arielcerdap/conformer-ctc-small-fluencybank

Conformer-CTC fine-tuned on FluencyBank Timestamped

Fine-tuned from nvidia/stt_en_conformer_ctc_small on the FluencyBank Timestamped dataset (arielcerdap/TimeStamped-Splits), targeting verbatim transcription of stuttered speech from adults who stutter (PWS).

Results on test set

Training details

• —Base model: nvidia/stt_en_conformer_ctc_small
• —Dataset: arielcerdap/TimeStamped-Splits (train=2,744 / val=342 / test=342)
• —Target: verbatim transcription (disfluencies preserved)
• —Optimizer: AdamW (lr=1e-5, weight_decay=1e-3)
• —Scheduler: CosineAnnealing (warmup_ratio=0.05)
• —Batch effective: 64
• —Early stopping: patience=10 on val_wer