{ "architectures": [ "StepAudio2ForCausalLM" ], "audio_encoder_config": { "adapter_stride": 2, "kernel_size": 3, "llm_dim": 5120, "model_type": "step_audio_2_encoder", "n_audio_ctx": 1500, "n_audio_head": 20, "n_audio_layer": 32, "n_audio_state": 1280, "n_codebook_size": 4096, "n_mels": 128 }, "auto_map": { "AutoConfig": "configuration_step_audio_2.StepAudio2Config", "AutoModelForCausalLM": "modeling_step_audio_2.StepAudio2ForCausalLM" }, "dtype": "bfloat16", "max_window_layers": null, "model_type": "step_audio_2", "quantization_config": { "config_groups": { "group_0": { "format": "nvfp4-pack-quantized", "input_activations": null, "output_activations": null, "targets": [ "Linear" ], "weights": { "actorder": null, "block_structure": null, "dynamic": false, "group_size": 16, "num_bits": 4, "observer": "minmax", "observer_kwargs": {}, "strategy": "tensor_group", "symmetric": true, "type": "float" } } }, "format": "nvfp4-pack-quantized", "global_compression_ratio": null, "ignore": [ "encoder.blocks.0.attn.query", "encoder.blocks.0.attn.key", "encoder.blocks.0.attn.value", "encoder.blocks.0.attn.out", "encoder.blocks.0.mlp.0", "encoder.blocks.0.mlp.2", "encoder.blocks.1.attn.query", "encoder.blocks.1.attn.key", "encoder.blocks.1.attn.value", "encoder.blocks.1.attn.out", "encoder.blocks.1.mlp.0", "encoder.blocks.1.mlp.2", "encoder.blocks.2.attn.query", "encoder.blocks.2.attn.key", "encoder.blocks.2.attn.value", "encoder.blocks.2.attn.out", "encoder.blocks.2.mlp.0", "encoder.blocks.2.mlp.2", "encoder.blocks.3.attn.query", "encoder.blocks.3.attn.key", "encoder.blocks.3.attn.value", "encoder.blocks.3.attn.out", "encoder.blocks.3.mlp.0", "encoder.blocks.3.mlp.2", "encoder.blocks.4.attn.query", "encoder.blocks.4.attn.key", "encoder.blocks.4.attn.value", "encoder.blocks.4.attn.out", "encoder.blocks.4.mlp.0", "encoder.blocks.4.mlp.2", "encoder.blocks.5.attn.query", "encoder.blocks.5.attn.key", "encoder.blocks.5.attn.value", "encoder.blocks.5.attn.out", "encoder.blocks.5.mlp.0", "encoder.blocks.5.mlp.2", "encoder.blocks.6.attn.query", "encoder.blocks.6.attn.key", "encoder.blocks.6.attn.value", "encoder.blocks.6.attn.out", "encoder.blocks.6.mlp.0", "encoder.blocks.6.mlp.2", "encoder.blocks.7.attn.query", "encoder.blocks.7.attn.key", "encoder.blocks.7.attn.value", "encoder.blocks.7.attn.out", "encoder.blocks.7.mlp.0", "encoder.blocks.7.mlp.2", "encoder.blocks.8.attn.query", "encoder.blocks.8.attn.key", "encoder.blocks.8.attn.value", "encoder.blocks.8.attn.out", "encoder.blocks.8.mlp.0", "encoder.blocks.8.mlp.2", "encoder.blocks.9.attn.query", "encoder.blocks.9.attn.key", "encoder.blocks.9.attn.value", "encoder.blocks.9.attn.out", "encoder.blocks.9.mlp.0", "encoder.blocks.9.mlp.2", "encoder.blocks.10.attn.query", "encoder.blocks.10.attn.key", "encoder.blocks.10.attn.value", "encoder.blocks.10.attn.out", "encoder.blocks.10.mlp.0", "encoder.blocks.10.mlp.2", "encoder.blocks.11.attn.query", "encoder.blocks.11.attn.key", "encoder.blocks.11.attn.value", "encoder.blocks.11.attn.out", "encoder.blocks.11.mlp.0", "encoder.blocks.11.mlp.2", "encoder.blocks.12.attn.query", "encoder.blocks.12.attn.key", "encoder.blocks.12.attn.value", "encoder.blocks.12.attn.out", "encoder.blocks.12.mlp.0", "encoder.blocks.12.mlp.2", "encoder.blocks.13.attn.query", "encoder.blocks.13.attn.key", "encoder.blocks.13.attn.value", "encoder.blocks.13.attn.out", "encoder.blocks.13.mlp.0", "encoder.blocks.13.mlp.2", "encoder.blocks.14.attn.query", "encoder.blocks.14.attn.key", "encoder.blocks.14.attn.value", "encoder.blocks.14.attn.out", "encoder.blocks.14.mlp.0", "encoder.blocks.14.mlp.2", "encoder.blocks.15.attn.query", "encoder.blocks.15.attn.key", "encoder.blocks.15.attn.value", "encoder.blocks.15.attn.out", "encoder.blocks.15.mlp.0", "encoder.blocks.15.mlp.2", "encoder