| --- |
| license: apache-2.0 |
| language: |
| - ca |
| - es |
| - gl |
| - eu |
| base_model: |
| - nvidia/stt_es_conformer_transducer_large |
| tags: |
| - automatic-speech-recognition |
| - NeMo |
| --- |
| # NVIDIA Conformer-Transducer Large (LoS) |
|
|
| ## Table of Contents |
| <details> |
| <summary>Click to expand</summary> |
|
|
| - [Model Description](#model-description) |
| - [Intended Uses and Limitations](#intended-uses-and-limitations) |
| - [How to Get Started with the Model](#how-to-get-started-with-the-model) |
| - [Training Details](#training-details) |
| - [Citation](#citation) |
| - [Additional Information](#additional-information) |
|
|
| </details> |
|
|
| ## Summary |
|
|
| The "stt_los_conformer_transducer_large_punctuated" is an acoustic model based on ["NVIDIA/stt_es_conformer_transducer_large"](https://huggingface.co/nvidia/stt_es_conformer_transducer_large/), suitable for Multilingual Automatic Speech Recognition in the languages for Spain (LoS): Catalan, Spanish, Galician, and Euskera |
| |
| ## Model Description |
| |
| This model transcribes speech in Catalan, Spanish, Galician, and Euskera alphabet, including punctuation, and was fine-tuned on a multilingual LoS dataset comprising 2700 hours. It is a "large" variant of Conformer-Transducer, with around 120 million parameters. |
| See the [model architecture](#model-architecture) section and [NeMo documentation](https://docs.nvidia.com/deeplearning/nemo/user-guide/docs/en/main/asr/models.html#conformer-transducer) for complete architecture details. |
| |
| ## Intended Uses and Limitations |
| |
| This model can be used for Automatic Speech Recognition (ASR) in Catalan, Spanish, Galician, and Euskera. It is intended to transcribe audio files in those languages to text with punctuation. |
| |
| ### Installation |
| |
| To use this model, install [NVIDIA NeMo](https://github.com/NVIDIA/NeMo). We recommend you install it after you've installed the latest PyTorch version. |
| ``` |
| pip install nemo_toolkit['all'] |
| ``` |
| |
| ### For Inference |
| To transcribe audio using this model, you can follow this example: |
| |
| |
| ```python |
| import nemo.collections.asr as nemo_asr |
| |
| nemo_asr_model = nemo_asr.models.EncDecRNNTBPEModel.restore_from(model) |
| transcription = nemo_asr_model.transcribe([audio_path])[0].text |
| print(transcription) |
| ``` |
| |
| ## Training Details |
| |
| ### Training data |
| |
| The specific datasets used to create the model are: |
| In Catalan: |
| - ["3CatParla"](https://huggingface.co/datasets/projecte-aina/3catparla_asr) (To be published soon) |
| - [Parlament-Parla-v3](https://huggingface.co/datasets/projecte-aina/parlament_parla_v3) (Only the anonymized version of the dataset is public. We trained the model with the non-anonymized version) |
| - [Corts Valencianes](https://huggingface.co/datasets/projecte-aina/corts_valencianes_asr_a) (Only the anonymized version of the dataset is public. We trained the model with the non-anonymized version) |
| - [IB3](https://huggingface.co/datasets/projecte-aina/ib3_ca_asr) (To be published soon) |
| - [Common Voice ca 17 Benchmark](https://huggingface.co/datasets/projecte-aina/commonvoice_benchmark_catalan_accents) |
| |
| |
| In Spanish: |
| - [ciempiess light](https://huggingface.co/datasets/ciempiess/ciempiess_light) |
| - [ciempiess fem](https://huggingface.co/datasets/ciempiess/ciempiess_fem) |
| - [ciempiess complementary](https://huggingface.co/datasets/ciempiess/ciempiess_complementary) |
| - [ciempiess balance](https://huggingface.co/datasets/ciempiess/ciempiess_balance) |
| - [CHM150](https://huggingface.co/datasets/carlosdanielhernandezmena/chm150_asr) |
