Conformer2
Speech-To-Text

Introducing Conformer-2: Superior Speech Recognition with Enhanced Accuracy and Speed
Average rated: 0.00/5 with 0 ratings
Favorited 2 times
Rate this tool
About Conformer2
We are introducing Conformer-2, our latest AI model for automatic speech recognition. Conformer-2 is trained on 1.1M hours of English audio data, extending Conformer-1 to provide improvements on proper nouns, alphanumerics, and robustness to noise. Conformer-2 builds on our original release of Conformer-1, improving both model performance and speed. This model update achieves a 31.7% improvement on alphanumerics, a 6.8% improvement on Proper Noun Error Rate, and a 12.0% improvement in robustness to noise. These improvements were made by increasing the amount of training data to 1.1M hours and increasing the number of models used to pseudo-label data. Our Conformer-1 model achieved state-of-the-art performance and demonstrated strong noise robustness, making it well-suited for real-world audio conditions. Conformer-2 maintains parity with Conformer-1 in terms of word error rate while taking a step forward in user-oriented metrics. Since the release of Conformer-1, our engineering team decreased the latency of our inference pipeline by up to 53.7%.
Key Features
- 31.7% improvement on alphanumerics
- 6.8% improvement on Proper Noun Error Rate
- 12.0% boost in noise robustness
- Trained on 1.1M hours of English audio
- Maintains word error rate parity with Conformer-1
- Up to 53.7% reduction in latency
- Enhanced performance in real-world audio conditions
- Improved transcription accuracy
- Increased number of models used for pseudo-labeling data
- Developed by AssemblyAI