RESULTS.md

RESULTS

Environments

• date: Fri Jan 17 07:59:11 CST 2025
• python version: 3.10.14 | packaged by conda-forge | (main, Mar 20 2024, 12:45:18) [GCC 12.3.0]
• espnet version: espnet 202409
• pytorch version: pytorch 2.4.0+cu121
• Git hash: e0dd1cf248a6b1b652f9c3e4c05d7516de37d965
  • Commit date: Tue Jul 23 07:19:00 2024 -0400

exp_owsm/s2t_train_05b_ds_raw_bpe50000

WER

dataset	Snt	Wrd	Corr	Sub	Del	Ins	Err	S.Err
decode_s2t_s2t_model_valid.total_count.ave_5best/eval1	250	250	28.8	71.2	0.0	0.0	71.2	71.2
decode_s2t_s2t_model_valid.total_count.ave_5best/test_zh_TW	4825	4825	6.0	94.0	0.0	0.0	94.0	94.0

CER

dataset	Snt	Wrd	Corr	Sub	Del	Ins	Err	S.Err
decode_s2t_s2t_model_valid.total_count.ave_5best/eval1	250	5372	86.6	11.9	1.5	3.3	16.7	71.2
decode_s2t_s2t_model_valid.total_count.ave_5best/test_zh_TW	4825	38305	60.6	39.1	0.3	0.3	39.7	94.0

TER

dataset	Snt	Wrd	Corr	Sub	Del	Ins	Err	S.Err
decode_s2t_s2t_model_valid.total_count.ave_5best/eval1	250	3372	79.9	17.3	2.8	5.6	25.7	71.2
decode_s2t_s2t_model_valid.total_count.ave_5best/test_zh_TW	4825	36530	49.9	37.7	12.5	1.1	51.2	94.0

exp_owsm/s2t_train_05b_ds_raw_bpe50000/decode_s2t_s2t_model_checkpoint_34_34_mp_rank_00_model_states

WER

dataset	Snt	Wrd	Corr	Sub	Del	Ins	Err	S.Err
ksponspeech/eval_clean_whisper	3000	23401	48.5	37.1	14.4	9.9	61.4	100.0

CER

dataset	Snt	Wrd	Corr	Sub	Del	Ins	Err	S.Err
ksponspeech/eval_clean_whisper	3000	134596	40.7	6.8	52.5	2.9	62.2	100.0

TER

dataset	Snt	Wrd	Corr	Sub	Del	Ins	Err	S.Err
ksponspeech/eval_clean_whisper	3000	59657	58.4	16.0	25.6	7.9	49.5	100.0