{ "model_name": "models--microsoft--Phi-4-mini-instruct_AutoRound", "model_type": "llm-phi3", "vm_cfg": null, "mm_cfg": null, "lm_cfg": { "model_type": "phi3", "data_type": "bfloat16", "arch": "phi", "token_cfg": { "vocab_size": 200064 }, "rope_cfg": { "rope_theta": 10000.0, "rope_local_base_freq": 10000.0, "rope_dimension_count": 96, "sliding_rope_dimension_count": null, "rope_scaling": { "factor": 1.0, "low_freq_factor": 0, "high_freq_factor": 0, "original_max_position_embeddings": 4096, "attention_factor": 1.1902380714238083, "long_factor": [ 1, 1.118320672, 1.250641126, 1.398617824, 1.564103225, 1.74916897, 1.956131817, 2.187582649, 2.446418898, 2.735880826, 3.059592084, 3.421605075, 3.826451687, 4.279200023, 4.785517845, 5.351743533, 5.984965424, 6.693110555, 7.485043894, 8.370679318, 9.36110372, 10.4687158, 11.70738129, 13.09260651, 14.64173252, 16.37415215, 18.31155283, 20.47818807, 22.90118105, 25.61086418, 28.64115884, 32.03, 32.1, 32.13, 32.23, 32.6, 32.61, 32.64, 32.66, 32.7, 32.71, 32.93, 32.97, 33.28, 33.49, 33.5, 44.16, 47.77 ], "short_factor": [ 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0 ], "rope_type": "longrope", "mrope_section": null, "mrope_interleaved": false } }, "attn_cfg": { "num_attention_heads": 24, "num_key_value_heads": 8, "head_dim": 128, "sliding_head_dim": null, "swa_enable": false, "sliding_window": 262144, "attention_bias": false, "attention_dropout": 0.0, "query_pre_attn_scalar": 0 }, "mlp_cfg": { "intermediate_size": 8192, "act": "silu", "num_layers": 3, "mlp_bias": false }, "hidden_size": 3072, "num_hidden_layers": 32, "max_position_embeddings": 2048, "rms_norm_eps": 1e-05, "rms_norm_unit_offset": false, "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" ], "hidden_size_per_layer_input": 0, "num_kv_shared_layers": 0, "use_double_wide_mlp": false, "attn_logit_softcapping": null, "final_logit_softcapping": null, "lm_head_num_splits": 4, "lm_head_split_dim": 50016, "draft_vocab_size": 0, "conv_L_cache": 3, "conv_bias": false, "lora_cfg": null, "speculative_decoding_cfg": null }, "pipeline_cfg": { "system_prompt": null, "chat_template": null, "max_num_tokens": 2048, "input_token_group_size": 128, "input_token_group_offsets": [ 0, 128, 256, 384, 512, 640, 768, 896, 1024, 1152, 1280, 1408, 1536, 1664, 1792, 1920 ], "future_token_mask_size": 128, "return_logits": false, "enable_filter_sharing": false, "quantize_embeddings": false, "split_mlp": true }, "language_model_name": "models--microsoft--Phi-4-mini-instruct_AutoRound_language" }