--- tags: - espnet - audio - automatic-speech-recognition language: mr datasets: - marathi_lrec2020 license: cc-by-4.0 --- ## ESPnet2 ASR model ### `espnet/marathi_lrec2020` This model was trained by [Aniket Tathe](https://github.com/Aniket-Tathe) using marathi_lrec2020 recipe in [espnet](https://github.com/espnet/espnet/). ### Demo: How to use in ESPnet2 Follow the [ESPnet installation instructions](https://espnet.github.io/espnet/installation.html) if you haven't done that already. ```bash cd espnet git checkout 6241a3e3ad9fef6a686ac82c6a7799d40d96cd27 pip install -e . cd egs2/marathi_lrec2020/asr1 ./run.sh --skip_data_prep false --skip_train true --download_model espnet/marathi_lrec2020 ``` # RESULTS ## Environment - **Date:** `Sat Oct 25 08:19:46 UTC 2025` - **Python version:** `3.9.23` - **ESPnet version:** `202509` - **PyTorch version:** `2.3.0+cu121` - **CUDA version:** `12.1` - **ESPnet Git hash (upstream):** `53e09761cb164b28f299e178262bf2056d8059d7` - **Commit date:** `Fri Oct 24 11:26:46 2025 +0900` --- ## Marathi ASR — `marathi_lrec2020` Recipe for **Marathi** ASR on the [**IndicCorpora Marathi subset**](https://www.cse.iitb.ac.in/~pjyothi/indiccorpora/#marathi). Training uses **`conf/train_asr_transformer.yaml`** (character Conformer: 3 blocks, 256-dim encoder, `batch_bins: 16000000`, `accum_grad: 4`, Adam `lr: 0.0005`, warmup 20k, SpecAugment, hybrid CTC/attention `ctc_weight: 0.3`). Decoding without LM: **`conf/decode_asr.yaml`** (`lm_weight: 0.0`). Decoding with LM (match reported fusion): **`conf/decode_asr_lm.yaml`** (beam 20, `ctc_weight: 0.5`, `lm_weight: 0.3`). --- ### Test-set decoding (`marathi_test`) Beam **20**, **CTC weight 0.5** unless noted. #### Beam 20, CTC 0.5 (no LM) | | Corr | Sub | Del | Ins | Err | S.Err | |--------|-----:|----:|----:|----:|----:|------:| | **CER** | 88.9 | 7.1 | 4.0 | 1.9 | 13.0 | 77.7 | | **WER** | 73.8 | 23.8 | 2.4 | 3.2 | 29.4 | 78.5 | #### Beam 20, CTC 0.5, LM weight 0.3 | | Corr | Sub | Del | Ins | Err | S.Err | |--------|-----:|----:|----:|----:|----:|------:| | **CER** | 89.0 | 6.6 | 4.4 | 1.7 | 12.6 | 74.3 | | **WER** | 76.0 | 21.6 | 2.4 | 3.0 | 27.0 | 75.0 | --- ### Dataset reference > P. Jyothi et al., *“IndicCorpora: A Large Multilingual Corpus for Indic Languages.”* > [IIT Bombay IndicCorpora — Marathi](https://www.cse.iitb.ac.in/~pjyothi/indiccorpora/#marathi) ## ASR config
expand ``` config: conf/tuning/train_asr_transformer.yaml print_config: false log_level: INFO drop_last_iter: false dry_run: false iterator_type: sequence valid_iterator_type: null output_dir: exp/asr_marathi_conf_old_try_3conf_16Mbin_4grad ngpu: 1 seed: 777 num_workers: 8 num_att_plot: 3 dist_backend: nccl dist_init_method: env:// dist_world_size: null dist_rank: null local_rank: 0 dist_master_addr: null dist_master_port: null dist_launcher: null multiprocessing_distributed: false unused_parameters: false sharded_ddp: false use_deepspeed: false deepspeed_config: null gradient_as_bucket_view: true ddp_comm_hook: null cudnn_enabled: true cudnn_benchmark: false cudnn_deterministic: true use_tf32: false collect_stats: false write_collected_feats: false max_epoch: 60 patience: null val_scheduler_criterion: - valid - loss early_stopping_criterion: - valid - loss - min best_model_criterion: - - valid - loss - min keep_nbest_models: 10 nbest_averaging_interval: 0 grad_clip: 5.0 grad_clip_type: 2.0 grad_noise: false accum_grad: 4 no_forward_run: false resume: false train_dtype: float32 use_amp: false log_interval: null use_matplotlib: true use_tensorboard: true create_graph_in_tensorboard: false use_wandb: false wandb_project: null wandb_id: null wandb_entity: null wandb_name: null wandb_model_log_interval: -1 detect_anomaly: false use_adapter: false adapter: lora save_strategy: all adapter_conf: {} pretrain_path: null init_param: [] ignore_init_mismatch: false freeze_param: [] num_iters_per_epoch: null batch_size: 20 valid_batch_size: null batch_bins: 16000000 valid_batch_bins: null category_sample_size: 10 upsampling_factor: 0.5 