Open LLM List
morsecode_en_quartznet_10x5
1-800-BAD-CODEIntroduction
The Morse Code recognition model, developed by 1-800-BAD-CODE, is designed to transcribe Morse code audio signals into English text. This model uses the NeMo framework and is trained specifically for decoding Morse code rather than regular speech signals.
Architecture
This model is based on the QuartzNet architecture within the NeMo framework. It requires only the base NeMo package for inference, as it functions as an ASR (Automatic Speech Recognition) model tailored for Morse code signals.
Training
The model was trained using synthetic Morse code data generated with the MorseCodeToolkit. The data consists of audio signals sampled at 8 kHz, representing various Morse code messages. The performance of the model is optimized for Morse code generated within specific parameter ranges used during training.
Guide: Running Locally
- Install NeMo: Ensure that the NeMo package is installed in your Python environment.
- Load the Model:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("1-800-BAD-CODE/morsecode_en_quartznet_10x5") - Download an Example File:
- Use the command:
wget https://upload.wikimedia.org/wikipedia/commons/0/04/Wikipedia-Morse.ogg
- Use the command:
- Transcribe Audio:
- Execute the transcription:
asr_model.transcribe(['Wikipedia-Morse.ogg']) - Expected output:
['WELCOME TO WIKIPEDIA, THE FREE ENCYCLOPEDIA THAT ANYONE CAN EDIT.']
- Execute the transcription:
For enhanced performance, consider using cloud-based GPUs from services like AWS, Google Cloud, or Azure.
License
Details regarding the specific license for this model are not provided in the document. Please refer to the original repository or contact the author for licensing information.
