Audio-Text-to-Text
ONNX
Safetensors
English
Chinese
Model Optimizer
bailing_moe_v3_audio
audio
speech
automatic-speech-recognition
audio-understanding
audio-reasoning
speech-to-text
conversational
mixture-of-experts
quantized
nvfp4
fp4
4-bit precision
modelopt
blackwell
custom_code
Eval Results (legacy)
Eval Results
8-bit precision
| { | |
| "architectures": [ | |
| "BailingMoeV3AudioForConditionalGeneration" | |
| ], | |
| "audio_config": { | |
| "downsampler_kernel_size_list": [ | |
| 3, | |
| 3, | |
| 3 | |
| ], | |
| "downsampler_padding_list": [ | |
| 1, | |
| 1, | |
| 1 | |
| ], | |
| "downsampler_stride_list": [ | |
| 2, | |
| 2, | |
| 2 | |
| ], | |
| "model_type": "bailing_moe_v3_audio_encoder", | |
| "n_audio_ctx": 3000, | |
| "n_audio_head": 20, | |
| "n_audio_layer": 32, | |
| "n_audio_state": 1280, | |
| "n_mels": 128, | |
| "whisper_chunk_size": 3000 | |
| }, | |
| "audio_token_id": 157168, | |
| "auto_map": { | |
| "AutoConfig": "configuration_bailing_moe_v3_audio.BailingMoeV3AudioConfig", | |
| "AutoModel": "modeling_bailing_moe_v3_audio.BailingMoeV3AudioForConditionalGeneration", | |
| "AutoModelForCausalLM": "modeling_bailing_moe_v3_audio.BailingMoeV3AudioForConditionalGeneration", | |
| "AutoProcessor": "processing_bailing_moe_v3_audio.BailingMoeV3AudioProcessor" | |
| }, | |
| "dtype": "bfloat16", | |
| "model_type": "bailing_moe_v3_audio", | |
| "norm_query_embeds": true, | |
| "text_config": { | |
| "attention_dropout": 0.0, | |
| "embedding_dropout": 0.0, | |
| "eos_token_id": 156892, | |
| "expert_swiglu_limit_list": null, | |
| "first_k_dense_replace": 1, | |
| "gated_attention_proj_granularity_type": "head_wise", | |
| "group_norm_size": 1, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_size": 1536, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 4608, | |
| "kda_lower_bound": -5, | |
| "kda_safe_gate": true, | |
| "kv_lora_rank": 512, | |
| "layer_group_size": 4, | |
| "linear_silu": true, | |
| "max_position_embeddings": 131072, | |
| "max_window_layers": 20, | |
| "model_type": "", | |
| "moe_intermediate_size": 512, | |
| "moe_router_enable_expert_bias": true, | |
| "moe_shared_expert_intermediate_size": 512, | |
| "mtp_loss_scaling_factor": 0, | |
| "mtp_use_kda": false, | |
| "n_group": 8, | |
| "no_kda_lora": true, | |
| "norm_topk_prob": true, | |
| "num_attention_heads": 16, | |
| "num_experts": 128, | |
| "num_experts_per_tok": 8, | |
| "num_hidden_layers": 24, | |
| "num_key_value_heads": 16, | |
| "num_kv_heads_for_linear_attn": 0, | |
| "num_nextn_predict_layers": 1, | |
| "num_shared_experts": 1, | |
| "output_dropout": 0.0, | |
| "output_router_logits": false, | |
| "pad_token_id": 156892, | |
| "partial_rotary_factor": 0.5, | |
| "q_lora_rank": 256, | |
| "qk_head_dim": 192, | |
| "qk_nope_head_dim": 128, | |
| "qk_rope_head_dim": 64, | |
| "rms_norm_eps": 1e-06, | |
| "rope_interleave": true, | |
| "rope_parameters": { | |
| "partial_rotary_factor": 0.5, | |
| "rope_theta": 6000000, | |
| "rope_type": "default" | |
| }, | |
| "rope_theta": 6000000, | |
| "rotary_dim": 64, | |
| "routed_scaling_factor": 2.5, | |
| "router_dtype": "fp32", | |
| "scale_router_input": false, | |
| "score_function": "sigmoid", | |
| "scoring_func": "sigmoid", | |
| "seq_aux": true, | |
