, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 52, "visually_grounded": false}, "metrics": {"CER": 5.2463211772232885, "LER": 22.839506149291992, "SER": 5.419766206163656}, "nll": 0.3408219516277313, "top1_error_count": 51, "top1_error_rate": 0.054140127388535034}}
+{"dataset_index": 6, "sample_id": "polish_00011", "source_ids": [], "target_token_count": 1093, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 7.6923076923076925, "LER": 41.578948974609375, "SER": 9.24908424908425}, "nll": 0.5074254274368286, "top1_error_count": 101, "top1_error_rate": 0.09240622140896615}}
+{"dataset_index": 7, "sample_id": "polish_00013", "source_ids": [], "target_token_count": 1877, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 47, "char_start": 43, "component_category": "duration_and_pitch_and_tie_slur", "component_flags": ["duration", "pitch", "tie_slur"], "component_group": "visual", "field_char_end": 49, "field_char_start": 43, "field_index": 1, "field_text": "16.g(L", "gold_field": "16.g(L", "gold_line": "8CC 8GG 8CL\t16.g(L\t4en", "gold_note_token": "16.g(L", "gold_piece": "16.g", "gold_token_id": 1474, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 53, "line_char_start": 31, "line_index": 3, "line_text": "8CC 8GG 8CL\t16.g(L\t4en", "operation": "replace", "pred_field": "16a(LL(L", "pred_local_replacement_line": "8CC 8GG 8CL\t16a(LL(L\t4en", "pred_note_token": "16a(LL(L", "pred_piece": "16a(LL", "pred_token_id": 1061, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 7.7532167601451665, "LER": 49.77973556518555, "SER": 7.995735607675907}, "nll": 0.4430171549320221, "top1_error_count": 150, "top1_error_rate": 0.07991475759190197}}
+{"dataset_index": 8, "sample_id": "polish_00014", "source_ids": [], "target_token_count": 1099, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t16bn(L", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 57, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t16bn(L", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t16bn(L", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 7.538691962056915, "LER": 33.33333206176758, "SER": 9.836065573770492}, "nll": 0.5544245839118958, "top1_error_count": 108, "top1_error_rate": 0.09827115559599636}}
+{"dataset_index": 9, "sample_id": "polish_00015", "source_ids": [], "target_token_count": 1086, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 38, "char_start": 37, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 38, "field_char_start": 37, "field_index": 1, "field_text": "*", "gold_field": "*", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "*^", "pred_local_replacement_line": "*^\t*^", "pred_note_token": null, "pred_piece": "*^", "pred_token_id": 191, "pred_trimmed_index": 10, "visually_grounded": false}, "metrics": {"CER": 8.019375672766415, "LER": 39.87730026245117, "SER": 9.76958525345622}, "nll": 0.4883825480937958, "top1_error_count": 106, "top1_error_rate": 0.09760589318600368}}
+{"dataset_index": 10, "sample_id": "polish_00017", "source_ids": [], "target_token_count": 1053, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 40, "char_start": 39, "component_category": "chord_structure_and_tie_slur", "component_flags": ["tie_slur", "chord_structure"], "component_group": "visual", "field_char_end": 44, "field_char_start": 34, "field_index": 0, "field_text": "4BBB- 4BB-", "gold_field": "4BBB- 4BB-", "gold_line": "4BBB- 4BB-\t4E 4cn", "gold_note_token": "4BBB-", "gold_piece": " ", "gold_token_id": 6, "gold_token_index": 9, "gold_trimmed_index": 9, "group": "visual", "line_char_end": 51, "line_char_start": 34, "line_index": 2, "line_text": "4BBB- 4BB-\t4E 4cn", "operation": "replace", "pred_field": "4BBB-(4BB-", "pred_local_replacement_line": "4BBB-(4BB-\t4E 4cn", "pred_note_token": "4BBB-(4BB-", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 9, "visually_grounded": true}, "metrics": {"CER": 2.3746701846965697, "LER": 15.55555534362793, "SER": 4.277566539923954}, "nll": 0.26790839433670044, "top1_error_count": 45, "top1_error_rate": 0.042735042735042736}}
+{"dataset_index": 11, "sample_id": "polish_00019", "source_ids": [], "target_token_count": 1006, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 177, "char_start": 175, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 177, "field_char_start": 170, "field_index": 1, "field_text": "16eee)J", "gold_field": "16eee)J", "gold_line": ".\t16eee)J", "gold_note_token": "16eee)J", "gold_piece": ")J", "gold_token_id": 186, "gold_token_index": 90, "gold_trimmed_index": 90, "group": "visual", "line_char_end": 177, "line_char_start": 168, "line_index": 12, "line_text": ".\t16eee)J", "operation": "replace", "pred_field": "16eee)JJ", "pred_local_replacement_line": ".\t16eee)JJ", "pred_note_token": "16eee)JJ", "pred_piece": ")JJ", "pred_token_id": 305, "pred_trimmed_index": 90, "visually_grounded": true}, "metrics": {"CER": 5.072463768115942, "LER": 33.52941131591797, "SER": 7.761194029850746}, "nll": 0.5277189016342163, "top1_error_count": 78, "top1_error_rate": 0.0775347912524851}}
+{"dataset_index": 12, "sample_id": "polish_00020", "source_ids": [], "target_token_count": 1183, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 5.78806767586821, "LER": 33.971290588378906, "SER": 7.698815566835872}, "nll": 0.4539109766483307, "top1_error_count": 91, "top1_error_rate": 0.07692307692307693}}
+{"dataset_index": 13, "sample_id": "polish_00022", "source_ids": [], "target_token_count": 695, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 56, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 52, "field_index": 0, "field_text": "8G( 8B(L", "gold_field": "8G( 8B(L", "gold_line": "8G( 8B(L\t8ddd)L", "gold_note_token": "8B(L", "gold_piece": "8B(L", "gold_token_id": 2168, "gold_token_index": 20, "gold_trimmed_index": 20, "group": "visual", "line_char_end": 67, "line_char_start": 52, "line_index": 4, "line_text": "8G( 8B(L\t8ddd)L", "operation": "replace", "pred_field": "8G( 8BL", "pred_local_replacement_line": "8G( 8BL\t8ddd)L", "pred_note_token": "8BL", "pred_piece": "8BL", "pred_token_id": 396, "pred_trimmed_index": 20, "visually_grounded": true}, "metrics": {"CER": 7.739938080495357, "LER": 41.3533821105957, "SER": 10.230547550432277}, "nll": 0.6753405332565308, "top1_error_count": 71, "top1_error_rate": 0.10215827338129496}}
+{"dataset_index": 14, "sample_id": "polish_00023", "source_ids": [], "target_token_count": 1332, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 51, "field_index": 1, "field_text": "4gg)", "gold_field": "4gg)", "gold_line": "8GG 8GL\t4gg)\t8bnL", "gold_note_token": "4gg)", "gold_piece": "4gg)", "gold_token_id": 2154, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 60, "line_char_start": 43, "line_index": 3, "line_text": "8GG 8GL\t4gg)\t8bnL", "operation": "replace", "pred_field": "4gg]", "pred_local_replacement_line": "8GG 8GL\t4gg]\t8bnL", "pred_note_token": "4gg]", "pred_piece": "4gg]", "pred_token_id": 1531, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 7.3012939001848425, "LER": 40.24390411376953, "SER": 7.738542449286251}, "nll": 0.40940192341804504, "top1_error_count": 103, "top1_error_rate": 0.07732732732732733}}
+{"dataset_index": 15, "sample_id": "polish_00024", "source_ids": [], "target_token_count": 1075, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 8.070757324488667, "LER": 36.99422073364258, "SER": 8.845437616387338}, "nll": 0.503007709980011, "top1_error_count": 95, "top1_error_rate": 0.08837209302325581}}
+{"dataset_index": 16, "sample_id": "polish_00025", "source_ids": [], "target_token_count": 879, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 53, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 55, "field_char_start": 49, "field_index": 1, "field_text": "24f#(L", "gold_field": "24f#(L", "gold_line": "8r\t24f#(L", "gold_note_token": "24f#(L", "gold_piece": "(L", "gold_token_id": 171, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 55, "line_char_start": 46, "line_index": 3, "line_text": "8r\t24f#(L", "operation": "replace", "pred_field": "24f#(LL", "pred_local_replacement_line": "8r\t24f#(LL", "pred_note_token": "24f#(LL", "pred_piece": "(LL", "pred_token_id": 279, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 6.993865030674846, "LER": 45.45454406738281, "SER": 10.250569476082005}, "nll": 0.5319342613220215, "top1_error_count": 90, "top1_error_rate": 0.10238907849829351}}
+{"dataset_index": 17, "sample_id": "polish_00027", "source_ids": [], "target_token_count": 1825, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 46, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 46, "field_index": 0, "field_text": "8BBB- 8BB-", "gold_field": "8BBB- 8BB-", "gold_line": "8BBB- 8BB-\t8b- 8dd", "gold_note_token": "8BBB-", "gold_piece": "8BBB-", "gold_token_id": 632, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 3, "line_text": "8BBB- 8BB-\t8b- 8dd", "operation": "replace", "pred_field": "8AAA 8BB-", "pred_local_replacement_line": "8AAA 8BB-\t8b- 8dd", "pred_note_token": "8AAA", "pred_piece": "8AAA", "pred_token_id": 490, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 8.82609962132246, "LER": 50.0, "SER": 9.758771929824562}, "nll": 0.5067708492279053, "top1_error_count": 178, "top1_error_rate": 0.09753424657534246}}
+{"dataset_index": 18, "sample_id": "polish_00029", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 39, "field_char_start": 30, "field_index": 0, "field_text": "8CC( 8C(L", "gold_field": "8CC( 8C(L", "gold_line": "8CC( 8C(L\t4gg([", "gold_note_token": "8CC(", "gold_piece": "8CC", "gold_token_id": 364, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 45, "line_char_start": 30, "line_index": 2, "line_text": "8CC( 8C(L\t4gg([", "operation": "replace", "pred_field": "=!|:( 8C(L", "pred_local_replacement_line": "=!|:( 8C(L\t4gg([", "pred_note_token": "=!|:(", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 7.7910958904109595, "LER": 36.59574508666992, "SER": 9.97586484312148}, "nll": 0.607370913028717, "top1_error_count": 124, "top1_error_rate": 0.09967845659163987}}
+{"dataset_index": 19, "sample_id": "polish_00030", "source_ids": [], "target_token_count": 1263, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "*met(c|)", "pred_local_replacement_line": "*met(c|)\t*^", "pred_note_token": null, "pred_piece": "*met(c|)", "pred_token_id": 1245, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 11.974110032362459, "LER": 53.488372802734375, "SER": 10.142630744849445}, "nll": 0.4626601040363312, "top1_error_count": 128, "top1_error_rate": 0.10134600158353127}}
+{"dataset_index": 20, "sample_id": "polish_00031", "source_ids": [], "target_token_count": 862, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 34, "field_index": 0, "field_text": "*M3/8", "gold_field": "*M3/8", "gold_line": "*M3/8\t*M3/8", "gold_note_token": null, "gold_piece": "*M3/8", "gold_token_id": 592, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 45, "line_char_start": 34, "line_index": 2, "line_text": "*M3/8\t*M3/8", "operation": "replace", "pred_field": "*M3/4", "pred_local_replacement_line": "*M3/4\t*M3/8", "pred_note_token": null, "pred_piece": "*M3/4", "pred_token_id": 283, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 9.001956947162427, "LER": 39.86014175415039, "SER": 9.40766550522648}, "nll": 0.5317352414131165, "top1_error_count": 81, "top1_error_rate": 0.09396751740139211}}
+{"dataset_index": 21, "sample_id": "polish_00032", "source_ids": [], "target_token_count": 1387, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 50, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 52, "field_char_start": 45, "field_index": 2, "field_text": "16ff#]L", "gold_field": "16ff#]L", "gold_line": "4G#(\t8G#L\t16ff#]L", "gold_note_token": "16ff#]L", "gold_piece": "16ff#", "gold_token_id": 180, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 52, "line_char_start": 35, "line_index": 3, "line_text": "4G#(\t8G#L\t16ff#]L", "operation": "replace", "pred_field": "16ff#)]L", "pred_local_replacement_line": "4G#(\t8G#L\t16ff#)]L", "pred_note_token": "16ff#)]L", "pred_piece": "16ff#)", "pred_token_id": 2192, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 11.356993736951983, "LER": 44.117645263671875, "SER": 11.76046176046176}, "nll": 0.7114558815956116, "top1_error_count": 163, "top1_error_rate": 0.11751982696467195}}
+{"dataset_index": 22, "sample_id": "polish_00034", "source_ids": [], "target_token_count": 788, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 45, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8f( 8a(L", "gold_field": "8f( 8a(L", "gold_line": "8f( 8a(L\t16r", "gold_note_token": null, "gold_piece": "8f(", "gold_token_id": 2288, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 42, "line_index": 3, "line_text": "8f( 8a(L\t16r", "operation": "replace", "pred_field": "* 8a(L", "pred_local_replacement_line": "* 8a(L\t16r", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 24.334600760456272, "LER": 55.90061950683594, "SER": 39.135959339263025}, "nll": 1.4968820810317993, "top1_error_count": 178, "top1_error_rate": 0.22588832487309646}}
+{"dataset_index": 23, "sample_id": "polish_00036", "source_ids": [], "target_token_count": 1317, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 451, "char_start": 447, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 451, "field_char_start": 447, "field_index": 2, "field_text": "16gJ", "gold_field": "16gJ", "gold_line": ".\t.\t16gJ", "gold_note_token": "16gJ", "gold_piece": "16gJ", "gold_token_id": 859, "gold_token_index": 280, "gold_trimmed_index": 280, "group": "visual", "line_char_end": 451, "line_char_start": 443, "line_index": 41, "line_text": ".\t.\t16gJ", "operation": "replace", "pred_field": "16gJJ", "pred_local_replacement_line": ".\t.\t16gJJ", "pred_note_token": "16gJJ", "pred_piece": "16gJJ", "pred_token_id": 351, "pred_trimmed_index": 280, "visually_grounded": true}, "metrics": {"CER": 5.703938433680398, "LER": 21.875, "SER": 4.711246200607903}, "nll": 0.29130154848098755, "top1_error_count": 62, "top1_error_rate": 0.047076689445709946}}
+{"dataset_index": 24, "sample_id": "polish_00037", "source_ids": [], "target_token_count": 871, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 49, "char_start": 46, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 46, "field_index": 0, "field_text": "4Fn 4dn", "gold_field": "4Fn 4dn", "gold_line": "4Fn 4dn\t2f#] 2ff#]", "gold_note_token": "4Fn", "gold_piece": "4Fn", "gold_token_id": 1918, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 2, "line_text": "4Fn 4dn\t2f#] 2ff#]", "operation": "replace", "pred_field": "4F# 4dn", "pred_local_replacement_line": "4F# 4dn\t2f#] 2ff#]", "pred_note_token": "4F#", "pred_piece": "4F#", "pred_token_id": 278, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 4.4473512099411385, "LER": 25.503355026245117, "SER": 6.091954022988506}, "nll": 0.28966453671455383, "top1_error_count": 53, "top1_error_rate": 0.060849598163030996}}
+{"dataset_index": 25, "sample_id": "polish_00038", "source_ids": [], "target_token_count": 1471, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 88, "char_start": 83, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 88, "field_char_start": 83, "field_index": 1, "field_text": "16a(L", "gold_field": "16a(L", "gold_line": "8CC 8FF# 8CJ\t16a(L\t8e-", "gold_note_token": "16a(L", "gold_piece": "16a(L", "gold_token_id": 2253, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 92, "line_char_start": 70, "line_index": 5, "line_text": "8CC 8FF# 8CJ\t16a(L\t8e-", "operation": "replace", "pred_field": "16aLL", "pred_local_replacement_line": "8CC 8FF# 8CJ\t16aLL\t8e-", "pred_note_token": "16aLL", "pred_piece": "16aLL", "pred_token_id": 338, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 6.914468425259793, "LER": 39.106143951416016, "SER": 7.687074829931973}, "nll": 0.42855408787727356, "top1_error_count": 113, "top1_error_rate": 0.07681849082256968}}
+{"dataset_index": 26, "sample_id": "polish_00039", "source_ids": [], "target_token_count": 1503, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 93, "char_start": 88, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 93, "field_char_start": 88, "field_index": 0, "field_text": "16A-L", "gold_field": "16A-L", "gold_line": "16A-L\t2.AA-[\t16cL\t8ee-", "gold_note_token": "16A-L", "gold_piece": "16A-L", "gold_token_id": 2834, "gold_token_index": 24, "gold_trimmed_index": 24, "group": "visual", "line_char_end": 110, "line_char_start": 88, "line_index": 4, "line_text": "16A-L\t2.AA-[\t16cL\t8ee-", "operation": "replace", "pred_field": "16A-LL", "pred_local_replacement_line": "16A-LL\t2.AA-[\t16cL\t8ee-", "pred_note_token": "16A-LL", "pred_piece": "16A-LL", "pred_token_id": 1109, "pred_trimmed_index": 24, "visually_grounded": true}, "metrics": {"CER": 9.421422986708366, "LER": 40.74074172973633, "SER": 9.454061251664447}, "nll": 0.5540184378623962, "top1_error_count": 142, "top1_error_rate": 0.09447771124417831}}
+{"dataset_index": 27, "sample_id": "polish_00040", "source_ids": [], "target_token_count": 1109, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 8.547430830039527, "LER": 44.318180084228516, "SER": 10.649819494584838}, "nll": 0.5462185144424438, "top1_error_count": 118, "top1_error_rate": 0.10640216411181244}}
+{"dataset_index": 28, "sample_id": "polish_00043", "source_ids": [], "target_token_count": 885, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "insert", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 14.143775569842198, "LER": 53.757225036621094, "SER": 17.307692307692307}, "nll": 0.9757447242736816, "top1_error_count": 153, "top1_error_rate": 0.17288135593220338}}
+{"dataset_index": 29, "sample_id": "polish_00044", "source_ids": [], "target_token_count": 677, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 74, "char_start": 73, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 82, "field_char_start": 70, "field_index": 1, "field_text": "16f( 16dd(Jk", "gold_field": "16f( 16dd(Jk", "gold_line": "16B-(Jk\t16f( 16dd(Jk", "gold_note_token": "16f(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 22, "gold_trimmed_index": 22, "group": "visual", "line_char_end": 82, "line_char_start": 62, "line_index": 4, "line_text": "16B-(Jk\t16f( 16dd(Jk", "operation": "delete", "pred_field": "16f  16dd(Jk", "pred_local_replacement_line": "16B-(Jk\t16f  16dd(Jk", "pred_note_token": "16f", "pred_piece": " ", "pred_token_id": 6, "pred_trimmed_index": 22, "visually_grounded": true}, "metrics": {"CER": 8.61132660977502, "LER": 47.10743713378906, "SER": 10.502958579881657}, "nll": 0.6614680290222168, "top1_error_count": 71, "top1_error_rate": 0.10487444608567208}}
+{"dataset_index": 30, "sample_id": "polish_00046", "source_ids": [], "target_token_count": 1069, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "clef", "char_end": 15, "char_start": 8, "component_category": "clef", "component_flags": [], "component_group": "visual", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2", "gold_note_token": null, "gold_piece": "*clefG2", "gold_token_id": 147, "gold_token_index": 2, "gold_trimmed_index": 2, "group": "visual", "line_char_end": 15, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2", "operation": "replace", "pred_field": "*clefG", "pred_local_replacement_line": "*clefF4\t*clefG", "pred_note_token": null, "pred_piece": "*clefG", "pred_token_id": 145, "pred_trimmed_index": 2, "visually_grounded": true}, "metrics": {"CER": 8.113695090439277, "LER": 39.2156867980957, "SER": 8.801498127340825}, "nll": 0.3846493065357208, "top1_error_count": 94, "top1_error_rate": 0.08793264733395698}}
+{"dataset_index": 31, "sample_id": "polish_00047", "source_ids": [], "target_token_count": 1243, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 66, "char_start": 63, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 68, "field_char_start": 63, "field_index": 1, "field_text": "8eeq(", "gold_field": "8eeq(", "gold_line": ".\t8eeq(\t.", "gold_note_token": "8eeq(", "gold_piece": "8ee", "gold_token_id": 84, "gold_token_index": 28, "gold_trimmed_index": 28, "group": "visual", "line_char_end": 70, "line_char_start": 61, "line_index": 4, "line_text": ".\t8eeq(\t.", "operation": "replace", "pred_field": "8ddq(", "pred_local_replacement_line": ".\t8ddq(\t.", "pred_note_token": "8ddq(", "pred_piece": "8dd", "pred_token_id": 79, "pred_trimmed_index": 28, "visually_grounded": true}, "metrics": {"CER": 5.326295585412668, "LER": 27.439023971557617, "SER": 5.394524959742351}, "nll": 0.34875616431236267, "top1_error_count": 67, "top1_error_rate": 0.053901850362027354}}
+{"dataset_index": 32, "sample_id": "polish_00048", "source_ids": [], "target_token_count": 1384, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 51, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 48, "field_index": 0, "field_text": "4c 4f", "gold_field": "4c 4f", "gold_line": "4c 4f\t12aa-(L", "gold_note_token": "4f", "gold_piece": "4f", "gold_token_id": 103, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 48, "line_index": 3, "line_text": "4c 4f\t12aa-(L", "operation": "replace", "pred_field": "4c 4a-", "pred_local_replacement_line": "4c 4a-\t12aa-(L", "pred_note_token": "4a-", "pred_piece": "4a-", "pred_token_id": 341, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 16.217287866772402, "LER": 49.8127326965332, "SER": 11.641359363702097}, "nll": 0.5684936046600342, "top1_error_count": 161, "top1_error_rate": 0.11632947976878613}}
+{"dataset_index": 33, "sample_id": "polish_00049", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 12.712712712712712, "LER": 40.9638557434082, "SER": 11.906677393403058}, "nll": 0.5594200491905212, "top1_error_count": 124, "top1_error_rate": 0.09967845659163987}}
+{"dataset_index": 34, "sample_id": "polish_00050", "source_ids": [], "target_token_count": 1242, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "replace", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 12.076172782164422, "LER": 51.87165832519531, "SER": 10.95890410958904}, "nll": 0.5614944100379944, "top1_error_count": 136, "top1_error_rate": 0.10950080515297907}}
