{ "aggregate": { "cer_normalized_micro": 0.05237288372939743, "character_errors": 3740, "corpora": 5, "reference_characters": 71411, "reference_words": 17425, "utterances": 1899, "wer_normalized_micro": 0.12975609756097561, "word_errors": 2261 }, "by_corpus": [ { "cer_normalized": "0.034468", "cer_raw": "0.034468", "char_deletions": "304", "char_deletions_raw": "304", "char_insertions": "177", "char_insertions_raw": "177", "char_substitutions": "240", "char_substitutions_raw": "240", "eval_corpus": "ILRDF", "eval_language": "seediq", "eval_utterances": "547", "reference_characters": "20918", "reference_characters_raw": "20918", "reference_words": "5155", "reference_words_raw": "5155", "run_id": "xlsr53_cpt_to_language_train-seediq_eval-seediq-all-corpora_dataset_v1_split_v2_formosan_safe_v1_seed13", "wer_normalized": "0.102425", "wer_raw": "0.102425", "word_deletions": "72", "word_deletions_raw": "72", "word_insertions": "32", "word_insertions_raw": "32", "word_substitutions": "424", "word_substitutions_raw": "424" }, { "cer_normalized": "0.150482", "cer_raw": "0.150482", "char_deletions": "430", "char_deletions_raw": "430", "char_insertions": "150", "char_insertions_raw": "150", "char_substitutions": "232", "char_substitutions_raw": "232", "eval_corpus": "NTU", "eval_language": "seediq", "eval_utterances": "99", "reference_characters": "5396", "reference_characters_raw": "5396", "reference_words": "1460", "reference_words_raw": "1460", "run_id": "xlsr53_cpt_to_language_train-seediq_eval-seediq-all-corpora_dataset_v1_split_v2_formosan_safe_v1_seed13", "wer_normalized": "0.339726", "wer_raw": "0.339726", "word_deletions": "124", "word_deletions_raw": "124", "word_insertions": "29", "word_insertions_raw": "29", "word_substitutions": "343", "word_substitutions_raw": "343" }, { "cer_normalized": "0.100073", "cer_raw": "0.100073", "char_deletions": "145", "char_deletions_raw": "145", "char_insertions": "263", "char_insertions_raw": "263", "char_substitutions": "139", "char_substitutions_raw": "139", "eval_corpus": "ePark1", "eval_language": "seediq", "eval_utterances": "289", "reference_characters": "5466", "reference_characters_raw": "5466", "reference_words": "1243", "reference_words_raw": "1243", "run_id": "xlsr53_cpt_to_language_train-seediq_eval-seediq-all-corpora_dataset_v1_split_v2_formosan_safe_v1_seed13", "wer_normalized": "0.193886", "wer_raw": "0.193886", "word_deletions": "25", "word_deletions_raw": "25", "word_insertions": "51", "word_insertions_raw": "51", "word_substitutions": "165", "word_substitutions_raw": "165" }, { "cer_normalized": "0.031188", "cer_raw": "0.031188", "char_deletions": "84", "char_deletions_raw": "84", "char_insertions": "130", "char_insertions_raw": "130", "char_substitutions": "85", "char_substitutions_raw": "85", "eval_corpus": "ePark2", "eval_language": "seediq", "eval_utterances": "242", "reference_characters": "9587", "reference_characters_raw": "9587", "reference_words": "2275", "reference_words_raw": "2275", "run_id": "xlsr53_cpt_to_language_train-seediq_eval-seediq-all-corpora_dataset_v1_split_v2_formosan_safe_v1_seed13", "wer_normalized": "0.06989", "wer_raw": "0.06989", "word_deletions": "13", "word_deletions_raw": "13", "word_insertions": "26", "word_insertions_raw": "26", "word_substitutions": "120", "word_substitutions_raw": "120" }, { "cer_normalized": "0.0453", "cer_raw": "0.0453", "char_deletions": "454", "char_deletions_raw": "454", "char_insertions": "443", "char_insertions_raw": "443", "char_substitutions": "464", "char_substitutions_raw": "464", "eval_corpus": "ePark3", "eval_language": "seediq", "eval_utterances": "722", "reference_characters": "30044", "reference_characters_raw": "30044", "reference_words": "7292", "reference_words_raw": "7292", "run_id": "xlsr53_cpt_to_language_train-seediq_eval-seediq-all-corpora_dataset_v1_split_v2_formosan_safe_v1_seed13", "wer_normalized": "0.114783", "wer_raw": "0.114783", "word_deletions": "105", "word_deletions_raw": "105", "word_insertions": "80", "word_insertions_raw": "80", "word_substitutions": "652", "word_substitutions_raw": "652" } ], "condition": "cpt_to_language", "dataset_version": "dataset_v1", "iso_639_3": "trv", "language": "Seediq", "language_id": "seediq", "model_family": "xlsr53", "normalization_version": "formosan_safe_v1", "readiness": { "actual_train_rows": 15137, "best_selection_cer": 0.06875724697793865, "best_selection_wer": 0.16030239596992973, "best_step": 56764, "condition": "cpt_to_language", "dominant_hypothesis_rate": 0.003686150605581885, "estimated_effective_epochs": 30.00013212657726, "failures": 0, "generated_at": "2026-06-29T12:22:12+00:00", "intended_train_rows": 15137, "manifest_dir": "/projects/prudlab/tacl_formosan_asr/experiments/manifests/gate6_xlsr_cpt_downstream_matrix/xlsr53_cpt_to_language_train-seediq_eval-seediq-all-corpora_dataset_v1_split_v2_formosan_safe_v1_seed13", "nonempty_hypothesis_rate": 1.0, "prediction_rows": 1899, "result_dir": "/scratch/scheppat/projects/tacl_formosan_asr/runs_xlsr_cpt_downstream_v1/xlsr53_cpt_to_language_train-seediq_eval-seediq-all-corpora_dataset_v1_split_v2_formosan_safe_v1_seed13", "result_label": "final", "run_id": "xlsr53_cpt_to_language_train-seediq_eval-seediq-all-corpora_dataset_v1_split_v2_formosan_safe_v1_seed13", "status": "review", "steps": 56764, "test_rows": 1899, "train_batch_size": 8, "warnings": 1 }, "seed": 13, "split_version": "split_v2" }