| "share_expert_swiglu_limit_list": null, | |
| "short_conv_kernel_size": 4, | |
| "tie_word_embeddings": false, | |
| "topk_group": 4, | |
| "topk_method": "noaux_tc", | |
| "up_proj_norm": false, | |
| "use_bias": false, | |
| "use_cache": true, | |
| "use_kda_lora": false, | |
| "use_mla_nope": false, | |
| "use_nGPT": false, | |
| "use_qk_norm": true, | |
| "use_qkv_bias": false, | |
| "v_head_dim": 128, | |
| "value_norm": false, | |
| "vocab_size": 157184 | |
| }, | |
| "tie_word_embeddings": false, | |
| "transformers_version": "5.14.1", | |
| "quantization_config": { | |
| "config_groups": { | |
| "group_0": { | |
| "input_activations": { | |
| "dynamic": false, | |
| "num_bits": 4, | |
| "type": "float", | |
| "group_size": 16 | |
| }, | |
| "weights": { | |
| "dynamic": false, | |
| "num_bits": 4, | |
| "type": "float", | |
| "group_size": 16 | |
| }, | |
| "targets": [ | |
| "Linear" | |
| ] | |
| } | |
| }, | |
| "ignore": [ | |
| "linear_proj_audio*", | |
| "lm_head", | |
| "model.audio*", | |
| "model.layers.0*", | |
| "model.layers.1.attention*", | |
| "model.layers.1.mlp.gate", | |
| "model.layers.1.mlp.shared_experts*", | |
| "model.layers.10*", | |
| "model.layers.11.attention*", | |
| "model.layers.11.mlp.gate", | |
| "model.layers.11.mlp.shared_experts*", | |
| "model.layers.12.attention*", | |
| "model.layers.12.mlp.gate", | |
| "model.layers.12.mlp.shared_experts*", | |
| "model.layers.13.attention*", | |
| "model.layers.13.mlp.gate", | |
| "model.layers.13.mlp.shared_experts*", | |
| "model.layers.14.attention*", | |
| "model.layers.14.mlp.gate", | |
| "model.layers.14.mlp.shared_experts*", | |
| "model.layers.15*", | |
| "model.layers.16.attention*", | |
| "model.layers.16.mlp.gate", | |
| "model.layers.16.mlp.shared_experts*", | |
| "model.layers.17.attention*", | |
| "model.layers.17.mlp.gate", | |
| "model.layers.17.mlp.shared_experts*", | |
| "model.layers.18*", | |
| "model.layers.19.attention*", | |
| "model.layers.19.mlp.gate", | |
| "model.layers.19.mlp.shared_experts*", | |
| "model.layers.2.attention*", | |
| "model.layers.2.mlp.gate", | |
| "model.layers.2.mlp.shared_experts*", | |
| "model.layers.20*", | |
| "model.layers.21*", | |
| "model.layers.22*", | |
| "model.layers.23.attention*", | |
| "model.layers.23.mlp.gate", | |
| "model.layers.23.mlp.shared_experts*", | |
| "model.layers.24*", | |
| "model.layers.3.attention*", | |
| "model.layers.3.mlp.gate", | |
| "model.layers.3.mlp.shared_experts*", | |
| "model.layers.4.attention*", | |
| "model.layers.4.mlp.gate", | |
| "model.layers.4.mlp.shared_experts*", | |
| "model.layers.5*", | |
| "model.layers.6.attention*", | |
| "model.layers.6.mlp.gate", | |
| "model.layers.6.mlp.shared_experts*", | |
| "model.layers.7.attention*", | |
| "model.layers.7.mlp.gate", | |
| "model.layers.7.mlp.shared_experts*", | |
| "model.layers.8.attention*", | |
| "model.layers.8.mlp.gate", | |
| "model.layers.8.mlp.shared_experts*", | |
| "model.layers.9.attention*", | |
| "model.layers.9.mlp.gate", | |
| "model.layers.9.mlp.shared_experts*", | |
| "model.word_embeddings", | |
| "model.layers.24.*" | |
| ], | |
| "quant_algo": "NVFP4", | |
| "producer": { | |
| "name": "modelopt", | |
| "version": "0.47.0" | |
| }, | |
| "quant_method": "modelopt" | |
| } | |
| } |