+{"dataset_index": 35, "sample_id": "polish_00052", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 52, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 59, "field_char_start": 47, "field_index": 0, "field_text": "2.DD-( 2.D-(", "gold_field": "2.DD-( 2.D-(", "gold_line": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "gold_note_token": "2.DD-(", "gold_piece": "-", "gold_token_id": 13, "gold_token_index": 13, "gold_trimmed_index": 13, "group": "visual", "line_char_end": 74, "line_char_start": 47, "line_index": 3, "line_text": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "operation": "replace", "pred_field": "2.DD-[( 2.D-(", "pred_local_replacement_line": "2.DD-[( 2.D-(\t8ff(L\t2.f 2.a-", "pred_note_token": "2.DD-[(", "pred_piece": "-[", "pred_token_id": 429, "pred_trimmed_index": 13, "visually_grounded": true}, "metrics": {"CER": 10.311614730878187, "LER": 45.35519027709961, "SER": 13.477366255144032}, "nll": 0.781216561794281, "top1_error_count": 131, "top1_error_rate": 0.13463514902363824}}
+{"dataset_index": 36, "sample_id": "polish_00053", "source_ids": [], "target_token_count": 1018, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 608, "char_start": 607, "component_category": "chord_structure_and_accidental_and_articulation", "component_flags": ["accidental", "articulation", "chord_structure"], "component_group": "visual", "field_char_end": 614, "field_char_start": 605, "field_index": 1, "field_text": "4d 4f# 4b", "gold_field": "4d 4f# 4b", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": "4d", "gold_piece": " ", "gold_token_id": 6, "gold_token_index": 353, "gold_trimmed_index": 353, "group": "visual", "line_char_end": 614, "line_char_start": 601, "line_index": 65, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4d`4f# 4b", "pred_local_replacement_line": "4BB\t4d`4f# 4b", "pred_note_token": "4d`4f#", "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 353, "visually_grounded": true}, "metrics": {"CER": 3.9809863339275102, "LER": 16.022098541259766, "SER": 4.424778761061947}, "nll": 0.2773802876472473, "top1_error_count": 45, "top1_error_rate": 0.04420432220039293}}
+{"dataset_index": 37, "sample_id": "polish_00054", "source_ids": [], "target_token_count": 689, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 46, "field_index": 0, "field_text": "8B-( 8d(L", "gold_field": "8B-( 8d(L", "gold_line": "8B-( 8d(L\t8fL", "gold_note_token": "8d(L", "gold_piece": "8d(L", "gold_token_id": 1103, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 59, "line_char_start": 46, "line_index": 3, "line_text": "8B-( 8d(L\t8fL", "operation": "replace", "pred_field": "8B-( 8dL", "pred_local_replacement_line": "8B-( 8dL\t8fL", "pred_note_token": "8dL", "pred_piece": "8dL", "pred_token_id": 206, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 9.815950920245399, "LER": 39.70588302612305, "SER": 11.773255813953488}, "nll": 0.7788732647895813, "top1_error_count": 81, "top1_error_rate": 0.11756168359941944}}
+{"dataset_index": 38, "sample_id": "polish_00056", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 45, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 45, "field_char_start": 42, "field_index": 0, "field_text": "8FL", "gold_field": "8FL", "gold_line": "8FL\t4cc 4ccc", "gold_note_token": null, "gold_piece": "8FL", "gold_token_id": 433, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 54, "line_char_start": 42, "line_index": 2, "line_text": "8FL\t4cc 4ccc", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t4cc 4ccc", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 14.988814317673377, "LER": 46.54377746582031, "SER": 14.412811387900357}, "nll": 0.6219017505645752, "top1_error_count": 162, "top1_error_rate": 0.144}}
+{"dataset_index": 39, "sample_id": "polish_00057", "source_ids": [], "target_token_count": 733, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 52, "component_category": "duration_and_pitch_and_accidental", "component_flags": ["duration", "pitch", "accidental"], "component_group": "visual", "field_char_end": 57, "field_char_start": 52, "field_index": 1, "field_text": "8ff#q", "gold_field": "8ff#q", "gold_line": ".\t8ff#q", "gold_note_token": "8ff#q", "gold_piece": "8ff#", "gold_token_id": 92, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": ".\t8ff#q", "operation": "replace", "pred_field": "ee-q", "pred_local_replacement_line": ".\tee-q", "pred_note_token": "ee-q", "pred_piece": "ee-", "pred_token_id": 1672, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 13.175675675675674, "LER": 47.55244827270508, "SER": 16.120218579234972}, "nll": 0.942428469657898, "top1_error_count": 118, "top1_error_rate": 0.16098226466575716}}
+{"dataset_index": 40, "sample_id": "polish_00058", "source_ids": [], "target_token_count": 900, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "rest", "char_end": 178, "char_start": 175, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 178, "field_char_start": 175, "field_index": 1, "field_text": "16r", "gold_field": "16r", "gold_line": "8D 8F 8A 8dL\t16r", "gold_note_token": "16r", "gold_piece": "16r", "gold_token_id": 168, "gold_token_index": 72, "gold_trimmed_index": 72, "group": "visual", "line_char_end": 178, "line_char_start": 162, "line_index": 12, "line_text": "8D 8F 8A 8dL\t16r", "operation": "replace", "pred_field": "8r", "pred_local_replacement_line": "8D 8F 8A 8dL\t8r", "pred_note_token": "8r", "pred_piece": "8r", "pred_token_id": 91, "pred_trimmed_index": 72, "visually_grounded": true}, "metrics": {"CER": 6.930091185410335, "LER": 32.85714340209961, "SER": 7.786429365962181}, "nll": 0.5073192119598389, "top1_error_count": 70, "top1_error_rate": 0.07777777777777778}}
+{"dataset_index": 41, "sample_id": "polish_00059", "source_ids": [], "target_token_count": 1838, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 60, "char_start": 56, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 60, "field_char_start": 56, "field_index": 1, "field_text": "8AAL", "gold_field": "8AAL", "gold_line": "4.AA\t8AAL\t24aa", "gold_note_token": "8AAL", "gold_piece": "8AAL", "gold_token_id": 512, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 65, "line_char_start": 51, "line_index": 4, "line_text": "4.AA\t8AAL\t24aa", "operation": "replace", "pred_field": "8.AAL", "pred_local_replacement_line": "4.AA\t8.AAL\t24aa", "pred_note_token": "8.AAL", "pred_piece": "8.AAL", "pred_token_id": 2335, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 12.548667265648398, "LER": 66.02316284179688, "SER": 12.955906369080022}, "nll": 0.7020101547241211, "top1_error_count": 238, "top1_error_rate": 0.1294885745375408}}
+{"dataset_index": 42, "sample_id": "polish_00060", "source_ids": [], "target_token_count": 1528, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 62, "char_start": 59, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 66, "field_char_start": 59, "field_index": 0, "field_text": "8FF 8FL", "gold_field": "8FF 8FL", "gold_line": "8FF 8FL\t2.F\t4.f", "gold_note_token": "8FF", "gold_piece": "8FF", "gold_token_id": 276, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 74, "line_char_start": 59, "line_index": 4, "line_text": "8FF 8FL\t2.F\t4.f", "operation": "replace", "pred_field": "8FL 8FL", "pred_local_replacement_line": "8FL 8FL\t2.F\t4.f", "pred_note_token": "8FL", "pred_piece": "8FL", "pred_token_id": 433, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 5.952380952380952, "LER": 42.52336502075195, "SER": 7.7930582842174205}, "nll": 0.5149219632148743, "top1_error_count": 119, "top1_error_rate": 0.07787958115183247}}
+{"dataset_index": 43, "sample_id": "polish_00061", "source_ids": [], "target_token_count": 653, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 50, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 59, "field_char_start": 50, "field_index": 0, "field_text": "8e-( 8g(L", "gold_field": "8e-( 8g(L", "gold_line": "8e-( 8g(L\t8gg 8bb-L", "gold_note_token": null, "gold_piece": "8e-", "gold_token_id": 263, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 69, "line_char_start": 50, "line_index": 3, "line_text": "8e-( 8g(L\t8gg 8bb-L", "operation": "replace", "pred_field": "*( 8g(L", "pred_local_replacement_line": "*( 8g(L\t8gg 8bb-L", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 12.053571428571429, "LER": 47.95918273925781, "SER": 16.87116564417178}, "nll": 0.807460367679596, "top1_error_count": 110, "top1_error_rate": 0.16845329249617153}}
+{"dataset_index": 44, "sample_id": "polish_00062", "source_ids": [], "target_token_count": 781, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 10.130718954248366, "LER": 37.3831787109375, "SER": 8.461538461538462}, "nll": 0.46052998304367065, "top1_error_count": 66, "top1_error_rate": 0.08450704225352113}}
+{"dataset_index": 45, "sample_id": "polish_00063", "source_ids": [], "target_token_count": 963, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 54, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 48, "field_index": 1, "field_text": "16qqddnL", "gold_field": "16qqddnL", "gold_line": ".\t16qqddnL", "gold_note_token": "16qqddnL", "gold_piece": "nL", "gold_token_id": 403, "gold_token_index": 11, "gold_trimmed_index": 11, "group": "visual", "line_char_end": 56, "line_char_start": 46, "line_index": 2, "line_text": ".\t16qqddnL", "operation": "replace", "pred_field": "16qqdd#LL", "pred_local_replacement_line": ".\t16qqdd#LL", "pred_note_token": "16qqdd#LL", "pred_piece": "#LL", "pred_token_id": 269, "pred_trimmed_index": 11, "visually_grounded": true}, "metrics": {"CER": 4.23572744014733, "LER": 28.289474487304688, "SER": 6.237006237006238}, "nll": 0.34785228967666626, "top1_error_count": 60, "top1_error_rate": 0.06230529595015576}}
+{"dataset_index": 46, "sample_id": "polish_00064", "source_ids": [], "target_token_count": 928, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 53, "char_start": 49, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 53, "field_char_start": 49, "field_index": 1, "field_text": "8dd[", "gold_field": "8dd[", "gold_line": "8r\t8dd[", "gold_note_token": "8dd[", "gold_piece": "8dd[", "gold_token_id": 1303, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 53, "line_char_start": 46, "line_index": 3, "line_text": "8r\t8dd[", "operation": "replace", "pred_field": "8dd(", "pred_local_replacement_line": "8r\t8dd(", "pred_note_token": "8dd(", "pred_piece": "8dd(", "pred_token_id": 1124, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 9.491137793024587, "LER": 45.19773864746094, "SER": 13.268608414239482}, "nll": 0.8054726123809814, "top1_error_count": 123, "top1_error_rate": 0.13254310344827586}}
+{"dataset_index": 47, "sample_id": "polish_00067", "source_ids": [], "target_token_count": 1138, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 43, "char_start": 39, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 43, "field_char_start": 39, "field_index": 1, "field_text": "16aL", "gold_field": "16aL", "gold_line": "8F(L\t16aL", "gold_note_token": "16aL", "gold_piece": "16aL", "gold_token_id": 602, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 43, "line_char_start": 34, "line_index": 2, "line_text": "8F(L\t16aL", "operation": "replace", "pred_field": "16aLL", "pred_local_replacement_line": "8F(L\t16aLL", "pred_note_token": "16aLL", "pred_piece": "16aLL", "pred_token_id": 338, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 12.06977840641207, "LER": 44.79166793823242, "SER": 14.775725593667547}, "nll": 0.9386909008026123, "top1_error_count": 167, "top1_error_rate": 0.14674868189806678}}
+{"dataset_index": 48, "sample_id": "polish_00068", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 57, "char_start": 53, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 57, "field_char_start": 53, "field_index": 1, "field_text": "4ff#", "gold_field": "4ff#", "gold_line": "4r\t4ff#", "gold_note_token": "4ff#", "gold_piece": "4ff#", "gold_token_id": 225, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": "4r\t4ff#", "operation": "replace", "pred_field": "4ff#(", "pred_local_replacement_line": "4r\t4ff#(", "pred_note_token": "4ff#(", "pred_piece": "4ff#(", "pred_token_id": 2286, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 22.732997481108313, "LER": 65.19336700439453, "SER": 20.473251028806587}, "nll": 0.993992805480957, "top1_error_count": 199, "top1_error_rate": 0.20452209660842754}}
+{"dataset_index": 49, "sample_id": "polish_00071", "source_ids": [], "target_token_count": 1433, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 48, "char_start": 46, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 48, "field_char_start": 46, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*^", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 51, "line_char_start": 46, "line_index": 2, "line_text": "*^\t*^", "operation": "replace", "pred_field": ".", "pred_local_replacement_line": ".\t*^", "pred_note_token": null, "pred_piece": ".", "pred_token_id": 14, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 10.909856781802864, "LER": 59.562843322753906, "SER": 13.268156424581006}, "nll": 0.6887483596801758, "top1_error_count": 190, "top1_error_rate": 0.13258897418004187}}
+{"dataset_index": 50, "sample_id": "polish_00072", "source_ids": [], "target_token_count": 1123, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 56, "char_start": 55, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 13, "gold_trimmed_index": 13, "group": "structural", "line_char_end": 65, "line_char_start": 52, "line_index": 3, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 13, "visually_grounded": false}, "metrics": {"CER": 4.740292486132123, "LER": 29.255319595336914, "SER": 6.506238859180035}, "nll": 0.2940332889556885, "top1_error_count": 73, "top1_error_rate": 0.06500445235975066}}
+{"dataset_index": 51, "sample_id": "polish_00074", "source_ids": [], "target_token_count": 1355, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 34, "char_start": 30, "component_category": "duration_and_pitch_and_accidental", "component_flags": ["duration", "pitch", "accidental"], "component_group": "visual", "field_char_end": 36, "field_char_start": 30, "field_index": 0, "field_text": "8BBnq(", "gold_field": "8BBnq(", "gold_line": "8BBnq(\t.", "gold_note_token": "8BBnq(", "gold_piece": "8BBn", "gold_token_id": 2042, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 38, "line_char_start": 30, "line_index": 2, "line_text": "8BBnq(\t.", "operation": "replace", "pred_field": "=!|:q(", "pred_local_replacement_line": "=!|:q(\t.", "pred_note_token": "=!|:q(", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 8.788282290279627, "LER": 42.56410217285156, "SER": 9.30576070901034}, "nll": 0.46926558017730713, "top1_error_count": 126, "top1_error_rate": 0.0929889298892989}}
+{"dataset_index": 52, "sample_id": "polish_00078", "source_ids": [], "target_token_count": 1172, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 49, "char_start": 48, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 49, "field_char_start": 48, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*\t*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 56, "line_char_start": 48, "line_index": 3, "line_text": "*\t*\t*^\t*", "operation": "insert", "pred_field": "4Dn", "pred_local_replacement_line": "4Dn\t*\t*^\t*", "pred_note_token": null, "pred_piece": "4Dn", "pred_token_id": 2294, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 16.15837346174425, "LER": 44.155845642089844, "SER": 12.5533731853117}, "nll": 0.5672224760055542, "top1_error_count": 147, "top1_error_rate": 0.12542662116040956}}
+{"dataset_index": 53, "sample_id": "polish_00082", "source_ids": [], "target_token_count": 1027, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "8AA-L", "pred_local_replacement_line": "8AA-L\t*^", "pred_note_token": null, "pred_piece": "8AA-L", "pred_token_id": 901, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 15.005537098560353, "LER": 51.31578826904297, "SER": 10.33138401559454}, "nll": 0.5895283222198486, "top1_error_count": 106, "top1_error_rate": 0.10321324245374879}}
+{"dataset_index": 54, "sample_id": "polish_00084", "source_ids": [], "target_token_count": 860, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 54, "char_start": 50, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 57, "field_char_start": 50, "field_index": 0, "field_text": "32qqCCL", "gold_field": "32qqCCL", "gold_line": "32qqCCL\t32qqe-L", "gold_note_token": "32qqCCL", "gold_piece": "32qq", "gold_token_id": 1235, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 65, "line_char_start": 50, "line_index": 3, "line_text": "32qqCCL\t32qqe-L", "operation": "replace", "pred_field": "32CCCCL", "pred_local_replacement_line": "32CCCCL\t32qqe-L", "pred_note_token": "32CCCCL", "pred_piece": "32CC", "pred_token_id": 1559, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 10.975609756097562, "LER": 43.312103271484375, "SER": 13.154831199068685}, "nll": 0.771187424659729, "top1_error_count": 113, "top1_error_rate": 0.1313953488372093}}
+{"dataset_index": 55, "sample_id": "polish_00085", "source_ids": [], "target_token_count": 1428, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 54, "char_start": 49, "component_category": "duration_and_tie_slur", "component_flags": ["duration", "tie_slur"], "component_group": "visual", "field_char_end": 54, "field_char_start": 49, "field_index": 2, "field_text": "8aa(L", "gold_field": "8aa(L", "gold_line": "4d\t4D 4A\t8aa(L\t4f# 4dd", "gold_note_token": "8aa(L", "gold_piece": "8aa(L", "gold_token_id": 1057, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 40, "line_index": 3, "line_text": "4d\t4D 4A\t8aa(L\t4f# 4dd", "operation": "replace", "pred_field": "8.aa", "pred_local_replacement_line": "4d\t4D 4A\t8.aa\t4f# 4dd", "pred_note_token": "8.aa", "pred_piece": "8.aa", "pred_token_id": 386, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 4.705882352941177, "LER": 23.60248374938965, "SER": 4.204625087596356}, "nll": 0.24162887036800385, "top1_error_count": 60, "top1_error_rate": 0.04201680672268908}}
+{"dataset_index": 56, "sample_id": "polish_00086", "source_ids": [], "target_token_count": 1170, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 44, "field_char_start": 42, "field_index": 0, "field_text": "4C", "gold_field": "4C", "gold_line": "4C\t4r", "gold_note_token": null, "gold_piece": "4C", "gold_token_id": 181, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 47, "line_char_start": 42, "line_index": 2, "line_text": "4C\t4r", "operation": "replace", "pred_field": "*M4/4", "pred_local_replacement_line": "*M4/4\t4r", "pred_note_token": null, "pred_piece": "*M4/4", "pred_token_id": 229, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 9.971374045801527, "LER": 40.85365676879883, "SER": 9.067579127459368}, "nll": 0.4045732617378235, "top1_error_count": 106, "top1_error_rate": 0.0905982905982906}}
+{"dataset_index": 57, "sample_id": "polish_00087", "source_ids": [], "target_token_count": 1300, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 47, "char_start": 44, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 56, "field_char_start": 44, "field_index": 1, "field_text": "8ccq( 8ee-q(", "gold_field": "8ccq( 8ee-q(", "gold_line": ".\t8ccq( 8ee-q(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 56, "line_char_start": 42, "line_index": 2, "line_text": ".\t8ccq( 8ee-q(", "operation": "replace", "pred_field": "8ddq( 8ee-q(", "pred_local_replacement_line": ".\t8ddq( 8ee-q(", "pred_note_token": "8ddq(", "pred_piece": "8dd", "pred_token_id": 79, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 13.968668407310705, "LER": 55.9782600402832, "SER": 12.16320246343341}, "nll": 0.5738366842269897, "top1_error_count": 158, "top1_error_rate": 0.12153846153846154}}
+{"dataset_index": 58, "sample_id": "polish_00089", "source_ids": [], "target_token_count": 782, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 37, "char_start": 35, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 43, "field_char_start": 35, "field_index": 0, "field_text": "8A( 8c(L", "gold_field": "8A( 8c(L", "gold_line": "8A( 8c(L\t2.D\t16g#(L", "gold_note_token": "8A(", "gold_piece": "8A", "gold_token_id": 109, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 35, "line_index": 3, "line_text": "8A( 8c(L\t2.D\t16g#(L", "operation": "replace", "pred_field": "8A(L( 8c(L", "pred_local_replacement_line": "8A(L( 8c(L\t2.D\t16g#(L", "pred_note_token": "8A(L(", "pred_piece": "8A(L", "pred_token_id": 1848, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 12.407282535401214, "LER": 48.44720458984375, "SER": 14.980793854033292}, "nll": 1.0558654069900513, "top1_error_count": 117, "top1_error_rate": 0.149616368286445}}
+{"dataset_index": 59, "sample_id": "polish_00090", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 276, "char_start": 272, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 276, "field_char_start": 272, "field_index": 0, "field_text": "8AAL", "gold_field": "8AAL", "gold_line": "8AAL\t8.r\t4ee", "gold_note_token": "8AAL", "gold_piece": "8AAL", "gold_token_id": 512, "gold_token_index": 164, "gold_trimmed_index": 164, "group": "visual", "line_char_end": 284, "line_char_start": 272, "line_index": 22, "line_text": "8AAL\t8.r\t4ee", "operation": "replace", "pred_field": "8AA", "pred_local_replacement_line": "8AA\t8.r\t4ee", "pred_note_token": "8AA", "pred_piece": "8AA", "pred_token_id": 210, "pred_trimmed_index": 164, "visually_grounded": true}, "metrics": {"CER": 9.105960264900661, "LER": 36.24161148071289, "SER": 7.6512455516014235}, "nll": 0.41781187057495117, "top1_error_count": 86, "top1_error_rate": 0.07644444444444444}}
+{"dataset_index": 60, "sample_id": "polish_00092", "source_ids": [], "target_token_count": 1153, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 62, "char_start": 61, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 62, "field_char_start": 52, "field_index": 1, "field_text": "4d-( 4dd-(", "gold_field": "4d-( 4dd-(", "gold_line": "8BB-\t4d-( 4dd-(", "gold_note_token": "4dd-(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 47, "line_index": 3, "line_text": "8BB-\t4d-( 4dd-(", "operation": "replace", "pred_field": "4d-( 4dd-\n", "pred_local_replacement_line": "8BB-\t4d-( 4dd-\n", "pred_note_token": "4dd-\n", "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 5.614714424007745, "LER": 34.07821273803711, "SER": 7.378472222222222}, "nll": 0.5313606262207031, "top1_error_count": 85, "top1_error_rate": 0.07372072853425846}}