| - [Tedx spanish](https://huggingface.co/datasets/ciempiess/tedx_spanish) |
| - [librivox spanish](https://huggingface.co/datasets/ciempiess/librivox_spanish) |
| - [Wikipedia spanish](https://huggingface.co/datasets/ciempiess/wikipedia_spanish) |
| - [voxforge spanish](https://huggingface.co/datasets/ciempiess/voxforge_spanish) |
| - [Tele con ciencia](https://huggingface.co/datasets/ciempiess/tele_con_ciencia) |
| - [Argentinian Spanish Speech Dataset](https://openslr.org/61/) |
| - [Dimex100 light](https://huggingface.co/datasets/carlosdanielhernandezmena/dimex100_light) |
| - [Glissando Spanish](https://glissando.labfon.uned.es/es) |
| - [Herico](https://openslr.org/39/) |
| - [Latino40](https://catalog.ldc.upenn.edu/LDC95S28) |
| - [Common voice 17 es](https://huggingface.co/datasets/mozilla-foundation/common_voice_17_0) |
| |
| |
| In Galician: |
| - [fleurs-galician](https://huggingface.co/datasets/google/fleurs) |
| - [google_crowdsourced](http://openslr.magicdatatech.com/77/) |
| - [Nos_ParlaSpeech-GL (clean)](https://huggingface.co/datasets/proxectonos/Nos_Parlaspeech-GL) |
| - [Nos_ParlaSpeech-GL (other)](https://huggingface.co/datasets/proxectonos/Nos_Parlaspeech-GL) |
| - [Nos_TranscriSpeech-GL](https://huggingface.co/datasets/proxectonos/Nos_Transcrispeech-GL) |
| - [Nos_RG-Podcast-GL](https://huggingface.co/datasets/proxectonos/Nos_RG-Podcast-GL) |
| - [FalAI (validated split)](https://huggingface.co/datasets/GTM-UVigo/FalAI) |
| - [Common Voice 22.0](https://huggingface.co/datasets/fsicoli/common_voice_22_0) |
| |
| |
| In Euskera: |
| - [composite_corpus_eu_v2.1](https://huggingface.co/datasets/HiTZ/composite_corpus_eu_v2.1) |
| |
| ### Training procedure |
| |
| This model is the result of finetuning the base model ["Nvidia/stt_es_conformer_transducer_large"](https://huggingface.co/nvidia/stt_es_conformer_transducer_large) by following this [tutorial](https://github.com/NVIDIA/NeMo/blob/main/tutorials/asr/Transducers_with_HF_Datasets.ipynb). |
| |
| ## Citation |
| If this model contributes to your research, please cite the work: |
| ```bibtex |
| @misc{LoS-conformer-transducer-punctuated-BSC-2026, |
| title={Languages of Spain ASR Model: stt_los_conformer_transducer_large_punctuated.}, |
| author={Messaoudi, Abir; Solito, Sarah; Hernandez Mena; Carlos, España-Bonet, Cristina}, |
| organization={Barcelona Supercomputing Center}, |
| url={https://huggingface.co/projecte-aina/stt_los_conformer_transducer_large_punctuated}, |
| year={2025} |
| } |
| ``` |
| |
| ## Additional Information |
|
|
| ### Author |
|
|
| The fine-tuning process was performed during 2025 in the [Language Technologies Laboratory](https://huggingface.co/BSC-LT) of the [Barcelona Supercomputing Center](https://www.bsc.es/) |
|
|
| For the Catalan Valencian data, we had the collaboration of [CENID](https://cenid.es/) within the framework of the [ILENIA](https://proyectoilenia.es) project. |
|
|
| ### Contact |
| For further information, please send an email to <langtech@bsc.es>. |
|
|
| ### Copyright |
| Copyright(c) 2025 by Language Technologies Laboratory, Barcelona Supercomputing Center. |
|
|
| ### License |
|
|
| [Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0) |
|
|
| ### Funding |
| This work is funded by the Ministerio para la Transformación Digital y de la Función Pública - Funded by EU – NextGenerationEU within the framework of the project ILENIA with reference 2022/TL22/00215337. |
|
|
| The training of the model was possible thanks to the computing time provided by [Barcelona Supercomputing Center](https://www.bsc.es/) through MareNostrum 5. |
