Wangyou Zhang Universal Train Enh Uses Refch0 2mem Raw

--- tags: - espnet - audio - audio-to-audio language: en datasets: - vctk_noisy - dns_ins20 - chime4 - reverb - whamr license: cc-by-4.0 --- ## ESPnet2 ENH model ### `espnet/Wangyou_Zhang_universal_train_enh_uses_refch0_2mem_raw` This model was trained by Wangyou Zhang using the universal_se 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 pip install -e . cd egs2/universal_se/enh1 ./run.sh --skip_data_prep false --skip_train true --is_tse_task false --download_model espnet/Wangyou_Zhang_universal_train_enh_uses_refch0_2mem_raw ``` <!-- Generated by scripts/utils/show_enh_score.sh --> # RESULTS ## Environments - date: `Sat Jul 15 12:50:47 CST 2023` - python version: `3.8.16 (default, Mar 2 2023, 03:21:46) [GCC 11.2.0]` - espnet version: `espnet 202301` - pytorch version: `pytorch 2.0.1` - Git hash: `` - Commit date: `` ## USES (ref_channel=0, 2 groups of memory tokens) |dataset|condition|PESQ_WB|STOI|SAR|SDR|SIR|SI_SNR|OVRL|SIG|BAK|P808_MOS| |---|---|---|---|---|---|---|---|---|---|---|---| |vctk_noisy_tt_2spk|1ch, 48kHz||93.05|10.97|10.97|0.00|8.36|3.14|3.39|4.05|3.57| |vctk_noisy_tt_2spk_16k|1ch, 16kHz|3.11|95.03|21.51|21.51|0.00|19.45|3.19|3.46|4.06|3.57| |dns20_tt_synthetic_no_reverb|1ch, 16kHz|3.23|97.77|19.63|19.63|0.00|19.72|3.32|3.56|4.10|4.04| |dns20_tt_synthetic_with_reverb|1ch, 16kHz|2.75|89.87|13.40|13.40|0.00|12.90|2.36|2.85|3.21|3.37| |chime4_et05_simu_isolated_6ch_track|5ch, 16kHz|2.95|97.82|18.30|18.30|0.00|17.24|3.22|3.47|4.07|3.75| |reverb_et_simu_8ch_multich|8ch, 16kHz|2.09|89.83|11.94|11.94|0.00|-10.12|2.98|3.35|3.79|3.90| |whamr_tt_mix_single_anechoic_max_16k|2ch, 16kHz|2.55|96.36|15.78|15.78|0.00|15.46|3.33|3.55|4.16|3.86| |whamr_tt_mix_single_reverb_max_16k|2ch, 16kHz|2.51|95.98|13.75|13.75|0.00|12.51|3.32|3.54|4.15|3.86| |chime4_et05_real_isolated_6ch_track_1ch|5ch, 16kHz|1.23|55.11|-2.34|-2.34|0.00|-30.45|3.07|3.36|3.98|3.75| |reverb_et_real_8ch_multich|8ch, 16kHz|1.17|75.30|4.39|4.39|0.00|1.62|3.11|3.42|3.97|3.99| ## ENH config <details><summary>expand</summary> ``` config: conf/tuning/train_enh_uses_refch0_2mem.yaml print_config: false log_level: INFO dry_run: false iterator_type: chunk output_dir: exp/enh_train_enh_uses_refch0_2me...

Source context: 8 downloads · 2 likes · Pipeline audio-to-audio · Library espnet · Repo espnet/Wangyou_Zhang_universal_train_enh_uses_refch0_2mem_raw