+{"dataset_index": 61, "sample_id": "polish_00093", "source_ids": [], "target_token_count": 1047, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 555, "char_start": 552, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 564, "field_char_start": 552, "field_index": 1, "field_text": "8e[ 8b[ 8ee[", "gold_field": "8e[ 8b[ 8ee[", "gold_line": "8r\t8e[ 8b[ 8ee[\t8ee", "gold_note_token": "8e[", "gold_piece": "8e[", "gold_token_id": 1374, "gold_token_index": 340, "gold_trimmed_index": 340, "group": "visual", "line_char_end": 568, "line_char_start": 549, "line_index": 45, "line_text": "8r\t8e[ 8b[ 8ee[\t8ee", "operation": "replace", "pred_field": "8e( 8b[ 8ee[", "pred_local_replacement_line": "8r\t8e( 8b[ 8ee[\t8ee", "pred_note_token": "8e(", "pred_piece": "8e(", "pred_token_id": 1873, "pred_trimmed_index": 340, "visually_grounded": true}, "metrics": {"CER": 8.870490833826139, "LER": 32.84671401977539, "SER": 7.45697896749522}, "nll": 0.3665979504585266, "top1_error_count": 78, "top1_error_rate": 0.07449856733524356}}
+{"dataset_index": 62, "sample_id": "polish_00094", "source_ids": [], "target_token_count": 773, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 40, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 47, "field_char_start": 40, "field_index": 1, "field_text": "32qqggL", "gold_field": "32qqggL", "gold_line": ".\t32qqggL", "gold_note_token": "32qqggL", "gold_piece": "32qq", "gold_token_id": 1235, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 47, "line_char_start": 38, "line_index": 2, "line_text": ".\t32qqggL", "operation": "replace", "pred_field": "32ggggL", "pred_local_replacement_line": ".\t32ggggL", "pred_note_token": "32ggggL", "pred_piece": "32gg", "pred_token_id": 477, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 9.932432432432432, "LER": 45.16128921508789, "SER": 12.305699481865284}, "nll": 0.8049496412277222, "top1_error_count": 95, "top1_error_rate": 0.12289780077619664}}
+{"dataset_index": 63, "sample_id": "polish_00095", "source_ids": [], "target_token_count": 1073, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "4D-", "pred_local_replacement_line": "4D-\t*^", "pred_note_token": null, "pred_piece": "4D-", "pred_token_id": 556, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 18.315018315018314, "LER": 44.845359802246094, "SER": 13.339552238805972}, "nll": 0.697178065776825, "top1_error_count": 143, "top1_error_rate": 0.13327120223671948}}
+{"dataset_index": 64, "sample_id": "polish_00096", "source_ids": [], "target_token_count": 935, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 52, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "2en", "gold_field": "2en", "gold_line": "2en\t4F#]\t4ggn\t2cc#", "gold_note_token": "2en", "gold_piece": "2en", "gold_token_id": 2374, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 70, "line_char_start": 52, "line_index": 3, "line_text": "2en\t4F#]\t4ggn\t2cc#", "operation": "replace", "pred_field": "2e", "pred_local_replacement_line": "2e\t4F#]\t4ggn\t2cc#", "pred_note_token": "2e", "pred_piece": "2e", "pred_token_id": 296, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 4.510309278350515, "LER": 36.42856979370117, "SER": 5.460385438972163}, "nll": 0.3092499077320099, "top1_error_count": 51, "top1_error_rate": 0.05454545454545454}}
+{"dataset_index": 65, "sample_id": "polish_00097", "source_ids": [], "target_token_count": 1050, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 61, "char_start": 57, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 63, "field_char_start": 57, "field_index": 1, "field_text": "16gg[L", "gold_field": "16gg[L", "gold_line": "8c 8eJ\t16gg[L\t16dd(L", "gold_note_token": "16gg[L", "gold_piece": "16gg", "gold_token_id": 146, "gold_token_index": 22, "gold_trimmed_index": 22, "group": "visual", "line_char_end": 70, "line_char_start": 50, "line_index": 3, "line_text": "8c 8eJ\t16gg[L\t16dd(L", "operation": "replace", "pred_field": "16ggLL[L", "pred_local_replacement_line": "8c 8eJ\t16ggLL[L\t16dd(L", "pred_note_token": "16ggLL[L", "pred_piece": "16ggLL", "pred_token_id": 382, "pred_trimmed_index": 22, "visually_grounded": true}, "metrics": {"CER": 7.276007215874925, "LER": 31.34328269958496, "SER": 6.48236415633937}, "nll": 0.3703092634677887, "top1_error_count": 68, "top1_error_rate": 0.06476190476190476}}
+{"dataset_index": 66, "sample_id": "polish_00100", "source_ids": [], "target_token_count": 1185, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 36, "char_start": 34, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 36, "field_char_start": 34, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "16DD", "pred_local_replacement_line": "16DD\t*", "pred_note_token": null, "pred_piece": "16DD", "pred_token_id": 526, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 13.939081053175013, "LER": 52.91005325317383, "SER": 13.429054054054054}, "nll": 0.7050472497940063, "top1_error_count": 159, "top1_error_rate": 0.1341772151898734}}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/summary.json

@@ -0,0 +1,454 @@
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_3.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.3157254796001081,
+      "errors": 2337,
+      "errors_per_1k_tokens": 31.36281285647185,
+      "errors_per_page_mean": 34.88059701492537,
+      "errors_per_page_median": 35.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.22466900837611456,
+      "errors": 1663,
+      "errors_per_1k_tokens": 22.31765416359122,
+      "errors_per_page_mean": 24.82089552238806,
+      "errors_per_page_median": 25.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.10497162928938125,
+      "errors": 777,
+      "errors_per_1k_tokens": 10.427430718647253,
+      "errors_per_page_mean": 11.597014925373134,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05498513915158065,
+      "errors": 407,
+      "errors_per_1k_tokens": 5.461987519291418,
+      "errors_per_page_mean": 6.074626865671642,
+      "errors_per_page_median": 5.0,
+      "page_share": 0.9104477611940298,
+      "pages_with_error": 61
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04228586868413942,
+      "errors": 313,
+      "errors_per_1k_tokens": 4.200496544319936,
+      "errors_per_page_mean": 4.6716417910447765,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03485544447446636,
+      "errors": 258,
+      "errors_per_1k_tokens": 3.4623901227940683,
+      "errors_per_page_mean": 3.8507462686567164,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.7611940298507462,
+      "pages_with_error": 51
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.03390975412050797,
+      "errors": 251,
+      "errors_per_1k_tokens": 3.3684493055089577,
+      "errors_per_page_mean": 3.746268656716418,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.835820895522388,
+      "pages_with_error": 56
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.026074034044852742,
+      "errors": 193,
+      "errors_per_1k_tokens": 2.5900825337180433,
+      "errors_per_page_mean": 2.8805970149253732,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.582089552238806,
+      "pages_with_error": 39
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.023372061604971628,
+      "errors": 173,
+      "errors_per_1k_tokens": 2.321680198617728,
+      "errors_per_page_mean": 2.582089552238806,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5074626865671642,
+      "pages_with_error": 34
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.023236962982977574,
+      "errors": 172,
+      "errors_per_1k_tokens": 2.3082600818627124,
+      "errors_per_page_mean": 2.5671641791044775,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.014860848419346123,
+      "errors": 110,
+      "errors_per_1k_tokens": 1.4762128430517345,
+      "errors_per_page_mean": 1.6417910447761195,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.012293974601459066,
+      "errors": 91,
+      "errors_per_1k_tokens": 1.2212306247064348,
+      "errors_per_page_mean": 1.3582089552238805,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.582089552238806,
+      "pages_with_error": 39
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.012158875979465009,
+      "errors": 90,
+      "errors_per_1k_tokens": 1.2078105079514192,
+      "errors_per_page_mean": 1.3432835820895523,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.373134328358209,
+      "pages_with_error": 25
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.008511213185625506,
+      "errors": 63,
+      "errors_per_1k_tokens": 0.8454673555659934,
+      "errors_per_page_mean": 0.9402985074626866,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.31343283582089554,
+      "pages_with_error": 21
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.008241015941637396,
+      "errors": 61,
+      "errors_per_1k_tokens": 0.8186271220559619,
+      "errors_per_page_mean": 0.9104477611940298,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.44776119402985076,
+      "pages_with_error": 30
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.007565522831667117,
+      "errors": 56,
+      "errors_per_1k_tokens": 0.751526538280883,
+      "errors_per_page_mean": 0.835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.373134328358209,
+      "pages_with_error": 25
+    },
+    {
+      "category": "termination",
+      "error_share": 0.005944339367738449,
+      "errors": 44,
+      "errors_per_1k_tokens": 0.5904851372206938,
+      "errors_per_page_mean": 0.6567164179104478,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.47761194029850745,
+      "pages_with_error": 32
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.005539043501756282,
+      "errors": 41,
+      "errors_per_1k_tokens": 0.5502247869556465,
+      "errors_per_page_mean": 0.6119402985074627,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3880597014925373,
+      "pages_with_error": 26
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.0045933531477978925,
+      "errors": 34,
+      "errors_per_1k_tokens": 0.45628396967053614,
+      "errors_per_page_mean": 0.5074626865671642,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.004323155903809781,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.26865671641791045,
+      "pages_with_error": 18
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.004323155903809781,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.004323155903809781,
+      "errors": 32,
+      "errors_per_1k_tokens": 0.4294437361605046,
+      "errors_per_page_mean": 0.47761194029850745,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0032423669278573357,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0032423669278573357,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.0031072683058632803,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.14925373134328357,
+      "pages_with_error": 10
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0031072683058632803,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.0029721696838692245,
+      "errors": 22,
+      "errors_per_1k_tokens": 0.2952425686103469,
+      "errors_per_page_mean": 0.3283582089552239,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0025668738178870575,
+      "errors": 19,
+      "errors_per_1k_tokens": 0.2549822183452996,
+      "errors_per_page_mean": 0.2835820895522388,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1044776119402985,
+      "pages_with_error": 7
+    },
+    {
+      "category": "clef",
+      "error_share": 0.0022966765738989463,
+      "errors": 17,
+      "errors_per_1k_tokens": 0.22814198483526807,
+      "errors_per_page_mean": 0.2537313432835821,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.16417910447761194,
+      "pages_with_error": 11
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.0013509862199405566,
+      "errors": 10,
+      "errors_per_1k_tokens": 0.13420116755015768,
+      "errors_per_page_mean": 0.14925373134328357,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0005403944879762226,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.00040529586598216696,
+      "errors": 3,
+      "errors_per_1k_tokens": 0.040260350265047304,
+      "errors_per_page_mean": 0.04477611940298507,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.04477611940298507,
+      "pages_with_error": 3
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0002701972439881113,
+      "errors": 2,
+      "errors_per_1k_tokens": 0.026840233510031536,
+      "errors_per_page_mean": 0.029850746268656716,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.029850746268656716,
+      "pages_with_error": 2
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.00013509862199405565,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 4,
+    "accidental_and_tie_slur": 41,
+    "accidental_only": 313,
+    "articulation_only": 32,
+    "barline": 24,
+    "chord_structure_and_accidental": 2,
+    "chord_structure_and_accidental_and_articulation": 19,
+    "chord_structure_and_accidental_and_tie_slur": 22,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 90,
+    "chord_structure_and_tie_slur": 91,
+    "chord_structure_and_tie_slur_and_articulation": 10,
+    "chord_structure_only": 32,
+    "clef": 17,
+    "column_or_line_break": 407,
+    "duration_and_accidental": 34,
+    "duration_and_pitch": 110,
+    "duration_and_pitch_and_accidental": 56,
+    "duration_and_pitch_and_accidental_and_tie_slur": 3,
+    "duration_and_pitch_and_tie_slur": 23,
+    "duration_and_tie_slur": 63,
+    "duration_only": 777,
+    "interpretation_or_control": 258,
+    "note_other": 2337,
+    "null_token": 61,
+    "pitch_and_accidental": 193,
+    "pitch_and_accidental_and_tie_slur": 32,
+    "pitch_and_tie_slur": 23,
+    "pitch_only": 251,
+    "rest_note_confusion": 173,
+    "spine_operation": 172,
+    "termination": 44,
+    "tie_slur_and_articulation": 24,
+    "tie_slur_only": 1663
+  },
+  "component_error_groups": {
+    "structural": 942,
+    "visual": 6460
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 12.819884218974039,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3151,
+    "articulation": 54,
+    "barline": 23,
+    "clef": 17,
+    "column_or_line_break": 408,
+    "duration": 994,
+    "interpretation_or_control": 146,
+    "note_other": 209,
+    "null_token": 62,
+    "pitch": 21,
+    "rest": 281,
+    "spine_operation": 205,
+    "termination": 32,
+    "tie_slur": 1799
+  },
+  "gold_error_groups": {
+    "structural": 853,
+    "visual": 6549
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.608084678649902,
+      "LER": 42.074729919433594,
+      "SER": 10.15070915222168,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.09933570422062672,
+      "top1_errors": 7402,
+      "top1_errors_per_page": 110.4776119402985
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 38,
+    "predicted_unk_literal": 104
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/summary.md

@@ -0,0 +1,43 @@
+# Transcoda Dataset Eval
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_3.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Dataset kind: `hf_parquet_split`
+- Mode: `teacher_forced`
+- Pages: `67`
+- Tokens: `74,515`
+- Elapsed: `12.8s`
+
+## Metrics
+
+| mode | CER | SER | LER | exact pages | top-1 error rate | top-1 errors/page |
+|---|---:|---:|---:|---:|---:|---:|
+| `teacher_forced` | 9.608085 | 10.150709 | 42.074730 | 0.000% | 9.9336% | 110.478 |
+
+## Teacher-Forced Error Categories
+
+| category | errors |
+|---|---:|
+| `accidental` | 3,151 |
+| `tie_slur` | 1,799 |
+| `duration` | 994 |
+| `column_or_line_break` | 408 |
+| `rest` | 281 |
+| `note_other` | 209 |
+| `spine_operation` | 205 |
+| `interpretation_or_control` | 146 |
+| `null_token` | 62 |
+| `articulation` | 54 |
+| `termination` | 32 |
+| `barline` | 23 |
+| `pitch` | 21 |
+| `clef` | 17 |
+
+## Outputs
+
+- `summary.json`
+- `per_page_records.jsonl`
+- `teacher_forced_predictions.jsonl` when teacher-forced mode is enabled
+- `free_generation_predictions.jsonl` when free-generation mode is enabled
+- `gold_prefix_errors.jsonl` when teacher-forced mode is enabled
+- `teacher_forced_error_report.md` when teacher-forced mode is enabled

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_3/teacher_forced_error_report.md

@@ -0,0 +1,524 @@
+# Teacher-Forced Error Report
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_3.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Split: `eval`
+- Pages: `67`
+- Tokens: `74,515`
+- Top-1 errors: `7,402`
+- Top-1 error rate: `9.934%`
+
+## Component Buckets
+
+| bucket | errors | error share | mean/page | median/page | pages | errors/1k tok |
+|---|---:|---:|---:|---:|---:|---:|
+| `note_other` | 2,337 | 31.57% | 34.881 | 35.000 | 67 (100.0%) | 31.363 |
+| `tie_slur_only` | 1,663 | 22.47% | 24.821 | 25.000 | 67 (100.0%) | 22.318 |
+| `duration_only` | 777 | 10.50% | 11.597 | 3.000 | 61 (91.0%) | 10.427 |
+| `column_or_line_break` | 407 | 5.50% | 6.075 | 5.000 | 61 (91.0%) | 5.462 |
+| `accidental_only` | 313 | 4.23% | 4.672 | 3.000 | 62 (92.5%) | 4.200 |
+| `interpretation_or_control` | 258 | 3.49% | 3.851 | 2.000 | 51 (76.1%) | 3.462 |
+| `pitch_only` | 251 | 3.39% | 3.746 | 2.000 | 56 (83.6%) | 3.368 |
+| `pitch_and_accidental` | 193 | 2.61% | 2.881 | 1.000 | 39 (58.2%) | 2.590 |
+| `rest_note_confusion` | 173 | 2.34% | 2.582 | 1.000 | 34 (50.7%) | 2.322 |
+| `spine_operation` | 172 | 2.32% | 2.567 | 1.000 | 41 (61.2%) | 2.308 |
+| `duration_and_pitch` | 110 | 1.49% | 1.642 | 1.000 | 41 (61.2%) | 1.476 |
+| `chord_structure_and_tie_slur` | 91 | 1.23% | 1.358 | 1.000 | 39 (58.2%) | 1.221 |
+| `chord_structure_and_articulation` | 90 | 1.22% | 1.343 | 0.000 | 25 (37.3%) | 1.208 |
+| `duration_and_tie_slur` | 63 | 0.85% | 0.940 | 0.000 | 21 (31.3%) | 0.845 |
+| `null_token` | 61 | 0.82% | 0.910 | 0.000 | 30 (44.8%) | 0.819 |
+| `duration_and_pitch_and_accidental` | 56 | 0.76% | 0.836 | 0.000 | 25 (37.3%) | 0.752 |
+| `termination` | 44 | 0.59% | 0.657 | 0.000 | 32 (47.8%) | 0.590 |
+| `accidental_and_tie_slur` | 41 | 0.55% | 0.612 | 0.000 | 26 (38.8%) | 0.550 |
+| `duration_and_accidental` | 34 | 0.46% | 0.507 | 0.000 | 18 (26.9%) | 0.456 |
+| `pitch_and_accidental_and_tie_slur` | 32 | 0.43% | 0.478 | 0.000 | 18 (26.9%) | 0.429 |
+| `articulation_only` | 32 | 0.43% | 0.478 | 0.000 | 12 (17.9%) | 0.429 |
+| `chord_structure_only` | 32 | 0.43% | 0.478 | 0.000 | 17 (25.4%) | 0.429 |
+| `tie_slur_and_articulation` | 24 | 0.32% | 0.358 | 0.000 | 17 (25.4%) | 0.322 |
+| `barline` | 24 | 0.32% | 0.358 | 0.000 | 17 (25.4%) | 0.322 |
+| `duration_and_pitch_and_tie_slur` | 23 | 0.31% | 0.343 | 0.000 | 10 (14.9%) | 0.309 |
+| `pitch_and_tie_slur` | 23 | 0.31% | 0.343 | 0.000 | 17 (25.4%) | 0.309 |
+| `chord_structure_and_accidental_and_tie_slur` | 22 | 0.30% | 0.328 | 0.000 | 12 (17.9%) | 0.295 |
+| `chord_structure_and_accidental_and_articulation` | 19 | 0.26% | 0.284 | 0.000 | 7 (10.4%) | 0.255 |
+| `clef` | 17 | 0.23% | 0.254 | 0.000 | 11 (16.4%) | 0.228 |
+| `chord_structure_and_tie_slur_and_articulation` | 10 | 0.14% | 0.149 | 0.000 | 8 (11.9%) | 0.134 |
+| `accidental_and_articulation` | 4 | 0.05% | 0.060 | 0.000 | 4 (6.0%) | 0.054 |
+| `duration_and_pitch_and_accidental_and_tie_slur` | 3 | 0.04% | 0.045 | 0.000 | 3 (4.5%) | 0.040 |
+| `chord_structure_and_accidental` | 2 | 0.03% | 0.030 | 0.000 | 2 (3.0%) | 0.027 |
+| `chord_structure_and_accidental_and_tie_slur_and_articulation` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+
+## Legacy Regex Buckets
+
+These are retained for compatibility and should not be treated as component diffs.
+
+| bucket | errors |
+|---|---:|
+| `accidental` | 3,151 |
+| `tie_slur` | 1,799 |
+| `duration` | 994 |
+| `column_or_line_break` | 408 |
+| `rest` | 281 |
+| `note_other` | 209 |
+| `spine_operation` | 205 |
+| `interpretation_or_control` | 146 |
+| `null_token` | 62 |
+| `articulation` | 54 |
+| `termination` | 32 |
+| `barline` | 23 |
+| `pitch` | 21 |
+| `clef` | 17 |
+
+## Error Samples
+
+### Sample 1
+
+```text
+.	.	16ee)J	.
+teacher-forced top-1 local replacement:
+.	.	16ee)JJ	.
+```
+
+### Sample 2
+
+```text
+8B( 8d( 8f(L	2G
+teacher-forced top-1 local replacement:
+8B( 8d( 8fL	2G
+```
+
+### Sample 3
+
+```text
+4BBB	8.ddW( 8.ff#W(L
+teacher-forced top-1 local replacement:
+4BBB	8.ddW( 8.ff#WL
+```
+
+### Sample 4
+
+```text
+.	16qqddJ
+teacher-forced top-1 local replacement:
+.	16qqddJJ
+```
+
+### Sample 5
+
+```text
+4FFn 4C# 4B-	8e
+teacher-forced top-1 local replacement:
+4FFn:4C# 4B-	8e
+```
+
+### Sample 6
+
+```text
+8CL	8ff) 8ccc)	8cc
+teacher-forced top-1 local replacement:
+8CL	8ff) 8ccc		8cc
+```
+
+### Sample 7
+
+```text
+.	12ee--	6ee--
+teacher-forced top-1 local replacement:
+.	4ee--	6ee--
+```
+
+### Sample 8
+
+```text
+.	16ff##)J
+teacher-forced top-1 local replacement:
+.	16ff##)JJ
+```
+
+### Sample 9
+
+```text
+.	16a)J	.
+teacher-forced top-1 local replacement:
+.	16a))J	.
+```
+
+### Sample 10
+
+```text
+.	32a)J	.
+teacher-forced top-1 local replacement:
+.	32a)JJJ	.
+```
+
+### Sample 11
+
+```text
+4BBB- 4BB-	16g(L
+teacher-forced top-1 local replacement:
+4BBB- 4BB-	16g(LL(L
+```
+
+### Sample 12
+
+```text
+*clefF4	*clefG2
+teacher-forced top-1 local replacement:
+
+clefF4	*clefG2
+```
+
+### Sample 13
+
+```text
+8BB- 8B-L	8cc 8een
+teacher-forced top-1 local replacement:
+8BB- 8dL	8cc 8een
+```
+
+### Sample 14
+
+```text
+4C	4r	8ccc)L
+teacher-forced top-1 local replacement:
+4C	4r<unk>8ccc)L
+```
+
+### Sample 15
+
+```text
+.	16ee)J
+teacher-forced top-1 local replacement:
+.	16ee)JJ)J
+```
+
+### Sample 16
+
+```text
+.	.	16qqccL	.
+teacher-forced top-1 local replacement:
+.	.	16qqccLL	.