.blocks.16.attn.query", "encoder.blocks.16.attn.key", "encoder.blocks.16.attn.value", "encoder.blocks.16.attn.out", "encoder.blocks.16.mlp.0", "encoder.blocks.16.mlp.2", "encoder.blocks.17.attn.query", "encoder.blocks.17.attn.key", "encoder.blocks.17.attn.value", "encoder.blocks.17.attn.out", "encoder.blocks.17.mlp.0", "encoder.blocks.17.mlp.2", "encoder.blocks.18.attn.query", "encoder.blocks.18.attn.key", "encoder.blocks.18.attn.value", "encoder.blocks.18.attn.out", "encoder.blocks.18.mlp.0", "encoder.blocks.18.mlp.2", "encoder.blocks.19.attn.query", "encoder.blocks.19.attn.key", "encoder.blocks.19.attn.value", "encoder.blocks.19.attn.out", "encoder.blocks.19.mlp.0", "encoder.blocks.19.mlp.2", "encoder.blocks.20.attn.query", "encoder.blocks.20.attn.key", "encoder.blocks.20.attn.value", "encoder.blocks.20.attn.out", "encoder.blocks.20.mlp.0", "encoder.blocks.20.mlp.2", "encoder.blocks.21.attn.query", "encoder.blocks.21.attn.key", "encoder.blocks.21.attn.value", "encoder.blocks.21.attn.out", "encoder.blocks.21.mlp.0", "encoder.blocks.21.mlp.2", "encoder.blocks.22.attn.query", "encoder.blocks.22.attn.key", "encoder.blocks.22.attn.value", "encoder.blocks.22.attn.out", "encoder.blocks.22.mlp.0", "encoder.blocks.22.mlp.2", "encoder.blocks.23.attn.query", "encoder.blocks.23.attn.key", "encoder.blocks.23.attn.value", "encoder.blocks.23.attn.out", "encoder.blocks.23.mlp.0", "encoder.blocks.23.mlp.2", "encoder.blocks.24.attn.query", "encoder.blocks.24.attn.key", "encoder.blocks.24.attn.value", "encoder.blocks.24.attn.out", "encoder.blocks.24.mlp.0", "encoder.blocks.24.mlp.2", "encoder.blocks.25.attn.query", "encoder.blocks.25.attn.key", "encoder.blocks.25.attn.value", "encoder.blocks.25.attn.out", "encoder.blocks.25.mlp.0", "encoder.blocks.25.mlp.2", "encoder.blocks.26.attn.query", "encoder.blocks.26.attn.key", "encoder.blocks.26.attn.value", "encoder.blocks.26.attn.out", "encoder.blocks.26.mlp.0", "encoder.blocks.26.mlp.2", "encoder.blocks.27.attn.query", "encoder.blocks.27.attn.key", "encoder.blocks.27.attn.value", "encoder.blocks.27.attn.out", "encoder.blocks.27.mlp.0", "encoder.blocks.27.mlp.2", "encoder.blocks.28.attn.query", "encoder.blocks.28.attn.key", "encoder.blocks.28.attn.value", "encoder.blocks.28.attn.out", "encoder.blocks.28.mlp.0", "encoder.blocks.28.mlp.2", "encoder.blocks.29.attn.query", "encoder.blocks.29.attn.key", "encoder.blocks.29.attn.value", "encoder.blocks.29.attn.out", "encoder.blocks.29.mlp.0", "encoder.blocks.29.mlp.2", "encoder.blocks.30.attn.query", "encoder.blocks.30.attn.key", "encoder.blocks.30.attn.value", "encoder.blocks.30.attn.out", "encoder.blocks.30.mlp.0", "encoder.blocks.30.mlp.2", "encoder.blocks.31.attn.query", "encoder.blocks.31.attn.key", "encoder.blocks.31.attn.value", "encoder.blocks.31.attn.out", "encoder.blocks.31.mlp.0", "encoder.blocks.31.mlp.2", "adapter.linear1", "adapter.linear2", "lm_head" ], "kv_cache_scheme": null, "quant_method": "compressed-tensors", "quantization_status": "compressed", "sparsity_config": {}, "transform_config": {}, "version": "0.12.2" }, "sliding_window": 2048, "text_config": { "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "dtype": "bfloat16", "hidden_act": "silu", "hidden_size": 5120, "initializer_range": 0.02, "intermediate_size": 27648, "layer_types": [ "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention", "full_attention" ], "max_position_embeddings": 65536, "max_window_layers": 28, "model_type": "qwen2", "num_attention_heads": 40, "num_hidden_layers": 64, "num_attention_groups": 8, "num_key_value_heads": 8, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": null, "use_cache": true, "use_sliding_window": false, "vocab_size": 158720 }, "tie_word_embeddings": false, "transformers_version": "4.56.2", "use_sliding_window": false }