Instructions to use AlexanderMaz/LanguageModel_Fusion with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use AlexanderMaz/LanguageModel_Fusion with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("AlexanderMaz/LanguageModel_Fusion") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
Commit History
-
Upload 2 files
Verified signature4650342 -
Delete optimize
Verified signature9be2cd2
-
Upload 2 files
Verified signatureChange optimization strategy.2217812 -
Delete optimize
Verified signature87031af
-
Update README.md
773f9bf -
Upload optimization
Add optimization results. data_all.pkl - contains validation_clear and validation_other decoding results; exp.db - database for optimization experiments0f3f3d4
-
Update README.md
b9f93e0 -
Update README
Add data description.b717995
-
Upload data
1. Trained word-level LSTM. 2. Trained token-level 4-gram arpa model. 3. Fine-tuned gpt2 model. 4. Baseline ASR model (https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/stt_en_conformer_transducer_small). 5. LibriSpeech text data for train, validation and test.36c6e03 -
initial commit
eb403da