{
    "model_name": "models--simaai--LFM2-VL-3B-Autoround-Safetensors",
    "model_type": "vlm-lfm2_vl",
    "vm_cfg": {
        "model_type": "siglip2_vision_model",
        "arch": "siglip2",
        "image_size": [
            512,
            512
        ],
        "patch_size": 16,
        "cls_embed": false,
        "hidden_size": 1152,
        "intermediate_size": 4304,
        "num_attention_heads": 16,
        "num_hidden_layers": 27,
        "hidden_act": "gelu_pytorch_tanh",
        "layer_norm_eps": 1e-06,
        "spatial_merge_size": 0,
        "temporal_patch_size": 1,
        "window_size": 0,
        "num_position_embeddings": 0,
        "fullatt_block_indexes": [],
        "deepstack_visual_indexes": [],
        "rope_theta": 10000.0
    },
    "mm_cfg": {
        "num_layers": 2,
        "hidden_act": "gelu",
        "mm_tokens_per_image": 256,
        "proj_dim": 2048,
        "downsample_factor": 2,
        "projector_use_layernorm": true
    },
    "lm_cfg": {
        "model_type": "lfm2",
        "data_type": "bfloat16",
        "arch": "lfm",
        "token_cfg": {
            "vocab_size": 65536
        },
        "rope_cfg": {
            "rope_theta": 1000000.0,
            "rope_local_base_freq": 1000000.0,
            "rope_dimension_count": 64,
            "sliding_rope_dimension_count": null,
            "rope_scaling": {
                "factor": 1.0,
                "low_freq_factor": 0,
                "high_freq_factor": 0,
                "original_max_position_embeddings": 0,
                "attention_factor": null,
                "long_factor": null,
                "short_factor": null,
                "rope_type": "default",
                "mrope_section": null,
                "mrope_interleaved": false
            }
        },
        "attn_cfg": {
            "num_attention_heads": 32,
            "num_key_value_heads": 8,
            "head_dim": 64,
            "sliding_head_dim": null,
            "swa_enable": false,
            "sliding_window": 0,
            "attention_bias": false,
            "attention_dropout": 0.0,
            "query_pre_attn_scalar": 0
        },
        "mlp_cfg": {
            "intermediate_size": 10752,
            "act": "silu",
            "num_layers": 3,
            "mlp_bias": false
        },
        "hidden_size": 2048,
        "num_hidden_layers": 30,
        "max_position_embeddings": 2048,
        "rms_norm_eps": 1e-05,
        "rms_norm_unit_offset": false,
        "layer_types": [
            "conv",
            "conv",
            "full_attention",
            "conv",
            "conv",
            "full_attention",
            "conv",
            "conv",
            "conv",
            "full_attention",
            "conv",
            "conv",
            "conv",
            "full_attention",
            "conv",
            "conv",
            "conv",
            "full_attention",
            "conv",
            "conv",
            "conv",
            "full_attention",
            "conv",
            "conv",
            "full_attention",
            "conv",
            "conv",
            "full_attention",
            "conv",
            "conv"
        ],
        "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": 1,
        "lm_head_split_dim": 65536,
        "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": 8192,
        "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,
            2048,
            2176,
            2304,
            2432,
            2560,
            2688,
            2816,
            2944,
            3072,
            3200,
            3328,
            3456,
            3584,
            3712,
            3840,
            3968,
            4096,
            4224,
            4352,
            4480,
            4608,
            4736,
            4864,
            4992,
            5120,
            5248,
            5376,
            5504,
            5632,
            5760,
            5888,
            6016,
            6144,
            6272,
            6400,
            6528,
            6656,
            6784,
            6912,
            7040,
            7168,
            7296,
            7424,
            7552,
            7680,
            7808,
            7936,
            8064
        ],
        "future_token_mask_size": 128,
        "long_context_future_token_mask_size": 1024,
        "return_logits": false,
        "enable_filter_sharing": true,
        "quantize_embeddings": true,
        "quantize_kv_cache": true,
        "split_mlp": true
    },
    "language_model_name": "models--simaai--LFM2-VL-3B-Autoround-Safetensors_language",
    "vision_model_name": "models--simaai--LFM2-VL-3B-Autoround-Safetensors_vision"
}