+```
+
+
+## Sample Metadata
+
+### Sample 1
+
+- Sample: `polish_00071`
+- Row/token: `49` / `1197`
+- Line/field: `157:2`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `)J` -> `)JJ`
+- Gold note -> top-1 note: `16ee)J` -> `16ee)JJ`
+- Gold token likelihood: rank `6`, p `0.124%`
+- Predicted token likelihood: p `93.450%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 305 | `")JJ"` | 93.450% |
+| 2 | 1570 | `")JJJ"` | 2.651% |
+| 3 | 10 | `")"` | 0.297% |
+| 4 | 2968 | `")JJk"` | 0.186% |
+| 5 | 741 | `"nJJ"` | 0.150% |
+
+### Sample 2
+
+- Sample: `polish_00084`
+- Row/token: `54` / `701`
+- Line/field: `124:0`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `8f(L` -> `8fL`
+- Gold note -> top-1 note: `8f(L` -> `8fL`
+- Gold token likelihood: rank `3`, p `0.070%`
+- Predicted token likelihood: p `98.681%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 265 | `"8fL"` | 98.681% |
+| 2 | 81 | `"8f"` | 0.077% |
+| 3 | 1384 | `"8f(L"` | 0.070% |
+| 4 | 502 | `"16fLL"` | 0.070% |
+| 5 | 640 | `"8.fL"` | 0.062% |
+
+### Sample 3
+
+- Sample: `polish_00025`
+- Row/token: `16` / `92`
+- Line/field: `15:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `(L` -> `L`
+- Gold note -> top-1 note: `8.ff#W(L` -> `8.ff#WL`
+- Gold token likelihood: rank `2`, p `0.120%`
+- Predicted token likelihood: p `99.209%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 38 | `"L"` | 99.209% |
+| 2 | 171 | `"(L"` | 0.120% |
+| 3 | 217 | `"KL"` | 0.099% |
+| 4 | 83 | `"LL"` | 0.088% |
+| 5 | 176 | `"#L"` | 0.048% |
+
+### Sample 4
+
+- Sample: `polish_00040`
+- Row/token: `27` / `838`
+- Line/field: `134:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `note_other`
+- Gold piece -> top-1 piece: `J` -> `JJ`
+- Gold note -> top-1 note: `16qqddJ` -> `16qqddJJ`
+- Gold token likelihood: rank `2`, p `1.911%`
+- Predicted token likelihood: p `92.079%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 80 | `"JJ"` | 92.079% |
+| 2 | 36 | `"J"` | 1.911% |
+| 3 | 363 | `"]JJ"` | 0.401% |
+| 4 | 258 | `"#JJ"` | 0.401% |
+| 5 | 741 | `"nJJ"` | 0.343% |
+
+### Sample 5
+
+- Sample: `polish_00017`
+- Row/token: `10` / `972`
+- Line/field: `164:0`
+- Component bucket: `chord_structure_only`; flags: `chord_structure`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: ` ` -> `:`
+- Gold note -> top-1 note: `4FFn` -> `4FFn:4C#`
+- Gold token likelihood: rank `3`, p `4.354%`
+- Predicted token likelihood: p `84.762%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 26 | `":"` | 84.762% |
+| 2 | 47 | `"\`"` | 4.492% |
+| 3 | 6 | `" "` | 4.354% |
+| 4 | 55 | `"i"` | 1.247% |
+| 5 | 4 | `"\t"` | 0.346% |
+
+### Sample 6
+
+- Sample: `polish_00049`
+- Row/token: `33` / `700`
+- Line/field: `95:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `)` -> `	`
+- Gold note -> top-1 note: `8ccc)` -> `8ccc`
+- Gold token likelihood: rank `5`, p `0.003%`
+- Predicted token likelihood: p `99.750%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 4 | `"\t"` | 99.750% |
+| 2 | 6 | `" "` | 0.015% |
+| 3 | 26 | `":"` | 0.005% |
+| 4 | 59 | `"n"` | 0.004% |
+| 5 | 10 | `")"` | 0.003% |
+
+### Sample 7
+
+- Sample: `polish_00095`
+- Row/token: `63` / `23`
+- Line/field: `4:1`
+- Component bucket: `duration_only`; flags: `duration`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `12ee-` -> `4ee-`
+- Gold note -> top-1 note: `12ee--` -> `4ee--`
+- Gold token likelihood: rank `3`, p `7.402%`
+- Predicted token likelihood: p `29.275%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 348 | `"4ee-"` | 29.275% |
+| 2 | 682 | `"8qqee"` | 9.504% |
+| 3 | 1363 | `"12ee-"` | 7.402% |
+| 4 | 14 | `"."` | 6.739% |
+| 5 | 1757 | `"8.ee-"` | 6.331% |
+
+### Sample 8
+
+- Sample: `polish_00020`
+- Row/token: `12` / `910`
+- Line/field: `156:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `)J` -> `)JJ`
+- Gold note -> top-1 note: `16ff##)J` -> `16ff##)JJ`
+- Gold token likelihood: rank `4`, p `0.477%`
+- Predicted token likelihood: p `90.958%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 305 | `")JJ"` | 90.958% |
+| 2 | 10 | `")"` | 2.580% |
+| 3 | 1570 | `")JJJ"` | 2.073% |
+| 4 | 186 | `")J"` | 0.477% |
+| 5 | 2968 | `")JJk"` | 0.233% |
+
+### Sample 9
+
+- Sample: `polish_00038`
+- Row/token: `25` / `1245`
+- Line/field: `151:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16a` -> `16a)`
+- Gold note -> top-1 note: `16a)J` -> `16a))J`
+- Gold token likelihood: rank `4`, p `1.501%`
+- Predicted token likelihood: p `70.099%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1749 | `"16a)"` | 70.099% |
+| 2 | 455 | `"32a"` | 12.967% |
+| 3 | 1094 | `"16a)JJ"` | 5.078% |
+| 4 | 98 | `"16a"` | 1.501% |
+| 5 | 1983 | `"16a("` | 1.064% |
+
+### Sample 10
+
+- Sample: `polish_00038`
+- Row/token: `25` / `160`
+- Line/field: `20:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `)J` -> `)JJJ`
+- Gold note -> top-1 note: `32a)J` -> `32a)JJJ`
+- Gold token likelihood: rank `8`, p `0.156%`
+- Predicted token likelihood: p `85.963%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1570 | `")JJJ"` | 85.963% |
+| 2 | 10 | `")"` | 4.280% |
+| 3 | 305 | `")JJ"` | 3.777% |
+| 4 | 2968 | `")JJk"` | 2.439% |
+| 5 | 2772 | `"]JJJ"` | 0.617% |
+
+### Sample 11
+
+- Sample: `polish_00052`
+- Row/token: `35` / `372`
+- Line/field: `64:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16g` -> `16g(LL`
+- Gold note -> top-1 note: `16g(L` -> `16g(LL(L`
+- Gold token likelihood: rank `13`, p `0.060%`
+- Predicted token likelihood: p `96.146%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1339 | `"16g(LL"` | 96.146% |
+| 2 | 2310 | `"16e(LL"` | 1.003% |
+| 3 | 2685 | `"16f(LL"` | 0.287% |
+| 4 | 2782 | `"16g("` | 0.198% |
+| 5 | 817 | `"8g(L"` | 0.154% |
+
+### Sample 12
+
+- Sample: `polish_00089`
+- Row/token: `58` / `781`
+- Line/field: `0:0`
+- Component bucket: `termination`; flags: `none`
+- Legacy bucket: `termination`
+- Gold piece -> top-1 piece: `<eos>` -> `
+`
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `2`, p `0.023%`
+- Predicted token likelihood: p `99.582%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 5 | `"\n"` | 99.582% |
+| 2 | 2 | `"<eos>"` | 0.023% |
+| 3 | 6 | `" "` | 0.009% |
+| 4 | 4 | `"\t"` | 0.002% |
+| 5 | 26 | `":"` | 0.002% |
+
+### Sample 13
+
+- Sample: `polish_00027`
+- Row/token: `17` / `1544`
+- Line/field: `222:0`
+- Component bucket: `pitch_and_accidental`; flags: `pitch, accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `8B-L` -> `8dL`
+- Gold note -> top-1 note: `8B-L` -> `8dL`
+- Gold token likelihood: rank `5`, p `0.933%`
+- Predicted token likelihood: p `46.372%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 206 | `"8dL"` | 46.372% |
+| 2 | 96 | `"8d"` | 31.871% |
+| 3 | 2115 | `"8BnL"` | 9.131% |
+| 4 | 828 | `"8d-L"` | 2.964% |
+| 5 | 430 | `"8B-L"` | 0.933% |
+
+### Sample 14
+
+- Sample: `polish_00087`
+- Row/token: `57` / `899`
+- Line/field: `122:1`
+- Component bucket: `column_or_line_break`; flags: `none`
+- Legacy bucket: `column_or_line_break`
+- Gold piece -> top-1 piece: `	` -> `<unk>`
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `2`, p `5.151%`
+- Predicted token likelihood: p `85.768%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 3 | `"<unk>"` | 85.768% |
+| 2 | 4 | `"\t"` | 5.151% |
+| 3 | 5 | `"\n"` | 0.351% |
+| 4 | 6 | `" "` | 0.182% |
+| 5 | 26 | `":"` | 0.137% |
+
+### Sample 15
+
+- Sample: `polish_00089`
+- Row/token: `58` / `291`
+- Line/field: `46:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16ee` -> `16ee)JJ`
+- Gold note -> top-1 note: `16ee)J` -> `16ee)JJ)J`
+- Gold token likelihood: rank `5`, p `0.125%`
+- Predicted token likelihood: p `97.220%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1119 | `"16ee)JJ"` | 97.220% |
+| 2 | 1797 | `"16ee)"` | 0.510% |
+| 3 | 1241 | `"16gg)JJ"` | 0.362% |
+| 4 | 332 | `"16eeJJ"` | 0.129% |
+| 5 | 126 | `"16ee"` | 0.125% |
+
+### Sample 16
+
+- Sample: `polish_00013`
+- Row/token: `7` / `1256`
+- Line/field: `149:2`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `note_other`
+- Gold piece -> top-1 piece: `L` -> `LL`
+- Gold note -> top-1 note: `16qqccL` -> `16qqccLL`
+- Gold token likelihood: rank `8`, p `0.166%`
+- Predicted token likelihood: p `91.529%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 83 | `"LL"` | 91.529% |
+| 2 | 279 | `"(LL"` | 0.768% |
+| 3 | 80 | `"JJ"` | 0.721% |
+| 4 | 308 | `"LLL"` | 0.320% |
+| 5 | 269 | `"#LL"` | 0.282% |

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/command.log

@@ -0,0 +1,491 @@
+Skipping import of cpp extensions due to incompatible torch version 2.11.0+cu129 for torchao version 0.15.0             Please see https://github.com/pytorch/ao/issues/2919 for more info
+[transformers] You are using a model of type `vit_mae` to instantiate a model of type `vit`. This may be expected if you are loading a checkpoint that shares a subset of the architecture (e.g., loading a `sam2_video` checkpoint into `Sam2Model`), but is otherwise not supported and can yield errors. Please verify that the checkpoint is compatible with the model you are instantiating.
+
+[transformers] ViTModel LOAD REPORT from: PRAIG/musvit
+Key                                                                     | Status     | 
+------------------------------------------------------------------------+------------+-
+decoder.decoder_embed.bias                                              | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.weight  | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_before.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.v_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.bias            | UNEXPECTED | 
+decoder.decoder_embed.weight                                            | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.o_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.k_proj.bias   | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.attention.q_proj.weight | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc1.weight          | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.weight          | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.layernorm_after.bias    | UNEXPECTED | 
+decoder.decoder_layers.{0, 1, 2, 3, 4, 5, 6, 7}.mlp.fc2.bias            | UNEXPECTED | 
+decoder.decoder_norm.weight                                             | UNEXPECTED | 
+decoder.decoder_norm.bias                                               | UNEXPECTED | 
+decoder.decoder_pred.bias                                               | UNEXPECTED | 
+decoder.decoder_pos_embed                                               | UNEXPECTED | 
+decoder.mask_token                                                      | UNEXPECTED | 
+decoder.decoder_pred.weight                                             | UNEXPECTED | 
+pooler.dense.weight                                                     | MISSING    | 
+pooler.dense.bias                                                       | MISSING    | 
+
+Notes:
+- UNEXPECTED:	can be ignored when loading from different task/architecture; not ok if you expect identical arch.
+- MISSING:	those params were newly initialized because missing from the checkpoint. Consider training on your downstream task.
+[eval] pages=32/67 mode=teacher_forced elapsed=9.1s
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_4.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.31903086085544125,
+      "errors": 2357,
+      "errors_per_1k_tokens": 31.631215191572167,
+      "errors_per_page_mean": 35.17910447761194,
+      "errors_per_page_median": 36.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.22387655657823496,
+      "errors": 1654,
+      "errors_per_1k_tokens": 22.19687311279608,
+      "errors_per_page_mean": 24.686567164179106,
+      "errors_per_page_median": 25.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.10327558202490525,
+      "errors": 763,
+      "errors_per_1k_tokens": 10.239549084077032,
+      "errors_per_page_mean": 11.388059701492537,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.8955223880597015,
+      "pages_with_error": 60
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05481862479696806,
+      "errors": 405,
+      "errors_per_1k_tokens": 5.435147285781387,
+      "errors_per_page_mean": 6.044776119402985,
+      "errors_per_page_median": 5.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04155387114239307,
+      "errors": 307,
+      "errors_per_1k_tokens": 4.119975843789841,
+      "errors_per_page_mean": 4.582089552238806,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.0334325933946941,
+      "errors": 247,
+      "errors_per_1k_tokens": 3.3147688384888947,
+      "errors_per_page_mean": 3.6865671641791047,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.8208955223880597,
+      "pages_with_error": 55
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03316188413643747,
+      "errors": 245,
+      "errors_per_1k_tokens": 3.2879286049788634,
+      "errors_per_page_mean": 3.656716417910448,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.746268656716418,
+      "pages_with_error": 50
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.025446670276123445,
+      "errors": 188,
+      "errors_per_1k_tokens": 2.5229819499429644,
+      "errors_per_page_mean": 2.8059701492537314,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5522388059701493,
+      "pages_with_error": 37
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.023280996210070383,
+      "errors": 172,
+      "errors_per_1k_tokens": 2.3082600818627124,
+      "errors_per_page_mean": 2.5671641791044775,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.02301028695181375,
+      "errors": 170,
+      "errors_per_1k_tokens": 2.2814198483526806,
+      "errors_per_page_mean": 2.537313432835821,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5223880597014925,
+      "pages_with_error": 35
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.014347590687601516,
+      "errors": 106,
+      "errors_per_1k_tokens": 1.4225323760316715,
+      "errors_per_page_mean": 1.5820895522388059,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.582089552238806,
+      "pages_with_error": 39
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.013264753654574987,
+      "errors": 98,
+      "errors_per_1k_tokens": 1.3151714419915452,
+      "errors_per_page_mean": 1.462686567164179,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.012723335138061722,
+      "errors": 94,
+      "errors_per_1k_tokens": 1.2614909749714822,
+      "errors_per_page_mean": 1.4029850746268657,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.373134328358209,
+      "pages_with_error": 25
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.008662696264212236,
+      "errors": 64,
+      "errors_per_1k_tokens": 0.8588874723210091,
+      "errors_per_page_mean": 0.9552238805970149,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.417910447761194,
+      "pages_with_error": 28
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.008662696264212236,
+      "errors": 64,
+      "errors_per_1k_tokens": 0.8588874723210091,
+      "errors_per_page_mean": 0.9552238805970149,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.40298507462686567,
+      "pages_with_error": 27
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.00852734163508392,
+      "errors": 63,
+      "errors_per_1k_tokens": 0.8454673555659934,
+      "errors_per_page_mean": 0.9402985074626866,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.31343283582089554,
+      "pages_with_error": 21
+    },
+    {
+      "category": "termination",
+      "error_share": 0.00568489442338928,
+      "errors": 42,
+      "errors_per_1k_tokens": 0.5636449037106623,
+      "errors_per_page_mean": 0.6268656716417911,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.0055495397942609635,
+      "errors": 41,
+      "errors_per_1k_tokens": 0.5502247869556465,
+      "errors_per_page_mean": 0.6119402985074627,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3283582089552239,
+      "pages_with_error": 22
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.005278830536004331,
+      "errors": 39,
+      "errors_per_1k_tokens": 0.523384553445615,
+      "errors_per_page_mean": 0.582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2835820895522388,
+      "pages_with_error": 19
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004466702761234434,
+      "errors": 33,
+      "errors_per_1k_tokens": 0.4428638529155204,
+      "errors_per_page_mean": 0.4925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.208955223880597,
+      "pages_with_error": 14
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.004466702761234434,
+      "errors": 33,
+      "errors_per_1k_tokens": 0.4428638529155204,
+      "errors_per_page_mean": 0.4925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.003789929615592853,
+      "errors": 28,
+      "errors_per_1k_tokens": 0.3757632691404415,
+      "errors_per_page_mean": 0.417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0032485110990795887,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.22388059701492538,
+      "pages_with_error": 15
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0031131564699512725,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2835820895522388,
+      "pages_with_error": 19
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0031131564699512725,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0031131564699512725,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.00284244721169464,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1044776119402985,
+      "pages_with_error": 7
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.00284244721169464,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "clef",
+      "error_share": 0.002030319436924743,
+      "errors": 15,
+      "errors_per_1k_tokens": 0.20130175132523653,
+      "errors_per_page_mean": 0.22388059701492538,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.14925373134328357,
+      "pages_with_error": 10
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.001488900920411478,
+      "errors": 11,
+      "errors_per_1k_tokens": 0.14762128430517346,
+      "errors_per_page_mean": 0.16417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0006767731456415809,
+      "errors": 5,
+      "errors_per_1k_tokens": 0.06710058377507884,
+      "errors_per_page_mean": 0.07462686567164178,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0005414185165132648,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0005414185165132648,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.0001353546291283162,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 4,
+    "accidental_and_tie_slur": 41,
+    "accidental_only": 307,
+    "articulation_only": 33,
+    "barline": 23,
+    "chord_structure_and_accidental": 5,
+    "chord_structure_and_accidental_and_articulation": 23,
+    "chord_structure_and_accidental_and_tie_slur": 21,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 94,
+    "chord_structure_and_tie_slur": 98,
+    "chord_structure_and_tie_slur_and_articulation": 11,
+    "chord_structure_only": 39,
+    "clef": 15,
+    "column_or_line_break": 405,
+    "duration_and_accidental": 33,
+    "duration_and_pitch": 106,
+    "duration_and_pitch_and_accidental": 64,
+    "duration_and_pitch_and_accidental_and_tie_slur": 4,
+    "duration_and_pitch_and_tie_slur": 21,
+    "duration_and_tie_slur": 63,
+    "duration_only": 763,
+    "interpretation_or_control": 245,
+    "note_other": 2357,
+    "null_token": 64,
+    "pitch_and_accidental": 188,
+    "pitch_and_accidental_and_tie_slur": 28,
+    "pitch_and_tie_slur": 23,
+    "pitch_only": 247,
+    "rest_note_confusion": 170,
+    "spine_operation": 172,
+    "termination": 42,
+    "tie_slur_and_articulation": 24,
+    "tie_slur_only": 1654
+  },
+  "component_error_groups": {
+    "structural": 928,
+    "visual": 6460
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 12.814844940090552,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3145,
+    "articulation": 53,
+    "barline": 22,
+    "clef": 15,
+    "column_or_line_break": 406,
+    "duration": 1011,
+    "interpretation_or_control": 138,
+    "note_other": 209,
+    "null_token": 65,
+    "pitch": 20,
+    "rest": 279,
+    "spine_operation": 205,
+    "termination": 28,
+    "tie_slur": 1792
+  },
+  "gold_error_groups": {
+    "structural": 842,
+    "visual": 6546
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.550797462463379,
+      "LER": 42.2453498840332,
+      "SER": 10.130560874938965,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.0991478225860565,
+      "top1_errors": 7388,
+      "top1_errors_per_page": 110.26865671641791
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 41,
+    "predicted_unk_literal": 104
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/per_page_records.jsonl

@@ -0,0 +1,67 @@
+{"dataset_index": 0, "sample_id": "polish_00002", "source_ids": [], "target_token_count": 1725, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 59, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 55, "field_index": 0, "field_text": "2.AA[", "gold_field": "2.AA[", "gold_line": "2.AA[\t4r\t12r", "gold_note_token": "2.AA[", "gold_piece": "[", "gold_token_id": 43, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 67, "line_char_start": 55, "line_index": 4, "line_text": "2.AA[\t4r\t12r", "operation": "replace", "pred_field": "2.AA(", "pred_local_replacement_line": "2.AA(\t4r\t12r", "pred_note_token": "2.AA(", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 13.612950699043413, "LER": 67.54386138916016, "SER": 9.976798143851507}, "nll": 0.44138890504837036, "top1_error_count": 172, "top1_error_rate": 0.09971014492753623}}