category_upsampling_factor: 0.5 dataset_upsampling_factor: 0.5 dataset_scaling_factor: 1.2 max_batch_size: null min_batch_size: 1 train_shape_file: - exp/asr_stats_raw_char/train/speech_shape - exp/asr_stats_raw_char/train/text_shape.char valid_shape_file: - exp/asr_stats_raw_char/valid/speech_shape - exp/asr_stats_raw_char/valid/text_shape.char batch_type: numel valid_batch_type: null fold_length: - 80000 - 150 sort_in_batch: descending shuffle_within_batch: false sort_batch: descending multiple_iterator: false chunk_length: 500 chunk_shift_ratio: 0.5 num_cache_chunks: 1024 chunk_excluded_key_prefixes: [] chunk_default_fs: null chunk_max_abs_length: null chunk_discard_short_samples: true train_data_path_and_name_and_type: - - dump/raw/marathi_train_sp/wav.scp - speech - sound - - dump/raw/marathi_train_sp/text - text - text valid_data_path_and_name_and_type: - - dump/raw/marathi_dev/wav.scp - speech - sound - - dump/raw/marathi_dev/text - text - text multi_task_dataset: false allow_variable_data_keys: false max_cache_size: 0.0 max_cache_fd: 32 allow_multi_rates: false valid_max_cache_size: null exclude_weight_decay: false exclude_weight_decay_conf: {} optim: adam optim_conf: lr: 0.0005 scheduler: warmuplr scheduler_conf: warmup_steps: 20000 token_list: - - - - ा - े - र - ् - क - स - ल - म - ि - ं - ी - य - त - न - व - ग - ह - ट - च - प - ड - आ - . - ज - श - ो - द - ब - अ - ू - ु - '?' - ण - इ - ध - ए - फ - ख - ॉ - ॅ - ळ - ँ - भ - थ - ठ - ई - झ - ष - उ - ऑ - ऊ - घ - ढ - ै - ओ - '2' - ( - ) - '0' - ृ - ौ - '-' - '3' - '1' - ऐ - ‍ - '5' - '8' - छ - '4' - '"' - ',' - '9' - औ - '!' - ़ - ञ - '7' - '6' - ९ - ऋ - e - g - ३ - X - १ - ० - init: xavier_uniform input_size: null ctc_conf: dropout_rate: 0.0 ctc_type: builtin reduce: true ignore_nan_grad: null zero_infinity: true brctc_risk_strategy: exp brctc_group_strategy: end brctc_risk_factor: 0.0 joint_net_conf: null use_preprocessor: true use_lang_prompt: false use_nlp_prompt: false token_type: char bpemodel: null non_linguistic_symbols: null cleaner: null g2p: null speech_volume_normalize: null rir_scp: null rir_apply_prob: 1.0 noise_scp: null noise_apply_prob: 1.0 noise_db_range: '13_15' short_noise_thres: 0.5 aux_ctc_tasks: [] frontend: default frontend_conf: fs: 16k specaug: specaug specaug_conf: apply_time_warp: true time_warp_window: 5 time_warp_mode: bicubic apply_freq_mask: true freq_mask_width_range: - 0 - 30 num_freq_mask: 2 apply_time_mask: true time_mask_width_range: - 0 - 40 num_time_mask: 2 normalize: global_mvn normalize_conf: stats_file: exp/asr_stats_raw_char/train/feats_stats.npz model: espnet model_conf: ctc_weight: 0.3 report_cer: true report_wer: true preencoder: null preencoder_conf: {} encoder: conformer encoder_conf: output_size: 256 attention_heads: 4 linear_units: 1024 num_blocks: 3 dropout_rate: 0.2 positional_dropout_rate: 0.2 attention_dropout_rate: 0.2 input_layer: conv2d normalize_before: true macaron_style: false pos_enc_layer_type: rel_pos selfattention_layer_type: rel_selfattn activation_type: swish use_cnn_module: true cnn_module_kernel: 17 postencoder: null postencoder_conf: {} decoder: transformer decoder_conf: attention_heads: 4 linear_units: 1024 num_blocks: 3 dropout_rate: 0.2 positional_dropout_rate: 0.2 self_attention_dropout_rate: 0.2 src_attention_dropout_rate: 0.2 preprocessor: default preprocessor_conf: {} required: - output_dir - token_list version: '202509' distributed: false ```
### Citing ESPnet ```BibTex @inproceedings{watanabe2018espnet, author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai}, title={{ESPnet}: End-to-End Speech Processing Toolkit}, year={2018}, booktitle={Proceedings of Interspeech}, pages={2207--2211}, doi={10.21437/Interspeech.2018-1456}, url={http://dx.doi.org/10.21437/Interspeech.2018-1456} } ``` or arXiv: ```bibtex @misc{watanabe2018espnet, title={ESPnet: End-to-End Speech Processing Toolkit}, author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai}, year={2018}, eprint={1804.00015}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```