{
  "architectures": [
    "FishOmniModel"
  ],
  "audio_decoder_config": {
    "_name_or_path": "",
    "architectures": null,
    "attention_o_bias": false,
    "attention_output_gate": true,
    "attention_qk_norm": false,
    "attention_qkv_bias": false,
    "audio_embed_dim": null,
    "audio_hidden_dim": 4096,
    "chunk_size_feed_forward": 0,
    "codebook_loss_weight_per_layer": null,
    "decoder_type": "codebook_mtp_shared",
    "dim": 2048,
    "dropout": 0.0,
    "dtype": null,
    "head_dim": 256,
    "id2label": {
      "0": "LABEL_0",
      "1": "LABEL_1"
    },
    "initializer_range": 0.02209708691207961,
    "intermediate_size": 6144,
    "is_encoder_decoder": false,
    "label2id": {
      "LABEL_0": 0,
      "LABEL_1": 1
    },
    "layernorm_zero_centered_gamma": true,
    "max_seq_len": 11,
    "model_type": "fish_qwen3_audio_decoder",
    "moe_intermediate_size": 768,
    "mtp_spec": {
      "attention_variant": "full_attention"
    },
    "mtp_use_swa": false,
    "n_head": 16,
    "n_layer": 4,
    "n_local_heads": 2,
    "norm_eps": 1e-06,
    "norm_topk_prob": true,
    "num_codebooks": 10,
    "num_experts": 1,
    "num_experts_per_tok": 1,
    "output_attentions": false,
    "output_hidden_states": false,
    "problem_type": null,
    "return_dict": true,
    "rope_base": 10000000.0,
    "rotary_interleaved": true,
    "rotary_percent": 0.25,
    "router_gamma": 0.001,
    "shared_enrichment_layers": 4,
    "text_dim": 2048,
    "tie_word_embeddings": true,
    "use_aux_loss_free": false,
    "use_gradient_checkpointing": true,
    "use_moe": false,
    "vocab_size": 4096,
    "vq_backbone_codebooks": 1,
    "vq_feature_placement": "backbone_partial",
    "vq_feature_scope": "all_vq",
    "window_attn_skip_freq": null,
    "window_rotary_base": null,
    "window_size": null
  },
  "audio_encoder_config": null,
  "audio_pad_token_id": 248076,
  "dtype": "bfloat16",
  "eos_token_id": 248046,
  "fish_schema_version": 1,
  "model_type": "fish_omni",
  "pad_token_id": 248044,
  "semantic_end_token_id": 252194,
  "semantic_start_token_id": 248099,
  "speech_template_version": "reference-bracket-v1",
  "text_config": {
    "_name_or_path": "",
    "architectures": null,
    "attention_bias": false,
    "attention_dropout": 0.0,
    "attn_output_gate": true,
    "bos_token_id": null,
    "chunk_size_feed_forward": 0,
    "dtype": null,
    "eos_token_id": 248046,
    "head_dim": 256,
    "hidden_act": "silu",
    "hidden_size": 2048,
    "id2label": {
      "0": "LABEL_0",
      "1": "LABEL_1"
    },
    "initializer_range": 0.02,
    "intermediate_size": 512,
    "is_encoder_decoder": false,
    "label2id": {
      "LABEL_0": 0,
      "LABEL_1": 1
    },
    "layer_types": [
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention"
    ],
    "layernorm_zero_centered_gamma": true,
    "linear_conv_kernel_dim": 4,
    "linear_key_head_dim": 128,
    "linear_num_key_heads": 16,
    "linear_num_value_heads": 32,
    "linear_value_head_dim": 128,
    "mamba_ssm_dtype": "float32",
    "max_position_embeddings": 262144,
    "mlp_layer_types": [
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe",
      "moe"
    ],
    "model_type": "fish_omni_text",
    "moe_intermediate_size": 512,
    "num_attention_heads": 16,
    "num_experts": 256,
    "num_experts_per_tok": 8,
    "num_hidden_layers": 40,
    "num_key_value_heads": 2,
    "output_attentions": false,
    "output_hidden_states": false,
    "pad_token_id": 248044,
    "problem_type": null,
    "return_dict": true,
    "rms_norm_eps": 1e-06,
    "rope_parameters": {
      "partial_rotary_factor": 0.25,
      "rope_theta": 10000000.0,
      "rope_type": "default"
    },
    "rotary_interleaved": false,
    "shared_expert_gate": true,
    "shared_expert_intermediate_size": 512,
    "tie_word_embeddings": false,
    "use_cache": true,
    "use_qk_norm": true,
    "vocab_size": 252288
  },
  "transformers_version": "5.14.1",
  "quantization_config": {
    "quant_method": "bsgemm-moe",
    "weight_block_size": [
      128,
      128
    ],
    "activation_scheme": "dynamic"
  }
}