+{"dataset_index": 1, "sample_id": "polish_00004", "source_ids": [], "target_token_count": 989, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 35, "field_char_start": 30, "field_index": 0, "field_text": "8AA(L", "gold_field": "8AA(L", "gold_line": "8AA(L\t4.ee(", "gold_note_token": "8AA(L", "gold_piece": "8AA", "gold_token_id": 210, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 41, "line_char_start": 30, "line_index": 2, "line_text": "8AA(L\t4.ee(", "operation": "replace", "pred_field": "=!|:(L", "pred_local_replacement_line": "=!|:(L\t4.ee(", "pred_note_token": "=!|:(L", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 6.369047619047619, "LER": 32.704402923583984, "SER": 7.186234817813765}, "nll": 0.4864768087863922, "top1_error_count": 71, "top1_error_rate": 0.0717896865520728}}
+{"dataset_index": 2, "sample_id": "polish_00005", "source_ids": [], "target_token_count": 742, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t2cc(", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 55, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t2cc(", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t2cc(", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 5.968077723802915, "LER": 40.1574821472168, "SER": 9.31174089068826}, "nll": 0.5857667922973633, "top1_error_count": 69, "top1_error_rate": 0.09299191374663072}}
+{"dataset_index": 3, "sample_id": "polish_00007", "source_ids": [], "target_token_count": 888, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 85, "char_start": 81, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 87, "field_char_start": 81, "field_index": 1, "field_text": "16dd#L", "gold_field": "16dd#L", "gold_line": "8E 8A 8cJ\t16dd#L", "gold_note_token": "16dd#L", "gold_piece": "16dd", "gold_token_id": 112, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 87, "line_char_start": 71, "line_index": 6, "line_text": "8E 8A 8cJ\t16dd#L", "operation": "replace", "pred_field": "16dd#LL#L", "pred_local_replacement_line": "8E 8A 8cJ\t16dd#LL#L", "pred_note_token": "16dd#LL#L", "pred_piece": "16dd#LL", "pred_token_id": 1164, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 7.578676942838793, "LER": 32.66666793823242, "SER": 8.342728297632469}, "nll": 0.4954758286476135, "top1_error_count": 74, "top1_error_rate": 0.08333333333333333}}
+{"dataset_index": 4, "sample_id": "polish_00008", "source_ids": [], "target_token_count": 1236, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 48, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 48, "field_char_start": 45, "field_index": 1, "field_text": "8d[", "gold_field": "8d[", "gold_line": "8r\t8d[", "gold_note_token": "8d[", "gold_piece": "8d[", "gold_token_id": 1247, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 48, "line_char_start": 42, "line_index": 3, "line_text": "8r\t8d[", "operation": "replace", "pred_field": "8d(", "pred_local_replacement_line": "8r\t8d(", "pred_note_token": "8d(", "pred_piece": "8d(", "pred_token_id": 1822, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 9.583133684714902, "LER": 45.549739837646484, "SER": 9.635627530364372}, "nll": 0.558600902557373, "top1_error_count": 119, "top1_error_rate": 0.09627831715210355}}
+{"dataset_index": 5, "sample_id": "polish_00009", "source_ids": [], "target_token_count": 942, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 111, "char_start": 110, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 110, "field_char_start": 107, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 52, "gold_trimmed_index": 52, "group": "structural", "line_char_end": 120, "line_char_start": 107, "line_index": 11, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 52, "visually_grounded": false}, "metrics": {"CER": 5.438259756877799, "LER": 23.456790924072266, "SER": 5.738575982996812}, "nll": 0.34805387258529663, "top1_error_count": 54, "top1_error_rate": 0.05732484076433121}}
+{"dataset_index": 6, "sample_id": "polish_00011", "source_ids": [], "target_token_count": 1093, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 7.338056680161944, "LER": 42.105262756347656, "SER": 9.24908424908425}, "nll": 0.5081247687339783, "top1_error_count": 101, "top1_error_rate": 0.09240622140896615}}
+{"dataset_index": 7, "sample_id": "polish_00013", "source_ids": [], "target_token_count": 1877, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 49, "char_start": 47, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 49, "field_char_start": 43, "field_index": 1, "field_text": "16.g(L", "gold_field": "16.g(L", "gold_line": "8CC 8GG 8CL\t16.g(L\t4en", "gold_note_token": "16.g(L", "gold_piece": "(L", "gold_token_id": 171, "gold_token_index": 19, "gold_trimmed_index": 19, "group": "visual", "line_char_end": 53, "line_char_start": 31, "line_index": 3, "line_text": "8CC 8GG 8CL\t16.g(L\t4en", "operation": "replace", "pred_field": "16.g(LL", "pred_local_replacement_line": "8CC 8GG 8CL\t16.g(LL\t4en", "pred_note_token": "16.g(LL", "pred_piece": "(LL", "pred_token_id": 279, "pred_trimmed_index": 19, "visually_grounded": true}, "metrics": {"CER": 7.720224348399868, "LER": 51.54185104370117, "SER": 8.208955223880597}, "nll": 0.4461793601512909, "top1_error_count": 154, "top1_error_rate": 0.08204581779435269}}
+{"dataset_index": 8, "sample_id": "polish_00014", "source_ids": [], "target_token_count": 1099, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 50, "char_start": 46, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8F( 8A(L", "gold_field": "8F( 8A(L", "gold_line": "8F( 8A(L\t16bn(L", "gold_note_token": "8A(L", "gold_piece": "8A(L", "gold_token_id": 1848, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 57, "line_char_start": 42, "line_index": 3, "line_text": "8F( 8A(L\t16bn(L", "operation": "replace", "pred_field": "8F( 8AL", "pred_local_replacement_line": "8F( 8AL\t16bn(L", "pred_note_token": "8AL", "pred_piece": "8AL", "pred_token_id": 343, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 7.488766849725412, "LER": 35.93073654174805, "SER": 10.018214936247723}, "nll": 0.5599254369735718, "top1_error_count": 110, "top1_error_rate": 0.10009099181073704}}
+{"dataset_index": 9, "sample_id": "polish_00015", "source_ids": [], "target_token_count": 1086, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 38, "char_start": 37, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 38, "field_char_start": 37, "field_index": 1, "field_text": "*", "gold_field": "*", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "*^", "pred_local_replacement_line": "*^\t*^", "pred_note_token": null, "pred_piece": "*^", "pred_token_id": 191, "pred_trimmed_index": 10, "visually_grounded": false}, "metrics": {"CER": 8.44994617868676, "LER": 39.2638053894043, "SER": 9.76958525345622}, "nll": 0.49212196469306946, "top1_error_count": 106, "top1_error_rate": 0.09760589318600368}}
+{"dataset_index": 10, "sample_id": "polish_00017", "source_ids": [], "target_token_count": 1053, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 40, "char_start": 39, "component_category": "chord_structure_and_tie_slur", "component_flags": ["tie_slur", "chord_structure"], "component_group": "visual", "field_char_end": 44, "field_char_start": 34, "field_index": 0, "field_text": "4BBB- 4BB-", "gold_field": "4BBB- 4BB-", "gold_line": "4BBB- 4BB-\t4E 4cn", "gold_note_token": "4BBB-", "gold_piece": " ", "gold_token_id": 6, "gold_token_index": 9, "gold_trimmed_index": 9, "group": "visual", "line_char_end": 51, "line_char_start": 34, "line_index": 2, "line_text": "4BBB- 4BB-\t4E 4cn", "operation": "replace", "pred_field": "4BBB-(4BB-", "pred_local_replacement_line": "4BBB-(4BB-\t4E 4cn", "pred_note_token": "4BBB-(4BB-", "pred_piece": "(", "pred_token_id": 9, "pred_trimmed_index": 9, "visually_grounded": true}, "metrics": {"CER": 2.321899736147757, "LER": 15.55555534362793, "SER": 4.3726235741444865}, "nll": 0.2698626220226288, "top1_error_count": 46, "top1_error_rate": 0.04368471035137702}}
+{"dataset_index": 11, "sample_id": "polish_00019", "source_ids": [], "target_token_count": 1006, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 29, "char_start": 26, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 33, "field_char_start": 26, "field_index": 0, "field_text": "8AA 8AL", "gold_field": "8AA 8AL", "gold_line": "8AA 8AL\t8.cc 8.cccL", "gold_note_token": "8AA", "gold_piece": "8AA", "gold_token_id": 210, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 45, "line_char_start": 26, "line_index": 2, "line_text": "8AA 8AL\t8.cc 8.cccL", "operation": "replace", "pred_field": "8C 8AL", "pred_local_replacement_line": "8C 8AL\t8.cc 8.cccL", "pred_note_token": "8C", "pred_piece": "8C", "pred_token_id": 166, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 5.124223602484472, "LER": 33.52941131591797, "SER": 7.6616915422885565}, "nll": 0.5398172736167908, "top1_error_count": 77, "top1_error_rate": 0.07654075546719682}}
+{"dataset_index": 12, "sample_id": "polish_00020", "source_ids": [], "target_token_count": 1183, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 5.520926090828139, "LER": 33.49282455444336, "SER": 7.4450084602368864}, "nll": 0.44378000497817993, "top1_error_count": 88, "top1_error_rate": 0.07438715131022823}}
+{"dataset_index": 13, "sample_id": "polish_00022", "source_ids": [], "target_token_count": 695, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 60, "char_start": 56, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 60, "field_char_start": 52, "field_index": 0, "field_text": "8G( 8B(L", "gold_field": "8G( 8B(L", "gold_line": "8G( 8B(L\t8ddd)L", "gold_note_token": "8B(L", "gold_piece": "8B(L", "gold_token_id": 2168, "gold_token_index": 20, "gold_trimmed_index": 20, "group": "visual", "line_char_end": 67, "line_char_start": 52, "line_index": 4, "line_text": "8G( 8B(L\t8ddd)L", "operation": "replace", "pred_field": "8G( 8BL", "pred_local_replacement_line": "8G( 8BL\t8ddd)L", "pred_note_token": "8BL", "pred_piece": "8BL", "pred_token_id": 396, "pred_trimmed_index": 20, "visually_grounded": true}, "metrics": {"CER": 7.894736842105263, "LER": 42.85714340209961, "SER": 10.518731988472622}, "nll": 0.6661466360092163, "top1_error_count": 73, "top1_error_rate": 0.10503597122302158}}
+{"dataset_index": 14, "sample_id": "polish_00023", "source_ids": [], "target_token_count": 1332, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 51, "field_index": 1, "field_text": "4gg)", "gold_field": "4gg)", "gold_line": "8GG 8GL\t4gg)\t8bnL", "gold_note_token": "4gg)", "gold_piece": "4gg)", "gold_token_id": 2154, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 60, "line_char_start": 43, "line_index": 3, "line_text": "8GG 8GL\t4gg)\t8bnL", "operation": "replace", "pred_field": "4gg]", "pred_local_replacement_line": "8GG 8GL\t4gg]\t8bnL", "pred_note_token": "4gg]", "pred_piece": "4gg]", "pred_token_id": 1531, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 6.839186691312385, "LER": 38.414634704589844, "SER": 7.2877535687453046}, "nll": 0.40376219153404236, "top1_error_count": 97, "top1_error_rate": 0.07282282282282282}}
+{"dataset_index": 15, "sample_id": "polish_00024", "source_ids": [], "target_token_count": 1075, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 7.628524046434494, "LER": 34.68207931518555, "SER": 8.472998137802607}, "nll": 0.5061269998550415, "top1_error_count": 91, "top1_error_rate": 0.08465116279069768}}
+{"dataset_index": 16, "sample_id": "polish_00025", "source_ids": [], "target_token_count": 879, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 34, "field_index": 0, "field_text": "*M3/4", "gold_field": "*M3/4", "gold_line": "*M3/4\t*M3/4", "gold_note_token": null, "gold_piece": "*M3/4", "gold_token_id": 283, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 45, "line_char_start": 34, "line_index": 2, "line_text": "*M3/4\t*M3/4", "operation": "replace", "pred_field": "*M2/4", "pred_local_replacement_line": "*M2/4\t*M3/4", "pred_note_token": null, "pred_piece": "*M2/4", "pred_token_id": 304, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 6.993865030674846, "LER": 44.05594253540039, "SER": 10.022779043280181}, "nll": 0.5387299060821533, "top1_error_count": 88, "top1_error_rate": 0.10011376564277588}}
+{"dataset_index": 17, "sample_id": "polish_00027", "source_ids": [], "target_token_count": 1825, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 46, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 46, "field_index": 0, "field_text": "8BBB- 8BB-", "gold_field": "8BBB- 8BB-", "gold_line": "8BBB- 8BB-\t8b- 8dd", "gold_note_token": "8BBB-", "gold_piece": "8BBB-", "gold_token_id": 632, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 3, "line_text": "8BBB- 8BB-\t8b- 8dd", "operation": "replace", "pred_field": "8AAA 8BB-", "pred_local_replacement_line": "8AAA 8BB-\t8b- 8dd", "pred_note_token": "8AAA", "pred_piece": "8AAA", "pred_token_id": 490, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 8.7678415380134, "LER": 49.22480773925781, "SER": 9.758771929824562}, "nll": 0.5116015672683716, "top1_error_count": 178, "top1_error_rate": 0.09753424657534246}}
+{"dataset_index": 18, "sample_id": "polish_00029", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 33, "char_start": 30, "component_category": "duration_and_pitch", "component_flags": ["duration", "pitch"], "component_group": "visual", "field_char_end": 39, "field_char_start": 30, "field_index": 0, "field_text": "8CC( 8C(L", "gold_field": "8CC( 8C(L", "gold_line": "8CC( 8C(L\t4gg([", "gold_note_token": "8CC(", "gold_piece": "8CC", "gold_token_id": 364, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 45, "line_char_start": 30, "line_index": 2, "line_text": "8CC( 8C(L\t4gg([", "operation": "replace", "pred_field": "=!|:( 8C(L", "pred_local_replacement_line": "=!|:( 8C(L\t4gg([", "pred_note_token": "=!|:(", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 7.876712328767123, "LER": 36.17021179199219, "SER": 9.895414320193082}, "nll": 0.6108992695808411, "top1_error_count": 123, "top1_error_rate": 0.09887459807073955}}
+{"dataset_index": 19, "sample_id": "polish_00030", "source_ids": [], "target_token_count": 1263, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 60, "char_start": 59, "component_category": "chord_structure_only", "component_flags": ["chord_structure"], "component_group": "visual", "field_char_end": 63, "field_char_start": 56, "field_index": 2, "field_text": "2dn 2fn", "gold_field": "2dn 2fn", "gold_line": "8r\t4.an(\t2dn 2fn", "gold_note_token": "2dn", "gold_piece": " ", "gold_token_id": 6, "gold_token_index": 20, "gold_trimmed_index": 20, "group": "visual", "line_char_end": 63, "line_char_start": 47, "line_index": 3, "line_text": "8r\t4.an(\t2dn 2fn", "operation": "replace", "pred_field": "2dn:2fn", "pred_local_replacement_line": "8r\t4.an(\t2dn:2fn", "pred_note_token": "2dn:2fn", "pred_piece": ":", "pred_token_id": 26, "pred_trimmed_index": 20, "visually_grounded": true}, "metrics": {"CER": 11.927877947295423, "LER": 52.55813980102539, "SER": 10.380348652931854}, "nll": 0.4622728228569031, "top1_error_count": 131, "top1_error_rate": 0.10372129849564529}}
+{"dataset_index": 20, "sample_id": "polish_00031", "source_ids": [], "target_token_count": 862, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 39, "char_start": 34, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 39, "field_char_start": 34, "field_index": 0, "field_text": "*M3/8", "gold_field": "*M3/8", "gold_line": "*M3/8\t*M3/8", "gold_note_token": null, "gold_piece": "*M3/8", "gold_token_id": 592, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 45, "line_char_start": 34, "line_index": 2, "line_text": "*M3/8\t*M3/8", "operation": "replace", "pred_field": "*M3/4", "pred_local_replacement_line": "*M3/4\t*M3/8", "pred_note_token": null, "pred_piece": "*M3/4", "pred_token_id": 283, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 8.545335942596216, "LER": 39.86014175415039, "SER": 9.059233449477352}, "nll": 0.5197849869728088, "top1_error_count": 78, "top1_error_rate": 0.09048723897911833}}
+{"dataset_index": 21, "sample_id": "polish_00032", "source_ids": [], "target_token_count": 1387, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 50, "char_start": 45, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 52, "field_char_start": 45, "field_index": 2, "field_text": "16ff#]L", "gold_field": "16ff#]L", "gold_line": "4G#(\t8G#L\t16ff#]L", "gold_note_token": "16ff#]L", "gold_piece": "16ff#", "gold_token_id": 180, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 52, "line_char_start": 35, "line_index": 3, "line_text": "4G#(\t8G#L\t16ff#]L", "operation": "replace", "pred_field": "16ff#)]L", "pred_local_replacement_line": "4G#(\t8G#L\t16ff#)]L", "pred_note_token": "16ff#)]L", "pred_piece": "16ff#)", "pred_token_id": 2192, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 11.356993736951983, "LER": 44.117645263671875, "SER": 11.76046176046176}, "nll": 0.7185993790626526, "top1_error_count": 163, "top1_error_rate": 0.11751982696467195}}
+{"dataset_index": 22, "sample_id": "polish_00034", "source_ids": [], "target_token_count": 788, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 45, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 50, "field_char_start": 42, "field_index": 0, "field_text": "8f( 8a(L", "gold_field": "8f( 8a(L", "gold_line": "8f( 8a(L\t16r", "gold_note_token": null, "gold_piece": "8f(", "gold_token_id": 2288, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 54, "line_char_start": 42, "line_index": 3, "line_text": "8f( 8a(L\t16r", "operation": "replace", "pred_field": "* 8a(L", "pred_local_replacement_line": "* 8a(L\t16r", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 24.58808618504436, "LER": 57.76397705078125, "SER": 39.135959339263025}, "nll": 1.5077717304229736, "top1_error_count": 179, "top1_error_rate": 0.22715736040609136}}
+{"dataset_index": 23, "sample_id": "polish_00036", "source_ids": [], "target_token_count": 1317, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 288, "char_start": 284, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 288, "field_char_start": 284, "field_index": 2, "field_text": "8fnL", "gold_field": "8fnL", "gold_line": "2.CC_ 2.C_\t2.g)\t8fnL", "gold_note_token": "8fnL", "gold_piece": "8fnL", "gold_token_id": 2196, "gold_token_index": 172, "gold_trimmed_index": 172, "group": "visual", "line_char_end": 288, "line_char_start": 268, "line_index": 25, "line_text": "2.CC_ 2.C_\t2.g)\t8fnL", "operation": "replace", "pred_field": "8fL", "pred_local_replacement_line": "2.CC_ 2.C_\t2.g)\t8fL", "pred_note_token": "8fL", "pred_piece": "8fL", "pred_token_id": 265, "pred_trimmed_index": 172, "visually_grounded": true}, "metrics": {"CER": 5.794477138976912, "LER": 21.35416603088379, "SER": 4.939209726443769}, "nll": 0.2921859920024872, "top1_error_count": 65, "top1_error_rate": 0.04935459377372817}}
+{"dataset_index": 24, "sample_id": "polish_00037", "source_ids": [], "target_token_count": 871, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 49, "char_start": 46, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 46, "field_index": 0, "field_text": "4Fn 4dn", "gold_field": "4Fn 4dn", "gold_line": "4Fn 4dn\t2f#] 2ff#]", "gold_note_token": "4Fn", "gold_piece": "4Fn", "gold_token_id": 1918, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 64, "line_char_start": 46, "line_index": 2, "line_text": "4Fn 4dn\t2f#] 2ff#]", "operation": "replace", "pred_field": "4F# 4dn", "pred_local_replacement_line": "4F# 4dn\t2f#] 2ff#]", "pred_note_token": "4F#", "pred_piece": "4F#", "pred_token_id": 278, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 4.251144538914323, "LER": 26.174495697021484, "SER": 5.977011494252873}, "nll": 0.27751046419143677, "top1_error_count": 52, "top1_error_rate": 0.05970149253731343}}
+{"dataset_index": 25, "sample_id": "polish_00038", "source_ids": [], "target_token_count": 1471, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 88, "char_start": 83, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 88, "field_char_start": 83, "field_index": 1, "field_text": "16a(L", "gold_field": "16a(L", "gold_line": "8CC 8FF# 8CJ\t16a(L\t8e-", "gold_note_token": "16a(L", "gold_piece": "16a(L", "gold_token_id": 2253, "gold_token_index": 38, "gold_trimmed_index": 38, "group": "visual", "line_char_end": 92, "line_char_start": 70, "line_index": 5, "line_text": "8CC 8FF# 8CJ\t16a(L\t8e-", "operation": "replace", "pred_field": "16aLL", "pred_local_replacement_line": "8CC 8FF# 8CJ\t16aLL\t8e-", "pred_note_token": "16aLL", "pred_piece": "16aLL", "pred_token_id": 338, "pred_trimmed_index": 38, "visually_grounded": true}, "metrics": {"CER": 6.754596322941647, "LER": 39.106143951416016, "SER": 7.551020408163265}, "nll": 0.4218112826347351, "top1_error_count": 111, "top1_error_rate": 0.07545887151597552}}
+{"dataset_index": 26, "sample_id": "polish_00039", "source_ids": [], "target_token_count": 1503, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 93, "char_start": 88, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 93, "field_char_start": 88, "field_index": 0, "field_text": "16A-L", "gold_field": "16A-L", "gold_line": "16A-L\t2.AA-[\t16cL\t8ee-", "gold_note_token": "16A-L", "gold_piece": "16A-L", "gold_token_id": 2834, "gold_token_index": 24, "gold_trimmed_index": 24, "group": "visual", "line_char_end": 110, "line_char_start": 88, "line_index": 4, "line_text": "16A-L\t2.AA-[\t16cL\t8ee-", "operation": "replace", "pred_field": "16A-LL", "pred_local_replacement_line": "16A-LL\t2.AA-[\t16cL\t8ee-", "pred_note_token": "16A-LL", "pred_piece": "16A-LL", "pred_token_id": 1109, "pred_trimmed_index": 24, "visually_grounded": true}, "metrics": {"CER": 9.343236903831118, "LER": 42.328041076660156, "SER": 9.52063914780293}, "nll": 0.5583821535110474, "top1_error_count": 143, "top1_error_rate": 0.09514304723885562}}
+{"dataset_index": 27, "sample_id": "polish_00040", "source_ids": [], "target_token_count": 1109, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 30, "char_start": 16, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 30, "field_char_start": 16, "field_index": 0, "field_text": "*k[f#c#g#d#a#]", "gold_field": "*k[f#c#g#d#a#]", "gold_line": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "gold_note_token": null, "gold_piece": "*k[f#c#g#d#a#]", "gold_token_id": 911, "gold_token_index": 4, "gold_trimmed_index": 4, "group": "structural", "line_char_end": 45, "line_char_start": 16, "line_index": 1, "line_text": "*k[f#c#g#d#a#]\t*k[f#c#g#d#a#]", "operation": "replace", "pred_field": "*k[f#c#g#]", "pred_local_replacement_line": "*k[f#c#g#]\t*k[f#c#g#d#a#]", "pred_note_token": null, "pred_piece": "*k[f#c#g#]", "pred_token_id": 472, "pred_trimmed_index": 4, "visually_grounded": false}, "metrics": {"CER": 8.201581027667984, "LER": 42.6136360168457, "SER": 10.469314079422382}, "nll": 0.5462478399276733, "top1_error_count": 116, "top1_error_rate": 0.10459873760144274}}
