This model was trained by Shunsuke Mitsumori using mlsuperb2 recipe in espnet.
Model source
Source excerpt
This model was trained by Shunsuke Mitsumori using mlsuperb2 recipe in espnet.
Sources
1 sourceVerified Sep 22
Model artifacts
1 artifacthome/data3/mitsumori/mlsuperb2/exp/asr_mms_baseline_refined_b8a4_i20k_gpu1/valid.loss.ave_2best.pth
pth · 3.61 GB · SHA-256 1d8922f08a42…5018 · Hugging Face
DownloadSource excerpts
2 excerpts--- tags: - espnet - audio - automatic-speech-recognition language: multilingual datasets: - ml_superb2 license: cc-by-4.0 --- ## ESPnet2 ASR model ### `shun3232/ml-superb2-mms-ctc-baseline-refined` This model was trained by Shunsuke Mitsumori using ml_superb2 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 e3b628d041733d5155f2d04d1acc76ff1faa8596 pip install -e . cd egs2/ml_superb2/asr1 ./run.sh --skip_data_prep false --skip_train true --download_model shun3232/ml-superb2-mms-ctc-baseline-refined ``` <!-- Generated by scripts/utils/show_asr_result.sh --> # RESULTS ## Environments - date: `Wed Jul 8 02:35:07 UTC 2026` - python version: `3.10.14 (tags/v3.10.14-25-ge98930d7387-dirty:e98930d7387, May 24 2024, 23:30:09) [GCC 13.2.0]` - espnet2 version: `espnet2 202604` - pytorch version: `pytorch 2.9.1+cu126` - Git hash: `e3b628d041733d5155f2d04d1acc76ff1faa8596` - Commit date: `Mon Jun 29 01:45:48 2026 +0900` ## exp/asr_mms_baseline_refined_b8a4_i20k_gpu1 ### WER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |decode_asr_asr_model_valid.loss.ave_2best/dev_dialect|7095|78862|30.3|63.3|6.4|4.9|74.6|99.8| ### CER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |decode_asr_asr_model_valid.loss.ave_2best/dev_dialect|7095|457230|77.5|9.9|12.6|3.6|26.1|99.8| ### TER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| ## exp/asr_mms_baseline_refined_b8a4_i20k_gpu1/decode_asr_asr_model_valid.loss.ave_2best ### WER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |org/dev|23609|234661|41.5|51.2|7.3|5.1|63.6|96.3| ### CER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |org/dev|23609|1495981|84.5|7.8|7.7|3.3|18.8|96.3| ### TER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| ## ASR config <details><summary>expand</summary> ``` config: conf/tuning/train_mms_baseline.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_mms_baseline_refined_b8a4_i20k_gpu1 ngp...
Source context: 3 downloads · 1 likes · Pipeline automatic-speech-recognition · Library espnet · Repo shun3232/ml-superb2-mms-ctc-baseline-refined