--- language: - eu - es dataset_info: features: - name: audio dtype: audio - name: sentence dtype: string - name: duration dtype: float64 - name: lang dtype: string configs: - config_name: default data_files: - split: train path: data/train* - split: dev path: data/dev* - split: dev_cv_es path: data/dev_cv_es* - split: dev_cv_eu path: data/dev_cv_eu* - split: dev_mls_es path: data/dev_mls_es* - split: dev_oslr_es path: data/dev_oslr_es* - split: dev_oslr_eu path: data/dev_oslr_eu* - split: dev_parl_es path: data/dev_parl_es* - split: dev_parl_eu path: data/dev_parl_eu* - split: dev_parl_bi path: data/dev_parl_bi* - split: dev_vp_es path: data/dev_vp_es* - split: test_cv_es path: data/test_cv_es* - split: test_cv_eu path: data/test_cv_eu* - split: test_mls_es path: data/test_mls_es* - split: test_oslr_es path: data/test_oslr_es* - split: test_oslr_eu path: data/test_oslr_eu* - split: test_parl_es path: data/test_parl_es* - split: test_parl_eu path: data/test_parl_eu* - split: test_parl_bi path: data/test_parl_bi* - split: test_vp_es path: data/test_vp_es* license: cc-by-4.0 task_categories: - automatic-speech-recognition tags: - asr - stt - dataset pretty_name: Composite dataset for bilingual Basque/Spanish v1.0 --- # Composite bilingual dataset for Spanish and Basque made from public available data This dataset is composed of the following public available data: ## Train split: The train split is composed of the following datasets combined: - **mozilla-foundation/common_voice_18_0/es**: a portion of the "validated" split removing "test_cv" and "dev_cv" split's sentences. (validated split contains official train + dev + test splits and more unique data) - **mozilla-foundation/common_voice_18_0/eu**: "validated" split removing "test_cv" and "dev_cv" split's sentences. (validated split contains official train + dev + test splits and more unique data) - **openslr/es**: a train split made from a portion of the SLR(39,61,67,71,72,73,74,75,108) subsets, this split has been cleaned from acronyms, numbers and sentences that are repeated in the following "test_oslr" and "dev_oslr" splits. - **openslr/eu**: a train split made from the SLR76 (Basque recordings) subset, this split has been cleaned from acronyms, numbers and sentences that are repeated in the following "test_oslr" and "dev_oslr" splits. - **mls/es**: a portion of the "train" split from the spanish dataset of Multilingual Librispeech. - **facebook/voxpopuli/es**: a portion of the "train" split from the spanish voxpopuli dataset, cleaned from acronyms and numeric characters. - **gttsehu/basque_parliament_1/es**: a portion of the official "train_clean" split for spanish. - **gttsehu/basque_parliament_1/eu**: the total official "train_clean" split for basque. - **gttsehu/basque_parliament_1/bi**: the total official "train_clean" split for bilingual basque spanish. The Total hours and sentences is slightly smaller because some of the sentences were removed due to be repeated in some of the test and dev splits. The Train split is balanced with the same amount of hours for Spanish and Basque. | Split tag | Source | Hours | Sentences | |:---------:|:-----------------------:|:-------------:|:-----------:| | - | common_voice_18_0_es | 139.98 h | 98424 | | - | common_voice_18_0_eu | 300.05 h | 198498 | | - | openslr_es | 11.87 h | 6355 | | - | openslr_eu | 6.28 h | 3229 | | - | mls_es | 240.07 h | 57682 | | - | voxpopuli_es | 36.62 h | 12424 | | - | basque_parliament_1_es | 247.27 h | 122183 | | - | basque_parliament_1_eu | 369.65 h | 185699 | | - | basque_parliament_1_bi | 14.42 h | 6984 | | train | **Total** | **1366.18 h** | **691471** | ## Test splits: Those test splits are separated, and it is recommended to not evaluate them together in a single split: - **mozilla-foundation/common_voice_18_0/es**: the official "test" split for spanish. - **mozilla-foundation/common_voice_18_0/eu**: the official "test" split for basque. - **openslr/es**: a test split made from the SLR(39,61,67,71,72,73,74,75,108) subsets, this split has been cleaned from acronyms and numbers. - **openslr/eu**: a test split made from the SLR76 (Basque recordings) subset, this split has been cleaned from acronyms and numbers. - **mls/es**: the "test" split from the spanish dataset of Multilingual Librispeech. - **facebook/voxpopuli/es**: the "test" split from the spanish voxpopuli dataset, cleaned from acronyms and numeric characters. - **gttsehu/basque_parliament_1/es**: the official "test" split for spanish. - **gttsehu/basque_parliament_1/eu**: the official "test" split for basque. - **gttsehu/basque_parliament_1/bi**: the official "test" split for bilingual basque spanish. | Split tag | Source | Hours | Sentences | |:------------:|:-----------------------:|:-----------:|:---------:| | test_cv_es | common_voice_18_0_es | 26.84 h | 15872 | | test_cv_eu | common_voice_18_0_eu | 24.2 h | 14312 | | test_oslr_es | openslr_es | 7.03 h | 4107 | | test_oslr_eu | openslr_eu | 1.03 h | 526 | | test_mls_es | mls_es | 10.0 h | 2385 | | test_vp_es | voxpopuli_es | 4.58 h | 1431 | | test_parl_es | basque_parliament_1_es | 6.56 h | 3450 | | test_parl_eu | basque_parliament_1_eu | 2.85 h | 1521 | | test_parl_bi | basque_parliament_1_bi | 0.36 h | 181 | | | **Total** | **83.46 h** | **43785** | ## Dev splits: There is a dev split composed by 9 dev subsplits that are also independently accesible. It is recommended to use the combined "dev" split for development tasks since it is accurately balanced in number of hours for spanish and basque (~3h spanish ~3h basque). - **mozilla-foundation/common_voice_18_0/es**: a small "dev" split made from the original dev split for spanish. - **mozilla-foundation/common_voice_18_0/eu**: a small "dev" split made from the original dev split for basque. - **openslr/es**: a dev split made from the SLR(39,61,67,71,72,73,74,75,108) subsets, this split has been cleaned from acronyms and numbers. - **openslr/eu**: a dev split made from the SLR76 (Basque recordings) subset, this split has been cleaned from acronyms and numbers. - **mls/es**: a small "dev" split from the original dev split of spanish dataset of Multilingual Librispeech. - **facebook/voxpopuli/es**: a small "dev" split from the spanish voxpopuli dev split dataset, cleaned from acronyms and numeric characters. - **gttsehu/basque_parliament_1/es**: a small dev split from the original dev split for spanish. - **gttsehu/basque_parliament_1/eu**: a small dev split from the original dev split for basque. - **gttsehu/basque_parliament_1/bi**: the official "dev" split for bilingual basque spanish. | Split tag | Source | Hours | Sentences | |:-----------:|:-----------------------:|:----------:|:---------:| | dev_cv_es | common_voice_18_0_es | 0.63 h | 375 | | dev_cv_eu | common_voice_18_0_eu | 1.02 h | 620 | | dev_oslr_es | openslr_es | 0.62 h | 379 | | dev_oslr_eu | openslr_eu | 1.01 h | 521 | | dev_mls_es | mls_es | 0.64 h | 152 | | dev_vp_es | voxpopuli_es | 0.59 h | 193 | | dev_parl_es | basque_parliament_1_es | 0.61 h | 320 | | dev_parl_eu | basque_parliament_1_eu | 1.01 h | 550 | | dev_parl_bi | basque_parliament_1_bi | 0.26 h | 131 | | dev | **Total** | **6.38 h** | **3241** | ## Funding: This project with reference 2022/TL22/00215335 has been parcially funded by the Ministerio de Transformación Digital and by the Plan de Recuperación, Transformación y Resiliencia – Funded by the European Union – NextGenerationEU [ILENIA](https://proyectoilenia.es/) and by the project [IkerGaitu](https://www.hitz.eus/iker-gaitu/) funded by the Basque Government.