+{"dataset_index": 28, "sample_id": "polish_00043", "source_ids": [], "target_token_count": 885, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "insert", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 14.202220923436586, "LER": 54.91329574584961, "SER": 17.307692307692307}, "nll": 0.9600445628166199, "top1_error_count": 153, "top1_error_rate": 0.17288135593220338}}
+{"dataset_index": 29, "sample_id": "polish_00044", "source_ids": [], "target_token_count": 677, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 74, "char_start": 73, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 82, "field_char_start": 70, "field_index": 1, "field_text": "16f( 16dd(Jk", "gold_field": "16f( 16dd(Jk", "gold_line": "16B-(Jk\t16f( 16dd(Jk", "gold_note_token": "16f(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 22, "gold_trimmed_index": 22, "group": "visual", "line_char_end": 82, "line_char_start": 62, "line_index": 4, "line_text": "16B-(Jk\t16f( 16dd(Jk", "operation": "delete", "pred_field": "16f  16dd(Jk", "pred_local_replacement_line": "16B-(Jk\t16f  16dd(Jk", "pred_note_token": "16f", "pred_piece": " ", "pred_token_id": 6, "pred_trimmed_index": 22, "visually_grounded": true}, "metrics": {"CER": 8.61132660977502, "LER": 49.58677673339844, "SER": 10.502958579881657}, "nll": 0.6658834218978882, "top1_error_count": 71, "top1_error_rate": 0.10487444608567208}}
+{"dataset_index": 30, "sample_id": "polish_00046", "source_ids": [], "target_token_count": 1069, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 57, "char_start": 54, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 57, "field_char_start": 54, "field_index": 0, "field_text": "4.F", "gold_field": "4.F", "gold_line": "4.F\t4.f", "gold_note_token": "4.F", "gold_piece": "4.F", "gold_token_id": 614, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": "4.F\t4.f", "operation": "replace", "pred_field": "4F", "pred_local_replacement_line": "4F\t4.f", "pred_note_token": "4F", "pred_piece": "4F", "pred_token_id": 153, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 8.527131782945736, "LER": 39.2156867980957, "SER": 8.801498127340825}, "nll": 0.3867547810077667, "top1_error_count": 94, "top1_error_rate": 0.08793264733395698}}
+{"dataset_index": 31, "sample_id": "polish_00047", "source_ids": [], "target_token_count": 1243, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 66, "char_start": 63, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 68, "field_char_start": 63, "field_index": 1, "field_text": "8eeq(", "gold_field": "8eeq(", "gold_line": ".\t8eeq(\t.", "gold_note_token": "8eeq(", "gold_piece": "8ee", "gold_token_id": 84, "gold_token_index": 28, "gold_trimmed_index": 28, "group": "visual", "line_char_end": 70, "line_char_start": 61, "line_index": 4, "line_text": ".\t8eeq(\t.", "operation": "replace", "pred_field": "8ddq(", "pred_local_replacement_line": ".\t8ddq(\t.", "pred_note_token": "8ddq(", "pred_piece": "8dd", "pred_token_id": 79, "pred_trimmed_index": 28, "visually_grounded": true}, "metrics": {"CER": 5.326295585412668, "LER": 28.04878044128418, "SER": 5.394524959742351}, "nll": 0.34744811058044434, "top1_error_count": 67, "top1_error_rate": 0.053901850362027354}}
+{"dataset_index": 32, "sample_id": "polish_00048", "source_ids": [], "target_token_count": 1384, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 51, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 53, "field_char_start": 48, "field_index": 0, "field_text": "4c 4f", "gold_field": "4c 4f", "gold_line": "4c 4f\t12aa-(L", "gold_note_token": "4f", "gold_piece": "4f", "gold_token_id": 103, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 48, "line_index": 3, "line_text": "4c 4f\t12aa-(L", "operation": "replace", "pred_field": "4c 4a-", "pred_local_replacement_line": "4c 4a-\t12aa-(L", "pred_note_token": "4a-", "pred_piece": "4a-", "pred_token_id": 341, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 16.256938937351308, "LER": 50.1872673034668, "SER": 11.641359363702097}, "nll": 0.568595290184021, "top1_error_count": 161, "top1_error_rate": 0.11632947976878613}}
+{"dataset_index": 33, "sample_id": "polish_00049", "source_ids": [], "target_token_count": 1244, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 12.912912912912914, "LER": 42.16867446899414, "SER": 12.067578439259854}, "nll": 0.5660963654518127, "top1_error_count": 126, "top1_error_rate": 0.10128617363344052}}
+{"dataset_index": 34, "sample_id": "polish_00050", "source_ids": [], "target_token_count": 1242, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 59, "char_start": 56, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 61, "field_char_start": 56, "field_index": 1, "field_text": "8ccq(", "gold_field": "8ccq(", "gold_line": ".\t8ccq(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 61, "line_char_start": 54, "line_index": 3, "line_text": ".\t8ccq(", "operation": "replace", "pred_field": "8ccqq(", "pred_local_replacement_line": ".\t8ccqq(", "pred_note_token": "8ccqq(", "pred_piece": "8ccq", "pred_token_id": 2776, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 11.79749187180678, "LER": 52.94117736816406, "SER": 10.717163577759871}, "nll": 0.5500451922416687, "top1_error_count": 133, "top1_error_rate": 0.107085346215781}}
+{"dataset_index": 35, "sample_id": "polish_00052", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 52, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 59, "field_char_start": 47, "field_index": 0, "field_text": "2.DD-( 2.D-(", "gold_field": "2.DD-( 2.D-(", "gold_line": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "gold_note_token": "2.DD-(", "gold_piece": "-", "gold_token_id": 13, "gold_token_index": 13, "gold_trimmed_index": 13, "group": "visual", "line_char_end": 74, "line_char_start": 47, "line_index": 3, "line_text": "2.DD-( 2.D-(\t8ff(L\t2.f 2.a-", "operation": "replace", "pred_field": "2.DD-[( 2.D-(", "pred_local_replacement_line": "2.DD-[( 2.D-(\t8ff(L\t2.f 2.a-", "pred_note_token": "2.DD-[(", "pred_piece": "-[", "pred_token_id": 429, "pred_trimmed_index": 13, "visually_grounded": true}, "metrics": {"CER": 10.708215297450426, "LER": 47.540985107421875, "SER": 13.991769547325102}, "nll": 0.8005555868148804, "top1_error_count": 136, "top1_error_rate": 0.1397738951695786}}
+{"dataset_index": 36, "sample_id": "polish_00053", "source_ids": [], "target_token_count": 1018, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 51, "char_start": 48, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 62, "field_char_start": 48, "field_index": 0, "field_text": "4FFn: 4Cn: 4A:", "gold_field": "4FFn: 4Cn: 4A:", "gold_line": "4FFn: 4Cn: 4A:\t8fn", "gold_note_token": "4FFn:", "gold_piece": "4FF", "gold_token_id": 369, "gold_token_index": 17, "gold_trimmed_index": 17, "group": "visual", "line_char_end": 66, "line_char_start": 48, "line_index": 4, "line_text": "4FFn: 4Cn: 4A:\t8fn", "operation": "replace", "pred_field": "4FF#n: 4Cn: 4A:", "pred_local_replacement_line": "4FF#n: 4Cn: 4A:\t8fn", "pred_note_token": "4FF#n:", "pred_piece": "4FF#", "pred_token_id": 607, "pred_trimmed_index": 17, "visually_grounded": true}, "metrics": {"CER": 4.09982174688057, "LER": 17.127071380615234, "SER": 4.818092428711897}, "nll": 0.28008151054382324, "top1_error_count": 49, "top1_error_rate": 0.0481335952848723}}
+{"dataset_index": 37, "sample_id": "polish_00054", "source_ids": [], "target_token_count": 689, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 51, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 55, "field_char_start": 46, "field_index": 0, "field_text": "8B-( 8d(L", "gold_field": "8B-( 8d(L", "gold_line": "8B-( 8d(L\t8fL", "gold_note_token": "8d(L", "gold_piece": "8d(L", "gold_token_id": 1103, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 59, "line_char_start": 46, "line_index": 3, "line_text": "8B-( 8d(L\t8fL", "operation": "replace", "pred_field": "8B-( 8dL", "pred_local_replacement_line": "8B-( 8dL\t8fL", "pred_note_token": "8dL", "pred_piece": "8dL", "pred_token_id": 206, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 9.049079754601227, "LER": 38.97058868408203, "SER": 11.19186046511628}, "nll": 0.7770552039146423, "top1_error_count": 77, "top1_error_rate": 0.11175616835994194}}
+{"dataset_index": 38, "sample_id": "polish_00056", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 45, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 45, "field_char_start": 42, "field_index": 0, "field_text": "8FL", "gold_field": "8FL", "gold_line": "8FL\t4cc 4ccc", "gold_note_token": null, "gold_piece": "8FL", "gold_token_id": 433, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 54, "line_char_start": 42, "line_index": 2, "line_text": "8FL\t4cc 4ccc", "operation": "replace", "pred_field": "*", "pred_local_replacement_line": "*\t4cc 4ccc", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 14.451901565995525, "LER": 46.54377746582031, "SER": 14.323843416370108}, "nll": 0.5860705375671387, "top1_error_count": 161, "top1_error_rate": 0.1431111111111111}}
+{"dataset_index": 39, "sample_id": "polish_00057", "source_ids": [], "target_token_count": 733, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 52, "component_category": "duration_and_pitch_and_accidental", "component_flags": ["duration", "pitch", "accidental"], "component_group": "visual", "field_char_end": 57, "field_char_start": 52, "field_index": 1, "field_text": "8ff#q", "gold_field": "8ff#q", "gold_line": ".\t8ff#q", "gold_note_token": "8ff#q", "gold_piece": "8ff#", "gold_token_id": 92, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": ".\t8ff#q", "operation": "replace", "pred_field": "ee-q", "pred_local_replacement_line": ".\tee-q", "pred_note_token": "ee-q", "pred_piece": "ee-", "pred_token_id": 1672, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 12.635135135135133, "LER": 46.853145599365234, "SER": 15.710382513661203}, "nll": 0.9459651708602905, "top1_error_count": 115, "top1_error_rate": 0.15688949522510232}}
+{"dataset_index": 40, "sample_id": "polish_00058", "source_ids": [], "target_token_count": 900, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "rest", "char_end": 178, "char_start": 175, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 178, "field_char_start": 175, "field_index": 1, "field_text": "16r", "gold_field": "16r", "gold_line": "8D 8F 8A 8dL\t16r", "gold_note_token": "16r", "gold_piece": "16r", "gold_token_id": 168, "gold_token_index": 72, "gold_trimmed_index": 72, "group": "visual", "line_char_end": 178, "line_char_start": 162, "line_index": 12, "line_text": "8D 8F 8A 8dL\t16r", "operation": "replace", "pred_field": "8r", "pred_local_replacement_line": "8D 8F 8A 8dL\t8r", "pred_note_token": "8r", "pred_piece": "8r", "pred_token_id": 91, "pred_trimmed_index": 72, "visually_grounded": true}, "metrics": {"CER": 6.990881458966565, "LER": 32.85714340209961, "SER": 7.897664071190211}, "nll": 0.5019165277481079, "top1_error_count": 71, "top1_error_rate": 0.07888888888888888}}
+{"dataset_index": 41, "sample_id": "polish_00059", "source_ids": [], "target_token_count": 1838, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 60, "char_start": 56, "component_category": "duration_only", "component_flags": ["duration"], "component_group": "visual", "field_char_end": 60, "field_char_start": 56, "field_index": 1, "field_text": "8AAL", "gold_field": "8AAL", "gold_line": "4.AA\t8AAL\t24aa", "gold_note_token": "8AAL", "gold_piece": "8AAL", "gold_token_id": 512, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 65, "line_char_start": 51, "line_index": 4, "line_text": "4.AA\t8AAL\t24aa", "operation": "replace", "pred_field": "8.AAL", "pred_local_replacement_line": "4.AA\t8.AAL\t24aa", "pred_note_token": "8.AAL", "pred_piece": "8.AAL", "pred_token_id": 2335, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 12.548667265648398, "LER": 66.02316284179688, "SER": 12.792596624931955}, "nll": 0.6928354501724243, "top1_error_count": 235, "top1_error_rate": 0.12785636561479868}}
+{"dataset_index": 42, "sample_id": "polish_00060", "source_ids": [], "target_token_count": 1528, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 62, "char_start": 59, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 66, "field_char_start": 59, "field_index": 0, "field_text": "8FF 8FL", "gold_field": "8FF 8FL", "gold_line": "8FF 8FL\t2.F\t4.f", "gold_note_token": "8FF", "gold_piece": "8FF", "gold_token_id": 276, "gold_token_index": 16, "gold_trimmed_index": 16, "group": "visual", "line_char_end": 74, "line_char_start": 59, "line_index": 4, "line_text": "8FF 8FL\t2.F\t4.f", "operation": "replace", "pred_field": "8FL 8FL", "pred_local_replacement_line": "8FL 8FL\t2.F\t4.f", "pred_note_token": "8FL", "pred_piece": "8FL", "pred_token_id": 433, "pred_trimmed_index": 16, "visually_grounded": true}, "metrics": {"CER": 6.0606060606060606, "LER": 43.457942962646484, "SER": 7.989521938441388}, "nll": 0.5164713859558105, "top1_error_count": 122, "top1_error_rate": 0.07984293193717278}}
+{"dataset_index": 43, "sample_id": "polish_00061", "source_ids": [], "target_token_count": 653, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 53, "char_start": 50, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 59, "field_char_start": 50, "field_index": 0, "field_text": "8e-( 8g(L", "gold_field": "8e-( 8g(L", "gold_line": "8e-( 8g(L\t8gg 8bb-L", "gold_note_token": null, "gold_piece": "8e-", "gold_token_id": 263, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 69, "line_char_start": 50, "line_index": 3, "line_text": "8e-( 8g(L\t8gg 8bb-L", "operation": "replace", "pred_field": "*( 8g(L", "pred_local_replacement_line": "*( 8g(L\t8gg 8bb-L", "pred_note_token": null, "pred_piece": "*", "pred_token_id": 11, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 12.202380952380953, "LER": 47.95918273925781, "SER": 17.33128834355828}, "nll": 0.8379897475242615, "top1_error_count": 113, "top1_error_rate": 0.17304747320061256}}
+{"dataset_index": 44, "sample_id": "polish_00062", "source_ids": [], "target_token_count": 781, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 16, "char_start": 15, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 15, "field_char_start": 8, "field_index": 1, "field_text": "*clefG2", "gold_field": "*clefG2", "gold_line": "*clefF4\t*clefG2\t*clefG2", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 3, "gold_trimmed_index": 3, "group": "structural", "line_char_end": 23, "line_char_start": 0, "line_index": 0, "line_text": "*clefF4\t*clefG2\t*clefG2", "operation": "replace", "pred_field": "*clefG2\n*clefG2", "pred_local_replacement_line": "*clefF4\t*clefG2\n*clefG2", "pred_note_token": null, "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 3, "visually_grounded": false}, "metrics": {"CER": 9.88562091503268, "LER": 40.18691635131836, "SER": 8.717948717948717}, "nll": 0.4546618163585663, "top1_error_count": 68, "top1_error_rate": 0.08706786171574904}}
+{"dataset_index": 45, "sample_id": "polish_00063", "source_ids": [], "target_token_count": 963, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 56, "char_start": 54, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 48, "field_index": 1, "field_text": "16qqddnL", "gold_field": "16qqddnL", "gold_line": ".\t16qqddnL", "gold_note_token": "16qqddnL", "gold_piece": "nL", "gold_token_id": 403, "gold_token_index": 11, "gold_trimmed_index": 11, "group": "visual", "line_char_end": 56, "line_char_start": 46, "line_index": 2, "line_text": ".\t16qqddnL", "operation": "replace", "pred_field": "16qqdd#LL", "pred_local_replacement_line": ".\t16qqdd#LL", "pred_note_token": "16qqdd#LL", "pred_piece": "#LL", "pred_token_id": 269, "pred_trimmed_index": 11, "visually_grounded": true}, "metrics": {"CER": 3.9287906691221606, "LER": 27.63157844543457, "SER": 5.925155925155925}, "nll": 0.34782788157463074, "top1_error_count": 57, "top1_error_rate": 0.059190031152647975}}
+{"dataset_index": 46, "sample_id": "polish_00064", "source_ids": [], "target_token_count": 928, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 53, "char_start": 49, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 53, "field_char_start": 49, "field_index": 1, "field_text": "8dd[", "gold_field": "8dd[", "gold_line": "8r\t8dd[", "gold_note_token": "8dd[", "gold_piece": "8dd[", "gold_token_id": 1303, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 53, "line_char_start": 46, "line_index": 3, "line_text": "8r\t8dd[", "operation": "replace", "pred_field": "8dd(", "pred_local_replacement_line": "8r\t8dd(", "pred_note_token": "8dd(", "pred_piece": "8dd(", "pred_token_id": 1124, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 9.605488850771868, "LER": 44.06779479980469, "SER": 13.160733549083064}, "nll": 0.815976083278656, "top1_error_count": 122, "top1_error_rate": 0.1314655172413793}}
+{"dataset_index": 47, "sample_id": "polish_00067", "source_ids": [], "target_token_count": 1138, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 43, "char_start": 39, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 43, "field_char_start": 39, "field_index": 1, "field_text": "16aL", "gold_field": "16aL", "gold_line": "8F(L\t16aL", "gold_note_token": "16aL", "gold_piece": "16aL", "gold_token_id": 602, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 43, "line_char_start": 34, "line_index": 2, "line_text": "8F(L\t16aL", "operation": "replace", "pred_field": "16aLL", "pred_local_replacement_line": "8F(L\t16aLL", "pred_note_token": "16aLL", "pred_piece": "16aLL", "pred_token_id": 338, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 12.164073550212164, "LER": 43.75, "SER": 14.951627088830255}, "nll": 0.9464731216430664, "top1_error_count": 169, "top1_error_rate": 0.14850615114235502}}
+{"dataset_index": 48, "sample_id": "polish_00068", "source_ids": [], "target_token_count": 973, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 57, "char_start": 53, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 57, "field_char_start": 53, "field_index": 1, "field_text": "4ff#", "gold_field": "4ff#", "gold_line": "4r\t4ff#", "gold_note_token": "4ff#", "gold_piece": "4ff#", "gold_token_id": 225, "gold_token_index": 14, "gold_trimmed_index": 14, "group": "visual", "line_char_end": 57, "line_char_start": 50, "line_index": 3, "line_text": "4r\t4ff#", "operation": "replace", "pred_field": "4ff#(", "pred_local_replacement_line": "4r\t4ff#(", "pred_note_token": "4ff#(", "pred_piece": "4ff#(", "pred_token_id": 2286, "pred_trimmed_index": 14, "visually_grounded": true}, "metrics": {"CER": 22.418136020151135, "LER": 64.64088439941406, "SER": 20.16460905349794}, "nll": 0.9882350564002991, "top1_error_count": 196, "top1_error_rate": 0.2014388489208633}}
+{"dataset_index": 49, "sample_id": "polish_00071", "source_ids": [], "target_token_count": 1433, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 48, "char_start": 46, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 48, "field_char_start": 46, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*^", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 51, "line_char_start": 46, "line_index": 2, "line_text": "*^\t*^", "operation": "replace", "pred_field": ".", "pred_local_replacement_line": ".\t*^", "pred_note_token": null, "pred_piece": ".", "pred_token_id": 14, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 11.836562763268745, "LER": 63.934425354003906, "SER": 13.756983240223464}, "nll": 0.7083817720413208, "top1_error_count": 197, "top1_error_rate": 0.1374738311235171}}
+{"dataset_index": 50, "sample_id": "polish_00072", "source_ids": [], "target_token_count": 1123, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "column_or_line_break", "char_end": 56, "char_start": 55, "component_category": "column_or_line_break", "component_flags": [], "component_group": "structural", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "4BB", "gold_field": "4BB", "gold_line": "4BB\t4d 4f# 4b", "gold_note_token": null, "gold_piece": "\t", "gold_token_id": 4, "gold_token_index": 13, "gold_trimmed_index": 13, "group": "structural", "line_char_end": 65, "line_char_start": 52, "line_index": 3, "line_text": "4BB\t4d 4f# 4b", "operation": "replace", "pred_field": "4BB`4d 4f# 4b", "pred_local_replacement_line": "4BB`4d 4f# 4b", "pred_note_token": null, "pred_piece": "`", "pred_token_id": 47, "pred_trimmed_index": 13, "visually_grounded": false}, "metrics": {"CER": 4.841149773071105, "LER": 29.255319595336914, "SER": 6.506238859180035}, "nll": 0.29712775349617004, "top1_error_count": 73, "top1_error_rate": 0.06500445235975066}}
+{"dataset_index": 51, "sample_id": "polish_00074", "source_ids": [], "target_token_count": 1355, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 34, "char_start": 30, "component_category": "duration_and_pitch_and_accidental", "component_flags": ["duration", "pitch", "accidental"], "component_group": "visual", "field_char_end": 36, "field_char_start": 30, "field_index": 0, "field_text": "8BBnq(", "gold_field": "8BBnq(", "gold_line": "8BBnq(\t.", "gold_note_token": "8BBnq(", "gold_piece": "8BBn", "gold_token_id": 2042, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 38, "line_char_start": 30, "line_index": 2, "line_text": "8BBnq(\t.", "operation": "replace", "pred_field": "=!|:q(", "pred_local_replacement_line": "=!|:q(\t.", "pred_note_token": "=!|:q(", "pred_piece": "=!|:", "pred_token_id": 400, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 8.832667554371948, "LER": 43.07692337036133, "SER": 9.527326440177253}, "nll": 0.47899264097213745, "top1_error_count": 129, "top1_error_rate": 0.0952029520295203}}
