Add model card, acoustic model checkpoint in safetensors format, ONNX model, tokenizer configs and KenLM model

initial commit