config.json

{
  "architectures": [
    "Echo"
  ],
  "base": 10000,
  "checkpointing": false,
  "cross_attention": false,
  "init_std": 0.02,
  "max_rel_dist": 1000,
  "model_type": "Echo",
  "n_audio_ctx": 1500,
  "n_audio_head": 16,
  "n_audio_layer": 24,
  "n_audio_state": 1024,
  "n_mels": 128,
  "n_text_ctx": 448,
  "n_text_head": 16,
  "n_text_layer": 16,
  "n_text_state": 1024,
  "torch_dtype": "float32",
  "transformers_version": "4.45.1",
  "unk_token_id": 113,
  "use_hybrid_attention": true,
  "vocab_size": 15141,
  "window_size": 100
}