+{"dataset_index": 52, "sample_id": "polish_00078", "source_ids": [], "target_token_count": 1172, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 49, "char_start": 48, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 49, "field_char_start": 48, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*\t*^\t*", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "structural", "line_char_end": 56, "line_char_start": 48, "line_index": 3, "line_text": "*\t*\t*^\t*", "operation": "insert", "pred_field": "4D#", "pred_local_replacement_line": "4D#\t*\t*^\t*", "pred_note_token": null, "pred_piece": "4D#", "pred_token_id": 708, "pred_trimmed_index": 12, "visually_grounded": false}, "metrics": {"CER": 16.211878009630816, "LER": 46.75324630737305, "SER": 12.894961571306576}, "nll": 0.560175359249115, "top1_error_count": 151, "top1_error_rate": 0.128839590443686}}
+{"dataset_index": 53, "sample_id": "polish_00082", "source_ids": [], "target_token_count": 1027, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "8AA-L", "pred_local_replacement_line": "8AA-L\t*^", "pred_note_token": null, "pred_piece": "8AA-L", "pred_token_id": 901, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 14.451827242524917, "LER": 51.97368240356445, "SER": 10.33138401559454}, "nll": 0.5927410125732422, "top1_error_count": 106, "top1_error_rate": 0.10321324245374879}}
+{"dataset_index": 54, "sample_id": "polish_00084", "source_ids": [], "target_token_count": 860, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 54, "char_start": 50, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 57, "field_char_start": 50, "field_index": 0, "field_text": "32qqCCL", "gold_field": "32qqCCL", "gold_line": "32qqCCL\t32qqe-L", "gold_note_token": "32qqCCL", "gold_piece": "32qq", "gold_token_id": 1235, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 65, "line_char_start": 50, "line_index": 3, "line_text": "32qqCCL\t32qqe-L", "operation": "replace", "pred_field": "32CCCCL", "pred_local_replacement_line": "32CCCCL\t32qqe-L", "pred_note_token": "32CCCCL", "pred_piece": "32CC", "pred_token_id": 1559, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 11.097560975609756, "LER": 44.58598709106445, "SER": 13.271245634458673}, "nll": 0.7674280405044556, "top1_error_count": 114, "top1_error_rate": 0.1325581395348837}}
+{"dataset_index": 55, "sample_id": "polish_00085", "source_ids": [], "target_token_count": 1428, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 54, "char_start": 49, "component_category": "duration_and_tie_slur", "component_flags": ["duration", "tie_slur"], "component_group": "visual", "field_char_end": 54, "field_char_start": 49, "field_index": 2, "field_text": "8aa(L", "gold_field": "8aa(L", "gold_line": "4d\t4D 4A\t8aa(L\t4f# 4dd", "gold_note_token": "8aa(L", "gold_piece": "8aa(L", "gold_token_id": 1057, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 40, "line_index": 3, "line_text": "4d\t4D 4A\t8aa(L\t4f# 4dd", "operation": "replace", "pred_field": "8.aa", "pred_local_replacement_line": "4d\t4D 4A\t8.aa\t4f# 4dd", "pred_note_token": "8.aa", "pred_piece": "8.aa", "pred_token_id": 386, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 4.793028322440087, "LER": 24.8447208404541, "SER": 4.274702172389628}, "nll": 0.24518053233623505, "top1_error_count": 61, "top1_error_rate": 0.042717086834733894}}
+{"dataset_index": 56, "sample_id": "polish_00086", "source_ids": [], "target_token_count": 1170, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 44, "field_char_start": 42, "field_index": 0, "field_text": "4C", "gold_field": "4C", "gold_line": "4C\t4r", "gold_note_token": null, "gold_piece": "4C", "gold_token_id": 181, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "visual", "line_char_end": 47, "line_char_start": 42, "line_index": 2, "line_text": "4C\t4r", "operation": "replace", "pred_field": "*M4/4", "pred_local_replacement_line": "*M4/4\t4r", "pred_note_token": null, "pred_piece": "*M4/4", "pred_token_id": 229, "pred_trimmed_index": 8, "visually_grounded": true}, "metrics": {"CER": 9.875954198473282, "LER": 40.85365676879883, "SER": 9.067579127459368}, "nll": 0.4193149209022522, "top1_error_count": 106, "top1_error_rate": 0.0905982905982906}}
+{"dataset_index": 57, "sample_id": "polish_00087", "source_ids": [], "target_token_count": 1300, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 47, "char_start": 44, "component_category": "pitch_and_accidental", "component_flags": ["pitch", "accidental"], "component_group": "visual", "field_char_end": 56, "field_char_start": 44, "field_index": 1, "field_text": "8ccq( 8ee-q(", "gold_field": "8ccq( 8ee-q(", "gold_line": ".\t8ccq( 8ee-q(", "gold_note_token": "8ccq(", "gold_piece": "8cc", "gold_token_id": 82, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 56, "line_char_start": 42, "line_index": 2, "line_text": ".\t8ccq( 8ee-q(", "operation": "replace", "pred_field": "8dd-q( 8ee-q(", "pred_local_replacement_line": ".\t8dd-q( 8ee-q(", "pred_note_token": "8dd-q(", "pred_piece": "8dd-", "pred_token_id": 550, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 13.533507397737162, "LER": 53.80434799194336, "SER": 11.70130869899923}, "nll": 0.5588526129722595, "top1_error_count": 152, "top1_error_rate": 0.11692307692307692}}
+{"dataset_index": 58, "sample_id": "polish_00089", "source_ids": [], "target_token_count": 782, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 43, "char_start": 39, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 43, "field_char_start": 35, "field_index": 0, "field_text": "8A( 8c(L", "gold_field": "8A( 8c(L", "gold_line": "8A( 8c(L\t2.D\t16g#(L", "gold_note_token": "8c(L", "gold_piece": "8c(L", "gold_token_id": 1647, "gold_token_index": 15, "gold_trimmed_index": 15, "group": "visual", "line_char_end": 54, "line_char_start": 35, "line_index": 3, "line_text": "8A( 8c(L\t2.D\t16g#(L", "operation": "replace", "pred_field": "8A( 8cL", "pred_local_replacement_line": "8A( 8cL\t2.D\t16g#(L", "pred_note_token": "8cL", "pred_piece": "8cL", "pred_token_id": 320, "pred_trimmed_index": 15, "visually_grounded": true}, "metrics": {"CER": 12.204989885367498, "LER": 49.6894416809082, "SER": 14.980793854033292}, "nll": 1.0461078882217407, "top1_error_count": 117, "top1_error_rate": 0.149616368286445}}
+{"dataset_index": 59, "sample_id": "polish_00090", "source_ids": [], "target_token_count": 1125, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 276, "char_start": 272, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 276, "field_char_start": 272, "field_index": 0, "field_text": "8AAL", "gold_field": "8AAL", "gold_line": "8AAL\t8.r\t4ee", "gold_note_token": "8AAL", "gold_piece": "8AAL", "gold_token_id": 512, "gold_token_index": 164, "gold_trimmed_index": 164, "group": "visual", "line_char_end": 284, "line_char_start": 272, "line_index": 22, "line_text": "8AAL\t8.r\t4ee", "operation": "replace", "pred_field": "8AA", "pred_local_replacement_line": "8AA\t8.r\t4ee", "pred_note_token": "8AA", "pred_piece": "8AA", "pred_token_id": 210, "pred_trimmed_index": 164, "visually_grounded": true}, "metrics": {"CER": 9.050772626931568, "LER": 37.583892822265625, "SER": 7.829181494661921}, "nll": 0.4268281161785126, "top1_error_count": 88, "top1_error_rate": 0.07822222222222222}}
+{"dataset_index": 60, "sample_id": "polish_00092", "source_ids": [], "target_token_count": 1153, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 62, "char_start": 61, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 62, "field_char_start": 52, "field_index": 1, "field_text": "4d-( 4dd-(", "gold_field": "4d-( 4dd-(", "gold_line": "8BB-\t4d-( 4dd-(", "gold_note_token": "4dd-(", "gold_piece": "(", "gold_token_id": 9, "gold_token_index": 18, "gold_trimmed_index": 18, "group": "visual", "line_char_end": 62, "line_char_start": 47, "line_index": 3, "line_text": "8BB-\t4d-( 4dd-(", "operation": "replace", "pred_field": "4d-( 4dd-\n", "pred_local_replacement_line": "8BB-\t4d-( 4dd-\n", "pred_note_token": "4dd-\n", "pred_piece": "\n", "pred_token_id": 5, "pred_trimmed_index": 18, "visually_grounded": true}, "metrics": {"CER": 5.421103581800581, "LER": 33.519554138183594, "SER": 7.118055555555555}, "nll": 0.5455198287963867, "top1_error_count": 82, "top1_error_rate": 0.07111882046834345}}
+{"dataset_index": 61, "sample_id": "polish_00093", "source_ids": [], "target_token_count": 1047, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 555, "char_start": 552, "component_category": "tie_slur_only", "component_flags": ["tie_slur"], "component_group": "visual", "field_char_end": 564, "field_char_start": 552, "field_index": 1, "field_text": "8e[ 8b[ 8ee[", "gold_field": "8e[ 8b[ 8ee[", "gold_line": "8r\t8e[ 8b[ 8ee[\t8ee", "gold_note_token": "8e[", "gold_piece": "8e[", "gold_token_id": 1374, "gold_token_index": 340, "gold_trimmed_index": 340, "group": "visual", "line_char_end": 568, "line_char_start": 549, "line_index": 45, "line_text": "8r\t8e[ 8b[ 8ee[\t8ee", "operation": "replace", "pred_field": "8e 8b[ 8ee[", "pred_local_replacement_line": "8r\t8e 8b[ 8ee[\t8ee", "pred_note_token": "8e", "pred_piece": "8e", "pred_token_id": 90, "pred_trimmed_index": 340, "visually_grounded": true}, "metrics": {"CER": 8.752217622708457, "LER": 34.30656814575195, "SER": 7.45697896749522}, "nll": 0.37567269802093506, "top1_error_count": 78, "top1_error_rate": 0.07449856733524356}}
+{"dataset_index": 62, "sample_id": "polish_00094", "source_ids": [], "target_token_count": 773, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "duration", "char_end": 44, "char_start": 40, "component_category": "pitch_only", "component_flags": ["pitch"], "component_group": "visual", "field_char_end": 47, "field_char_start": 40, "field_index": 1, "field_text": "32qqggL", "gold_field": "32qqggL", "gold_line": ".\t32qqggL", "gold_note_token": "32qqggL", "gold_piece": "32qq", "gold_token_id": 1235, "gold_token_index": 10, "gold_trimmed_index": 10, "group": "visual", "line_char_end": 47, "line_char_start": 38, "line_index": 2, "line_text": ".\t32qqggL", "operation": "replace", "pred_field": "32ggggL", "pred_local_replacement_line": ".\t32ggggL", "pred_note_token": "32ggggL", "pred_piece": "32gg", "pred_token_id": 477, "pred_trimmed_index": 10, "visually_grounded": true}, "metrics": {"CER": 10.067567567567567, "LER": 45.16128921508789, "SER": 12.176165803108809}, "nll": 0.7969435453414917, "top1_error_count": 94, "top1_error_rate": 0.12160413971539456}}
+{"dataset_index": 63, "sample_id": "polish_00095", "source_ids": [], "target_token_count": 1073, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "interpretation_or_control", "char_end": 43, "char_start": 42, "component_category": "interpretation_or_control", "component_flags": [], "component_group": "structural", "field_char_end": 43, "field_char_start": 42, "field_index": 0, "field_text": "*", "gold_field": "*", "gold_line": "*\t*^", "gold_note_token": null, "gold_piece": "*", "gold_token_id": 11, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 46, "line_char_start": 42, "line_index": 2, "line_text": "*\t*^", "operation": "replace", "pred_field": "4D-", "pred_local_replacement_line": "4D-\t*^", "pred_note_token": null, "pred_piece": "4D-", "pred_token_id": 556, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 18.15803244374673, "LER": 43.29896926879883, "SER": 13.059701492537313}, "nll": 0.6930850744247437, "top1_error_count": 140, "top1_error_rate": 0.13047530288909598}}
+{"dataset_index": 64, "sample_id": "polish_00096", "source_ids": [], "target_token_count": 935, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "accidental", "char_end": 55, "char_start": 52, "component_category": "accidental_only", "component_flags": ["accidental"], "component_group": "visual", "field_char_end": 55, "field_char_start": 52, "field_index": 0, "field_text": "2en", "gold_field": "2en", "gold_line": "2en\t4F#]\t4ggn\t2cc#", "gold_note_token": "2en", "gold_piece": "2en", "gold_token_id": 2374, "gold_token_index": 12, "gold_trimmed_index": 12, "group": "visual", "line_char_end": 70, "line_char_start": 52, "line_index": 3, "line_text": "2en\t4F#]\t4ggn\t2cc#", "operation": "replace", "pred_field": "2e", "pred_local_replacement_line": "2e\t4F#]\t4ggn\t2cc#", "pred_note_token": "2e", "pred_piece": "2e", "pred_token_id": 296, "pred_trimmed_index": 12, "visually_grounded": true}, "metrics": {"CER": 4.252577319587629, "LER": 34.28571319580078, "SER": 5.032119914346895}, "nll": 0.2973078191280365, "top1_error_count": 47, "top1_error_rate": 0.05026737967914439}}
+{"dataset_index": 65, "sample_id": "polish_00097", "source_ids": [], "target_token_count": 1050, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "tie_slur", "char_end": 61, "char_start": 57, "component_category": "note_other", "component_flags": [], "component_group": "visual", "field_char_end": 63, "field_char_start": 57, "field_index": 1, "field_text": "16gg[L", "gold_field": "16gg[L", "gold_line": "8c 8eJ\t16gg[L\t16dd(L", "gold_note_token": "16gg[L", "gold_piece": "16gg", "gold_token_id": 146, "gold_token_index": 22, "gold_trimmed_index": 22, "group": "visual", "line_char_end": 70, "line_char_start": 50, "line_index": 3, "line_text": "8c 8eJ\t16gg[L\t16dd(L", "operation": "replace", "pred_field": "16ggLL[L", "pred_local_replacement_line": "8c 8eJ\t16ggLL[L\t16dd(L", "pred_note_token": "16ggLL[L", "pred_piece": "16ggLL", "pred_token_id": 382, "pred_trimmed_index": 22, "visually_grounded": true}, "metrics": {"CER": 7.877330126277811, "LER": 31.34328269958496, "SER": 6.7683508102955185}, "nll": 0.37606099247932434, "top1_error_count": 71, "top1_error_rate": 0.06761904761904762}}
+{"dataset_index": 66, "sample_id": "polish_00100", "source_ids": [], "target_token_count": 1185, "teacher_forced": {"exact_match": false, "first_divergence": {"category": "spine_operation", "char_end": 36, "char_start": 34, "component_category": "spine_operation", "component_flags": [], "component_group": "structural", "field_char_end": 36, "field_char_start": 34, "field_index": 0, "field_text": "*^", "gold_field": "*^", "gold_line": "*^\t*", "gold_note_token": null, "gold_piece": "*^", "gold_token_id": 191, "gold_token_index": 8, "gold_trimmed_index": 8, "group": "structural", "line_char_end": 38, "line_char_start": 34, "line_index": 2, "line_text": "*^\t*", "operation": "replace", "pred_field": "16DD", "pred_local_replacement_line": "16DD\t*", "pred_note_token": null, "pred_piece": "16DD", "pred_token_id": 526, "pred_trimmed_index": 8, "visually_grounded": false}, "metrics": {"CER": 13.57769747031492, "LER": 52.91005325317383, "SER": 13.260135135135135}, "nll": 0.6973803043365479, "top1_error_count": 157, "top1_error_rate": 0.13248945147679325}}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/summary.json

@@ -0,0 +1,454 @@
+{
+  "checkpoint": "/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_4.ckpt",
+  "component_error_buckets": [
+    {
+      "category": "note_other",
+      "error_share": 0.31903086085544125,
+      "errors": 2357,
+      "errors_per_1k_tokens": 31.631215191572167,
+      "errors_per_page_mean": 35.17910447761194,
+      "errors_per_page_median": 36.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "tie_slur_only",
+      "error_share": 0.22387655657823496,
+      "errors": 1654,
+      "errors_per_1k_tokens": 22.19687311279608,
+      "errors_per_page_mean": 24.686567164179106,
+      "errors_per_page_median": 25.0,
+      "page_share": 1.0,
+      "pages_with_error": 67
+    },
+    {
+      "category": "duration_only",
+      "error_share": 0.10327558202490525,
+      "errors": 763,
+      "errors_per_1k_tokens": 10.239549084077032,
+      "errors_per_page_mean": 11.388059701492537,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.8955223880597015,
+      "pages_with_error": 60
+    },
+    {
+      "category": "column_or_line_break",
+      "error_share": 0.05481862479696806,
+      "errors": 405,
+      "errors_per_1k_tokens": 5.435147285781387,
+      "errors_per_page_mean": 6.044776119402985,
+      "errors_per_page_median": 5.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "accidental_only",
+      "error_share": 0.04155387114239307,
+      "errors": 307,
+      "errors_per_1k_tokens": 4.119975843789841,
+      "errors_per_page_mean": 4.582089552238806,
+      "errors_per_page_median": 3.0,
+      "page_share": 0.9253731343283582,
+      "pages_with_error": 62
+    },
+    {
+      "category": "pitch_only",
+      "error_share": 0.0334325933946941,
+      "errors": 247,
+      "errors_per_1k_tokens": 3.3147688384888947,
+      "errors_per_page_mean": 3.6865671641791047,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.8208955223880597,
+      "pages_with_error": 55
+    },
+    {
+      "category": "interpretation_or_control",
+      "error_share": 0.03316188413643747,
+      "errors": 245,
+      "errors_per_1k_tokens": 3.2879286049788634,
+      "errors_per_page_mean": 3.656716417910448,
+      "errors_per_page_median": 2.0,
+      "page_share": 0.746268656716418,
+      "pages_with_error": 50
+    },
+    {
+      "category": "pitch_and_accidental",
+      "error_share": 0.025446670276123445,
+      "errors": 188,
+      "errors_per_1k_tokens": 2.5229819499429644,
+      "errors_per_page_mean": 2.8059701492537314,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5522388059701493,
+      "pages_with_error": 37
+    },
+    {
+      "category": "spine_operation",
+      "error_share": 0.023280996210070383,
+      "errors": 172,
+      "errors_per_1k_tokens": 2.3082600818627124,
+      "errors_per_page_mean": 2.5671641791044775,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "rest_note_confusion",
+      "error_share": 0.02301028695181375,
+      "errors": 170,
+      "errors_per_1k_tokens": 2.2814198483526806,
+      "errors_per_page_mean": 2.537313432835821,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.5223880597014925,
+      "pages_with_error": 35
+    },
+    {
+      "category": "duration_and_pitch",
+      "error_share": 0.014347590687601516,
+      "errors": 106,
+      "errors_per_1k_tokens": 1.4225323760316715,
+      "errors_per_page_mean": 1.5820895522388059,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.582089552238806,
+      "pages_with_error": 39
+    },
+    {
+      "category": "chord_structure_and_tie_slur",
+      "error_share": 0.013264753654574987,
+      "errors": 98,
+      "errors_per_1k_tokens": 1.3151714419915452,
+      "errors_per_page_mean": 1.462686567164179,
+      "errors_per_page_median": 1.0,
+      "page_share": 0.6119402985074627,
+      "pages_with_error": 41
+    },
+    {
+      "category": "chord_structure_and_articulation",
+      "error_share": 0.012723335138061722,
+      "errors": 94,
+      "errors_per_1k_tokens": 1.2614909749714822,
+      "errors_per_page_mean": 1.4029850746268657,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.373134328358209,
+      "pages_with_error": 25
+    },
+    {
+      "category": "duration_and_pitch_and_accidental",
+      "error_share": 0.008662696264212236,
+      "errors": 64,
+      "errors_per_1k_tokens": 0.8588874723210091,
+      "errors_per_page_mean": 0.9552238805970149,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.417910447761194,
+      "pages_with_error": 28
+    },
+    {
+      "category": "null_token",
+      "error_share": 0.008662696264212236,
+      "errors": 64,
+      "errors_per_1k_tokens": 0.8588874723210091,
+      "errors_per_page_mean": 0.9552238805970149,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.40298507462686567,
+      "pages_with_error": 27
+    },
+    {
+      "category": "duration_and_tie_slur",
+      "error_share": 0.00852734163508392,
+      "errors": 63,
+      "errors_per_1k_tokens": 0.8454673555659934,
+      "errors_per_page_mean": 0.9402985074626866,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.31343283582089554,
+      "pages_with_error": 21
+    },
+    {
+      "category": "termination",
+      "error_share": 0.00568489442338928,
+      "errors": 42,
+      "errors_per_1k_tokens": 0.5636449037106623,
+      "errors_per_page_mean": 0.6268656716417911,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.43283582089552236,
+      "pages_with_error": 29
+    },
+    {
+      "category": "accidental_and_tie_slur",
+      "error_share": 0.0055495397942609635,
+      "errors": 41,
+      "errors_per_1k_tokens": 0.5502247869556465,
+      "errors_per_page_mean": 0.6119402985074627,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.3283582089552239,
+      "pages_with_error": 22
+    },
+    {
+      "category": "chord_structure_only",
+      "error_share": 0.005278830536004331,
+      "errors": 39,
+      "errors_per_1k_tokens": 0.523384553445615,
+      "errors_per_page_mean": 0.582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2835820895522388,
+      "pages_with_error": 19
+    },
+    {
+      "category": "duration_and_accidental",
+      "error_share": 0.004466702761234434,
+      "errors": 33,
+      "errors_per_1k_tokens": 0.4428638529155204,
+      "errors_per_page_mean": 0.4925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.208955223880597,
+      "pages_with_error": 14
+    },
+    {
+      "category": "articulation_only",
+      "error_share": 0.004466702761234434,
+      "errors": 33,
+      "errors_per_1k_tokens": 0.4428638529155204,
+      "errors_per_page_mean": 0.4925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "pitch_and_accidental_and_tie_slur",
+      "error_share": 0.003789929615592853,
+      "errors": 28,
+      "errors_per_1k_tokens": 0.3757632691404415,
+      "errors_per_page_mean": 0.417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2537313432835821,
+      "pages_with_error": 17
+    },
+    {
+      "category": "tie_slur_and_articulation",
+      "error_share": 0.0032485110990795887,
+      "errors": 24,
+      "errors_per_1k_tokens": 0.32208280212037843,
+      "errors_per_page_mean": 0.3582089552238806,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.22388059701492538,
+      "pages_with_error": 15
+    },
+    {
+      "category": "pitch_and_tie_slur",
+      "error_share": 0.0031131564699512725,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.2835820895522388,
+      "pages_with_error": 19
+    },
+    {
+      "category": "chord_structure_and_accidental_and_articulation",
