Introduction
1.
Installation
2.
Key Features
3.
Product overview
4.
Performance
5.
ML training flow
5.1.
Installation
5.2.
Model YAML configurations
5.3.
Training times
5.4.
Data preparation
5.4.1.
Supported dataset formats
5.4.1.1.
JSON format
5.4.1.2.
WebDataset format
5.4.1.3.
Hugging Face dataset format
5.4.1.4.
Directory of audio format
5.4.2.
Log mel feature normalization
5.5.
Training
5.5.1.
Batch size hyperparameters
5.5.2.
Heterogeneous CPUs
5.5.3.
Tensorboard
5.5.4.
Challenging target data
5.5.5.
Resuming and fine-tuning
5.5.6.
Profiling
5.5.7.
Changing the character set
5.5.8.
Delay penalty
5.5.9.
Endpointing
5.6.
Validation
5.6.1.
WER calculation
5.6.2.
State resets
5.6.3.
Beam search decoding
5.6.4.
N-gram language model
5.6.5.
Automatic batch size reduction
5.6.6.
Saving predictions
5.6.7.
User-perceived latency
5.6.8.
Emission latency
5.7.
Export inference checkpoint
6.
Inference flow
6.1.
Licensing
6.2.
Programming the FPGA
6.3.
CAIMAN-ASR server
6.3.1.
Compiling weights
6.3.2.
Websocket API
6.3.3.
Testing inference performance
6.3.4.
Live audio demonstration
6.3.5.
Hardware requirements
6.4.
CAIMAN-ASR demo
7.
Versions
Light (default)
Rust
Coal
Navy
Ayu
CAIMAN-ASR
Versions
Documentation for previous versions:
v1.9.0
v1.10.0
v1.10.1
v1.11.0
v1.12.0