2025-label-dependent-ILM-ctc

Author: Zijian Yang

2025-label-dependent-ILM-ctc/README.md

@@ -0,0 +1,10 @@
+This folder contains information related to the Interspeech 2025 publication:
+
+"Label-Context-Dependent Internal Language Model Estimation for CTC".
+
+We use [RETURNN](https://github.com/rwth-i6/returnn) to conduct our experiments and our setups are based on [Sisyphus](https://github.com/rwth-i6/sisyphus)
+
+Experiments for WER comparisons on in-domain and cross-domain tasks (Table 1 and Table 2 in the paper) can be found [here](https://github.com/rwth-i6/i6_experiments/blob/phan_ctc_10kbpe/users/phan/configs/interspeech2025/table12.py)
+Experiments for different context lengths can be found [here](https://github.com/rwth-i6/i6_experiments/blob/phan_ctc_10kbpe/users/phan/configs/interspeech2025/table3.py)
+
+For questions on how to setup Sisyphus and run experiments, please open an issue here or write to zyang@cs.rwth-aachen.de or minh.nghia.phan@rwth-aachen.de