+      "error_share": 0.0031131564699512725,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "barline",
+      "error_share": 0.0031131564699512725,
+      "errors": 23,
+      "errors_per_1k_tokens": 0.3086626853653627,
+      "errors_per_page_mean": 0.34328358208955223,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.23880597014925373,
+      "pages_with_error": 16
+    },
+    {
+      "category": "duration_and_pitch_and_tie_slur",
+      "error_share": 0.00284244721169464,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1044776119402985,
+      "pages_with_error": 7
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur",
+      "error_share": 0.00284244721169464,
+      "errors": 21,
+      "errors_per_1k_tokens": 0.28182245185533117,
+      "errors_per_page_mean": 0.31343283582089554,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.1791044776119403,
+      "pages_with_error": 12
+    },
+    {
+      "category": "clef",
+      "error_share": 0.002030319436924743,
+      "errors": 15,
+      "errors_per_1k_tokens": 0.20130175132523653,
+      "errors_per_page_mean": 0.22388059701492538,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.14925373134328357,
+      "pages_with_error": 10
+    },
+    {
+      "category": "chord_structure_and_tie_slur_and_articulation",
+      "error_share": 0.001488900920411478,
+      "errors": 11,
+      "errors_per_1k_tokens": 0.14762128430517346,
+      "errors_per_page_mean": 0.16417910447761194,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.11940298507462686,
+      "pages_with_error": 8
+    },
+    {
+      "category": "chord_structure_and_accidental",
+      "error_share": 0.0006767731456415809,
+      "errors": 5,
+      "errors_per_1k_tokens": 0.06710058377507884,
+      "errors_per_page_mean": 0.07462686567164178,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "duration_and_pitch_and_accidental_and_tie_slur",
+      "error_share": 0.0005414185165132648,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "accidental_and_articulation",
+      "error_share": 0.0005414185165132648,
+      "errors": 4,
+      "errors_per_1k_tokens": 0.05368046702006307,
+      "errors_per_page_mean": 0.05970149253731343,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.05970149253731343,
+      "pages_with_error": 4
+    },
+    {
+      "category": "chord_structure_and_accidental_and_tie_slur_and_articulation",
+      "error_share": 0.0001353546291283162,
+      "errors": 1,
+      "errors_per_1k_tokens": 0.013420116755015768,
+      "errors_per_page_mean": 0.014925373134328358,
+      "errors_per_page_median": 0.0,
+      "page_share": 0.014925373134328358,
+      "pages_with_error": 1
+    }
+  ],
+  "component_error_categories": {
+    "accidental_and_articulation": 4,
+    "accidental_and_tie_slur": 41,
+    "accidental_only": 307,
+    "articulation_only": 33,
+    "barline": 23,
+    "chord_structure_and_accidental": 5,
+    "chord_structure_and_accidental_and_articulation": 23,
+    "chord_structure_and_accidental_and_tie_slur": 21,
+    "chord_structure_and_accidental_and_tie_slur_and_articulation": 1,
+    "chord_structure_and_articulation": 94,
+    "chord_structure_and_tie_slur": 98,
+    "chord_structure_and_tie_slur_and_articulation": 11,
+    "chord_structure_only": 39,
+    "clef": 15,
+    "column_or_line_break": 405,
+    "duration_and_accidental": 33,
+    "duration_and_pitch": 106,
+    "duration_and_pitch_and_accidental": 64,
+    "duration_and_pitch_and_accidental_and_tie_slur": 4,
+    "duration_and_pitch_and_tie_slur": 21,
+    "duration_and_tie_slur": 63,
+    "duration_only": 763,
+    "interpretation_or_control": 245,
+    "note_other": 2357,
+    "null_token": 64,
+    "pitch_and_accidental": 188,
+    "pitch_and_accidental_and_tie_slur": 28,
+    "pitch_and_tie_slur": 23,
+    "pitch_only": 247,
+    "rest_note_confusion": 170,
+    "spine_operation": 172,
+    "termination": 42,
+    "tie_slur_and_articulation": 24,
+    "tie_slur_only": 1654
+  },
+  "component_error_groups": {
+    "structural": 928,
+    "visual": 6460
+  },
+  "dataset": {
+    "data_files": [
+      "data/eval-00000-of-00001.parquet"
+    ],
+    "image_column": "image_original_real",
+    "kind": "hf_parquet_split",
+    "name": "polish67-original",
+    "repo_id": "cminst/transcoda-polish-filtered-eval",
+    "rows_available": 67,
+    "source": "cminst/transcoda-polish-filtered-eval",
+    "split": "eval",
+    "transcription_column": "transcription"
+  },
+  "elapsed_sec": 12.814844940090552,
+  "generation_settings": {
+    "early_stopping": null,
+    "length_penalty": null,
+    "num_beams": null,
+    "repetition_penalty": null,
+    "strategy": null
+  },
+  "gold_error_categories": {
+    "accidental": 3145,
+    "articulation": 53,
+    "barline": 22,
+    "clef": 15,
+    "column_or_line_break": 406,
+    "duration": 1011,
+    "interpretation_or_control": 138,
+    "note_other": 209,
+    "null_token": 65,
+    "pitch": 20,
+    "rest": 279,
+    "spine_operation": 205,
+    "termination": 28,
+    "tie_slur": 1792
+  },
+  "gold_error_groups": {
+    "structural": 842,
+    "visual": 6546
+  },
+  "metrics": {
+    "teacher_forced": {
+      "CER": 9.550797462463379,
+      "LER": 42.2453498840332,
+      "SER": 10.130560874938965,
+      "exact_match_rate": 0.0,
+      "top1_error_rate": 0.0991478225860565,
+      "top1_errors": 7388,
+      "top1_errors_per_page": 110.26865671641791
+    }
+  },
+  "mode": "teacher_forced",
+  "outputs": {
+    "free_generation_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/free_generation_predictions.jsonl",
+    "gold_prefix_errors": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/gold_prefix_errors.jsonl",
+    "per_page_records": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/per_page_records.jsonl",
+    "summary": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/summary.json",
+    "summary_md": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/summary.md",
+    "teacher_forced_error_report": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/teacher_forced_error_report.md",
+    "teacher_forced_predictions": "/gpfs/scrubbed/qxli2/rescore/results/eval-artifacts/multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/teacher_forced_predictions.jsonl"
+  },
+  "pages": 67,
+  "reconstruction": {
+    "matched_unk_preserved": 2,
+    "predicted_eos": 41,
+    "predicted_unk_literal": 104
+  },
+  "tokens": 74515
+}

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/summary.md

@@ -0,0 +1,43 @@
+# Transcoda Dataset Eval
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_4.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Dataset kind: `hf_parquet_split`
+- Mode: `teacher_forced`
+- Pages: `67`
+- Tokens: `74,515`
+- Elapsed: `12.8s`
+
+## Metrics
+
+| mode | CER | SER | LER | exact pages | top-1 error rate | top-1 errors/page |
+|---|---:|---:|---:|---:|---:|---:|
+| `teacher_forced` | 9.550797 | 10.130561 | 42.245350 | 0.000% | 9.9148% | 110.269 |
+
+## Teacher-Forced Error Categories
+
+| category | errors |
+|---|---:|
+| `accidental` | 3,145 |
+| `tie_slur` | 1,792 |
+| `duration` | 1,011 |
+| `column_or_line_break` | 406 |
+| `rest` | 279 |
+| `note_other` | 209 |
+| `spine_operation` | 205 |
+| `interpretation_or_control` | 138 |
+| `null_token` | 65 |
+| `articulation` | 53 |
+| `termination` | 28 |
+| `barline` | 22 |
+| `pitch` | 20 |
+| `clef` | 15 |
+
+## Outputs
+
+- `summary.json`
+- `per_page_records.jsonl`
+- `teacher_forced_predictions.jsonl` when teacher-forced mode is enabled
+- `free_generation_predictions.jsonl` when free-generation mode is enabled
+- `gold_prefix_errors.jsonl` when teacher-forced mode is enabled
+- `teacher_forced_error_report.md` when teacher-forced mode is enabled

multiview-renderswitch-encoder-only/polish67-original-teacher-forced/runs/epoch_4/teacher_forced_error_report.md

@@ -0,0 +1,522 @@
+# Teacher-Forced Error Report
+
+- Checkpoint: `/gpfs/scrubbed/qxli2/rescore/results/runs/transcoda-multiview-renderswitch-encoder-only-from-musvit-243k-8ep-h200x4/checkpoint_epoch_4.ckpt`
+- Dataset: `cminst/transcoda-polish-filtered-eval`
+- Split: `eval`
+- Pages: `67`
+- Tokens: `74,515`
+- Top-1 errors: `7,388`
+- Top-1 error rate: `9.915%`
+
+## Component Buckets
+
+| bucket | errors | error share | mean/page | median/page | pages | errors/1k tok |
+|---|---:|---:|---:|---:|---:|---:|
+| `note_other` | 2,357 | 31.90% | 35.179 | 36.000 | 67 (100.0%) | 31.631 |
+| `tie_slur_only` | 1,654 | 22.39% | 24.687 | 25.000 | 67 (100.0%) | 22.197 |
+| `duration_only` | 763 | 10.33% | 11.388 | 3.000 | 60 (89.6%) | 10.240 |
+| `column_or_line_break` | 405 | 5.48% | 6.045 | 5.000 | 62 (92.5%) | 5.435 |
+| `accidental_only` | 307 | 4.16% | 4.582 | 3.000 | 62 (92.5%) | 4.120 |
+| `pitch_only` | 247 | 3.34% | 3.687 | 2.000 | 55 (82.1%) | 3.315 |
+| `interpretation_or_control` | 245 | 3.32% | 3.657 | 2.000 | 50 (74.6%) | 3.288 |
+| `pitch_and_accidental` | 188 | 2.54% | 2.806 | 1.000 | 37 (55.2%) | 2.523 |
+| `spine_operation` | 172 | 2.33% | 2.567 | 1.000 | 41 (61.2%) | 2.308 |
+| `rest_note_confusion` | 170 | 2.30% | 2.537 | 1.000 | 35 (52.2%) | 2.281 |
+| `duration_and_pitch` | 106 | 1.43% | 1.582 | 1.000 | 39 (58.2%) | 1.423 |
+| `chord_structure_and_tie_slur` | 98 | 1.33% | 1.463 | 1.000 | 41 (61.2%) | 1.315 |
+| `chord_structure_and_articulation` | 94 | 1.27% | 1.403 | 0.000 | 25 (37.3%) | 1.261 |
+| `duration_and_pitch_and_accidental` | 64 | 0.87% | 0.955 | 0.000 | 28 (41.8%) | 0.859 |
+| `null_token` | 64 | 0.87% | 0.955 | 0.000 | 27 (40.3%) | 0.859 |
+| `duration_and_tie_slur` | 63 | 0.85% | 0.940 | 0.000 | 21 (31.3%) | 0.845 |
+| `termination` | 42 | 0.57% | 0.627 | 0.000 | 29 (43.3%) | 0.564 |
+| `accidental_and_tie_slur` | 41 | 0.55% | 0.612 | 0.000 | 22 (32.8%) | 0.550 |
+| `chord_structure_only` | 39 | 0.53% | 0.582 | 0.000 | 19 (28.4%) | 0.523 |
+| `duration_and_accidental` | 33 | 0.45% | 0.493 | 0.000 | 14 (20.9%) | 0.443 |
+| `articulation_only` | 33 | 0.45% | 0.493 | 0.000 | 12 (17.9%) | 0.443 |
+| `pitch_and_accidental_and_tie_slur` | 28 | 0.38% | 0.418 | 0.000 | 17 (25.4%) | 0.376 |
+| `tie_slur_and_articulation` | 24 | 0.32% | 0.358 | 0.000 | 15 (22.4%) | 0.322 |
+| `pitch_and_tie_slur` | 23 | 0.31% | 0.343 | 0.000 | 19 (28.4%) | 0.309 |
+| `chord_structure_and_accidental_and_articulation` | 23 | 0.31% | 0.343 | 0.000 | 8 (11.9%) | 0.309 |
+| `barline` | 23 | 0.31% | 0.343 | 0.000 | 16 (23.9%) | 0.309 |
+| `duration_and_pitch_and_tie_slur` | 21 | 0.28% | 0.313 | 0.000 | 7 (10.4%) | 0.282 |
+| `chord_structure_and_accidental_and_tie_slur` | 21 | 0.28% | 0.313 | 0.000 | 12 (17.9%) | 0.282 |
+| `clef` | 15 | 0.20% | 0.224 | 0.000 | 10 (14.9%) | 0.201 |
+| `chord_structure_and_tie_slur_and_articulation` | 11 | 0.15% | 0.164 | 0.000 | 8 (11.9%) | 0.148 |
+| `chord_structure_and_accidental` | 5 | 0.07% | 0.075 | 0.000 | 4 (6.0%) | 0.067 |
+| `duration_and_pitch_and_accidental_and_tie_slur` | 4 | 0.05% | 0.060 | 0.000 | 4 (6.0%) | 0.054 |
+| `accidental_and_articulation` | 4 | 0.05% | 0.060 | 0.000 | 4 (6.0%) | 0.054 |
+| `chord_structure_and_accidental_and_tie_slur_and_articulation` | 1 | 0.01% | 0.015 | 0.000 | 1 (1.5%) | 0.013 |
+
+## Legacy Regex Buckets
+
+These are retained for compatibility and should not be treated as component diffs.
+
+| bucket | errors |
+|---|---:|
+| `accidental` | 3,145 |
+| `tie_slur` | 1,792 |
+| `duration` | 1,011 |
+| `column_or_line_break` | 406 |
+| `rest` | 279 |
+| `note_other` | 209 |
+| `spine_operation` | 205 |
+| `interpretation_or_control` | 138 |
+| `null_token` | 65 |
+| `articulation` | 53 |
+| `termination` | 28 |
+| `barline` | 22 |
+| `pitch` | 20 |
+| `clef` | 15 |
+
+## Error Samples
+
+### Sample 1
+
+```text
+.	.	16b)J	.
+teacher-forced top-1 local replacement:
+.	.	16b)JJ)J	.
+```
+
+### Sample 2
+
+```text
+.	32qqBBnL
+teacher-forced top-1 local replacement:
+.	32BBBBnL
+```
+
+### Sample 3
+
+```text
+.	16qqddJ
+teacher-forced top-1 local replacement:
+.	16qqddJJ
+```
+
+### Sample 4
+
+```text
+8f( 8a-(L	4aa-)
+teacher-forced top-1 local replacement:
+8f(	8a-(L	4aa-)
+```
+
+### Sample 5
+
+```text
+4FFn 4C# 4B-	8e
+teacher-forced top-1 local replacement:
+4FFn:4C# 4B-	8e
+```
+
+### Sample 6
+
+```text
+8c 8fJ	16cc[L	.
+teacher-forced top-1 local replacement:
+8c 8fJ	16cc[LL	.
+```
+
+### Sample 7
+
+```text
+.	12g-	.
+teacher-forced top-1 local replacement:
+.	8g-	.
+```
+
+### Sample 8
+
+```text
+8CC#	16cc#(L
+teacher-forced top-1 local replacement:
+8CC#	16cc#(LL
+```
+
+### Sample 9
+
+```text
+16A-L	2.AA-[	16cL	8ee-
+teacher-forced top-1 local replacement:
+16A-LL	2.AA-[	16cL	8ee-
+```
+
+### Sample 10
+
+```text
+8CC 8GG 8BBL	16g(L	4.f
+teacher-forced top-1 local replacement:
+8CC 8GG 8BBL	16g(LL	4.f
+```
+
+### Sample 11
+
+```text
+4B-( 4d-(	16g(L
+teacher-forced top-1 local replacement:
+4B-( 4d-(	16g(LL(L
+```
+
+### Sample 12
+
+```text
+.	48dddJ	.
+teacher-forced top-1 local replacement:
+.	32ddd8dddJ	.
+```
+
+### Sample 13
+
+```text
+.	16G 16B-J	16b- 16ffJ
+teacher-forced top-1 local replacement:
+.	16G 16cJ	16b- 16ffJ
+```
+
+### Sample 14
+
+```text
+4f) 4a-)	.	24cc]L
+teacher-forced top-1 local replacement:
+4f) 4a-		.	24cc]L
+```
+
+### Sample 15
+
+```text
+==	==
+teacher-forced top-1 local replacement:
+=	==
+```
+
+### Sample 16
+
+```text
+.	.	16qqccL	.
+teacher-forced top-1 local replacement:
+.	.	16qqccLL	.
+```
+
+
+## Sample Metadata
+
+### Sample 1
+
+- Sample: `polish_00071`
+- Row/token: `49` / `1365`
+- Line/field: `175:2`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16b` -> `16b)JJ`
+- Gold note -> top-1 note: `16b)J` -> `16b)JJ)J`
+- Gold token likelihood: rank `34`, p `0.006%`
+- Predicted token likelihood: p `97.342%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1274 | `"16b)JJ"` | 97.342% |
+| 2 | 2292 | `"16b-)JJ"` | 0.544% |
+| 3 | 1622 | `"16cc#)JJ"` | 0.257% |
+| 4 | 1910 | `"16b)"` | 0.213% |
+| 5 | 1094 | `"16a)JJ"` | 0.188% |
+
+### Sample 2
+
+- Sample: `polish_00084`
+- Row/token: `54` / `727`
+- Line/field: `129:1`
+- Component bucket: `pitch_only`; flags: `pitch`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `32qq` -> `32BB`
+- Gold note -> top-1 note: `32qqBBnL` -> `32BBBBnL`
+- Gold token likelihood: rank `2`, p `38.350%`
+- Predicted token likelihood: p `49.242%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1191 | `"32BB"` | 49.242% |
+| 2 | 1235 | `"32qq"` | 38.350% |
+| 3 | 20 | `"4"` | 1.633% |
+| 4 | 1414 | `"32BB-"` | 1.312% |
+| 5 | 528 | `"16qq"` | 1.158% |
+
+### Sample 3
+
+- Sample: `polish_00025`
+- Row/token: `16` / `241`
+- Line/field: `39:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `note_other`
+- Gold piece -> top-1 piece: `J` -> `JJ`
+- Gold note -> top-1 note: `16qqddJ` -> `16qqddJJ`
+- Gold token likelihood: rank `5`, p `0.430%`
+- Predicted token likelihood: p `92.892%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 80 | `"JJ"` | 92.892% |
+| 2 | 1487 | `"JJJJ"` | 0.626% |
+| 3 | 321 | `"JJJ"` | 0.472% |
+| 4 | 258 | `"#JJ"` | 0.444% |
+| 5 | 363 | `"]JJ"` | 0.430% |
+
+### Sample 4
+
+- Sample: `polish_00043`
+- Row/token: `28` / `18`
+- Line/field: `4:0`
+- Component bucket: `chord_structure_only`; flags: `chord_structure`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: ` ` -> `	`
+- Gold note -> top-1 note: `8f(` -> `8f(`
+- Gold token likelihood: rank `2`, p `14.948%`
+- Predicted token likelihood: p `80.807%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 4 | `"\t"` | 80.807% |
+| 2 | 6 | `" "` | 14.948% |
+| 3 | 5 | `"\n"` | 0.437% |
+| 4 | 251 | `"]L"` | 0.118% |
+| 5 | 26 | `":"` | 0.116% |
+
+### Sample 5
+
+- Sample: `polish_00017`
+- Row/token: `10` / `1020`
+- Line/field: `173:0`
+- Component bucket: `chord_structure_only`; flags: `chord_structure`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: ` ` -> `:`
+- Gold note -> top-1 note: `4FFn` -> `4FFn:4C#`
+- Gold token likelihood: rank `2`, p `1.519%`
+- Predicted token likelihood: p `93.987%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 26 | `":"` | 93.987% |
+| 2 | 6 | `" "` | 1.519% |
+| 3 | 47 | `"\`"` | 1.221% |
+| 4 | 55 | `"i"` | 0.525% |
+| 5 | 4 | `"\t"` | 0.155% |
+
+### Sample 6
+
+- Sample: `polish_00049`
+- Row/token: `33` / `898`
+- Line/field: `122:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `[L` -> `[LL`
+- Gold note -> top-1 note: `16cc[L` -> `16cc[LL`
+- Gold token likelihood: rank `12`, p `0.095%`
+- Predicted token likelihood: p `49.342%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 469 | `"[LL"` | 49.342% |
+| 2 | 410 | `"]LL"` | 31.857% |
+| 3 | 1330 | `"_LL"` | 13.280% |
+| 4 | 83 | `"LL"` | 0.469% |
+| 5 | 822 | `"nLL"` | 0.332% |
+
+### Sample 7
+
+- Sample: `polish_00095`
+- Row/token: `63` / `46`
+- Line/field: `7:1`
+- Component bucket: `duration_only`; flags: `duration`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `12g-` -> `8g-`
+- Gold note -> top-1 note: `12g-` -> `8g-`
+- Gold token likelihood: rank `2`, p `5.814%`
+- Predicted token likelihood: p `85.437%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 635 | `"8g-"` | 85.437% |
+| 2 | 2530 | `"12g-"` | 5.814% |
+| 3 | 832 | `"16g-"` | 3.996% |
+| 4 | 2302 | `"32g-"` | 0.349% |
+| 5 | 292 | `"8.g"` | 0.318% |
+
+### Sample 8
+
+- Sample: `polish_00020`
+- Row/token: `12` / `978`
+- Line/field: `169:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `(L` -> `(LL`
+- Gold note -> top-1 note: `16cc#(L` -> `16cc#(LL`
+- Gold token likelihood: rank `6`, p `0.131%`
+- Predicted token likelihood: p `95.713%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 279 | `"(LL"` | 95.713% |
+| 2 | 1633 | `"(LLL"` | 0.379% |
+| 3 | 269 | `"#LL"` | 0.345% |
+| 4 | 9 | `"("` | 0.314% |
+| 5 | 469 | `"[LL"` | 0.191% |
+
+### Sample 9
+
+- Sample: `polish_00039`
+- Row/token: `26` / `24`
+- Line/field: `4:0`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `16A-L` -> `16A-LL`
+- Gold note -> top-1 note: `16A-L` -> `16A-LL`
+- Gold token likelihood: rank `16`, p `0.037%`
+- Predicted token likelihood: p `97.914%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1109 | `"16A-LL"` | 97.914% |
+| 2 | 1702 | `"16A#LL"` | 0.189% |
+| 3 | 591 | `"16ALL"` | 0.126% |
+| 4 | 748 | `"16B-LL"` | 0.122% |
+| 5 | 598 | `"16A-"` | 0.111% |
+
+### Sample 10
+
+- Sample: `polish_00038`
+- Row/token: `25` / `286`
+- Line/field: `35:1`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `(L` -> `(LL`
+- Gold note -> top-1 note: `16g(L` -> `16g(LL`
+- Gold token likelihood: rank `12`, p `0.068%`
+- Predicted token likelihood: p `86.959%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 279 | `"(LL"` | 86.959% |
+| 2 | 9 | `"("` | 3.821% |
+| 3 | 1633 | `"(LLL"` | 2.467% |
+| 4 | 469 | `"[LL"` | 1.165% |
+| 5 | 2909 | `"(["` | 0.664% |
+
+### Sample 11
+
+- Sample: `polish_00052`
+- Row/token: `35` / `492`
+- Line/field: `87:1`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `tie_slur`
+- Gold piece -> top-1 piece: `16g` -> `16g(LL`
+- Gold note -> top-1 note: `16g(L` -> `16g(LL(L`
+- Gold token likelihood: rank `18`, p `0.042%`
+- Predicted token likelihood: p `97.638%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1339 | `"16g(LL"` | 97.638% |
+| 2 | 2782 | `"16g("` | 0.201% |
+| 3 | 1061 | `"16a(LL"` | 0.130% |
+| 4 | 2310 | `"16e(LL"` | 0.104% |
+| 5 | 817 | `"8g(L"` | 0.101% |
+
+### Sample 12
+
+- Sample: `polish_00090`
+- Row/token: `59` / `194`
+- Line/field: `26:1`
+- Component bucket: `duration_only`; flags: `duration`
+- Legacy bucket: `duration`
+- Gold piece -> top-1 piece: `4` -> `32ddd`
+- Gold note -> top-1 note: `48dddJ` -> `32ddd8dddJ`
+- Gold token likelihood: rank `3`, p `4.405%`
+- Predicted token likelihood: p `45.897%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1081 | `"32ddd"` | 45.897% |
+| 2 | 2342 | `"64ddd"` | 19.133% |
+| 3 | 20 | `"4"` | 4.405% |
+| 4 | 2786 | `"64bb-"` | 1.223% |
+| 5 | 22 | `"6"` | 1.223% |
+
+### Sample 13
+
+- Sample: `polish_00027`
+- Row/token: `17` / `1678`
+- Line/field: `239:1`
+- Component bucket: `pitch_and_accidental`; flags: `pitch, accidental`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `16B-J` -> `16cJ`
+- Gold note -> top-1 note: `16B-J` -> `16cJ`
+- Gold token likelihood: rank `2`, p `3.409%`
+- Predicted token likelihood: p `93.600%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 1940 | `"16cJ"` | 93.600% |
+| 2 | 2889 | `"16B-J"` | 3.409% |
+| 3 | 2534 | `"16BJ"` | 0.383% |
+| 4 | 1130 | `"16dJ"` | 0.338% |
+| 5 | 2021 | `"16AJ"` | 0.338% |
+
+### Sample 14
+
+- Sample: `polish_00087`
+- Row/token: `57` / `935`
+- Line/field: `127:0`
+- Component bucket: `tie_slur_only`; flags: `tie_slur`
+- Legacy bucket: `accidental`
+- Gold piece -> top-1 piece: `)` -> `	`
+- Gold note -> top-1 note: `4a-)` -> `4a-`
+- Gold token likelihood: rank `2`, p `1.571%`
+- Predicted token likelihood: p `97.178%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 4 | `"\t"` | 97.178% |
+| 2 | 10 | `")"` | 1.571% |
+| 3 | 55 | `"i"` | 0.094% |
+| 4 | 26 | `":"` | 0.028% |
+| 5 | 9 | `"("` | 0.027% |
+
+### Sample 15
+
+- Sample: `polish_00089`
+- Row/token: `58` / `324`
+- Line/field: `54:0`
+- Component bucket: `barline`; flags: `none`
+- Legacy bucket: `barline`
+- Gold piece -> top-1 piece: `==` -> `=`
+- Gold note -> top-1 note: `None` -> `None`
+- Gold token likelihood: rank `2`, p `13.442%`
+- Predicted token likelihood: p `48.405%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 28 | `"="` | 48.405% |
+| 2 | 212 | `"=="` | 13.442% |
+| 3 | 229 | `"*M4/4"` | 2.907% |
+| 4 | 283 | `"*M3/4"` | 2.647% |
+| 5 | 302 | `"*k[]"` | 1.819% |
+
+### Sample 16
+
+- Sample: `polish_00013`
+- Row/token: `7` / `1325`
+- Line/field: `157:2`
+- Component bucket: `note_other`; flags: `none`
+- Legacy bucket: `note_other`
+- Gold piece -> top-1 piece: `L` -> `LL`
+- Gold note -> top-1 note: `16qqccL` -> `16qqccLL`
+- Gold token likelihood: rank `9`, p `0.189%`
+- Predicted token likelihood: p `89.165%`
+
+| rank | token id | token | probability |
+|---:|---:|---|---:|
+| 1 | 83 | `"LL"` | 89.165% |
+| 2 | 279 | `"(LL"` | 2.032% |
+| 3 | 80 | `"JJ"` | 0.454% |
+| 4 | 269 | `"#LL"` | 0.426% |
+| 5 | 469 | `"[LL"` | 0.388% |