{
  "_model_size_label": "9B",
  "architectures": [
    "ZDTaichu5_0_ForConditionalGeneration"
  ],
  "auto_map": {
    "AutoConfig": "configuration.ZDTaichu5_0_Config",
    "AutoImageProcessor": "image_processing.ZDTaichu5_0_ImageProcessor",
    "AutoModel": "modeling.ZDTaichu5_0_ForConditionalGeneration",
    "AutoModelForCausalLM": "modeling.ZDTaichu5_0_ForConditionalGeneration",
    "AutoProcessor": "processing.ZDTaichu5_0_Processor"
  },
  "bos_token_id": null,
  "downsample_ratio": 0.5,
  "dtype": "bfloat16",
  "eos_token_id": 248046,
  "force_image_size": 512,
  "hidden_size": 4096,
  "image_tag_type": "internvl",
  "img_context_token": "<|image_pad|>",
  "img_context_token_id": 248056,
  "img_end_token": "<|vision_end|>",
  "img_start_token": "<|vision_start|>",
  "keys_to_ignore_at_inference": [
    "past_key_values"
  ],
  "llm_config": {
    "architectures": [
      "Qwen3_5ForCausalLM"
    ],
    "attention_bias": false,
    "attention_dropout": 0.0,
    "bos_token_id": null,
    "dtype": "bfloat16",
    "eos_token_id": null,
    "full_attention_interval": 4,
    "head_dim": 256,
    "hidden_act": "silu",
    "hidden_size": 4096,
    "initializer_range": 0.02,
    "intermediate_size": 12288,
    "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_conv_kernel_dim": 4,
    "linear_key_head_dim": 128,
    "linear_num_key_heads": 16,
    "linear_num_value_heads": 32,
    "linear_value_head_dim": 128,
    "max_position_embeddings": 262144,
    "model_type": "qwen3_5_text",
    "mtp_loss_scaling_factor": 0.1,
    "mtp_num_layers": 1,
    "num_attention_heads": 16,
    "num_hidden_layers": 32,
    "num_key_value_heads": 4,
    "pad_token_id": 248044,
    "partial_rotary_factor": 0.25,
    "rms_norm_eps": 1e-06,
    "rope_parameters": {
      "mrope_interleaved": true,
      "mrope_section": [
        11,
        11,
        10
      ],
      "partial_rotary_factor": 0.25,
      "rope_theta": 10000000,
      "rope_type": "default"
    },
    "tie_word_embeddings": false,
    "use_cache": false,
    "mamba_ssm_dtype": "float32",
    "vocab_size": 248320
  },
  "max_dynamic_patch": 12,
  "min_dynamic_patch": 1,
  "model_type": "zdtaichu5_0",
  "pad_token_id": 248044,
  "patch_size": 16,
  "projector_hidden_size": 20480,
  "ps_version": "v2",
  "template": "qwen3_5",
  "tie_word_embeddings": false,
  "transformers_version": "5.3.0",
  "use_thumbnail": true,
  "video_context_token": "<|video_pad|>",
  "video_context_token_id": 248057,
  "vision_config": {
    "adaptor_configs": {},
    "adaptor_names": null,
    "architectures": [
      "RADIOModel"
    ],
    "args": {
      "aa": null,
      "amp": true,
      "amp_dtype": "bfloat16",
      "amp_impl": "native",
      "aug_repeats": 0,
      "aug_splits": 0,
      "auto_workload_inspector": false,
      "bn_eps": null,
      "bn_momentum": null,
      "cache_dir": null,
      "channels_last": false,
      "checkpoint_folder": null,
      "checkpoint_hist": 10,
      "chk_keep_forever": 100,
      "class_map": "",
      "clip_grad": null,
      "clip_mode": "norm",
      "cls_token_per_teacher": true,
      "coco_annotations_file": null,
      "coco_image_dir": null,
      "color_jitter": 0.4,
      "cooldown_epochs": 0,
      "cpe_max_size": 2048,
      "cpe_num_registers": null,
      "crd_loss": false,
      "crd_loss_weight": 0.8,
      "crop_pct": null,
      "cutmix": 0.0,
      "cutmix_minmax": null,
      "dataset_download": false,
      "debug_full_knn": false,
      "decay_epochs": 90,
      "decay_milestones": [
        90,
        180,
        270
      ],
      "decay_rate": 0.1,
      "depchain": true,
      "detect_anomaly": false,
      "dist_bn": "reduce",
      "dist_norm_weight": 0.0,
      "distributed": true,
      "drop": 0.0,
      "drop_block": null,
      "drop_connect": null,
      "drop_path": null,
      "dtype": "float32",
      "epoch": 299,
      "epoch_repeats": 0.0,
      "eval": false,
      "eval_metric": "knn_top1",
      "eval_teacher": false,
      "eval_teacher_only": false,
      "eval_throughput": false,
      "fast_norm": false,
      "fd_loss_fn": "MSE",
      "feature_normalization": "PHI_STANDARDIZE",
      "feature_summarizer": "cls_token",
      "feature_upscale_factor": null,
      "force_disable_damp": false,
      "force_disable_spectral_reparam": false,
      "force_new_wandb_id": false,
      "force_spectral_reparam": false,
      "freeze_bn": false,
      "fsdp": true,
      "full_equivariance": false,
      "fuser": "",
      "gp": null,
      "grad_accum_steps": 1,
      "grad_checkpointing": false,
      "head_init_bias": null,
      "head_init_scale": null,
      "head_lr": null,
      "head_warmup": 3,
      "head_weight_decay": 0.0005,
      "hflip": 0.5,
      "img_size": null,
      "in_chans": null,
      "initial_checkpoint": null,
      "input_size": null,
      "interpolation": "",
      "layer_decay": null,
      "local_rank": 0,
      "log_interval": 50,
      "log_mlflow": false,
      "log_teacher_timings": true,
      "log_train_metrics_per_epoch": true,
      "log_train_metrics_per_log_interval": true,
      "log_wandb": true,
      "loss_auto_balance": false,
      "lr_base": 0.1,
      "lr_base_scale": "",
      "lr_base_size": 256,
      "lr_cycle_decay": 0.5,
      "lr_cycle_limit": 1,
      "lr_cycle_mul": 1.0,
      "lr_k_decay": 1.0,
      "lr_noise": null,
      "lr_noise_pct": 0.67,
      "lr_noise_std": 1.0,
      "mean": null,
      "mesa": false,
      "min_lr": 1e-05,
      "mixup": 0.0,
      "mixup_mode": "batch",
      "mixup_off_epoch": 0,
      "mixup_prob": 1.0,
      "mixup_switch_prob": 0.5,
      "mlp_hidden_size": 1520,
      "mlp_num_inner": 2,
      "mlp_version": "v2",
      "model": "vit_huge_patch16_224",
      "model_kwargs": {},
      "model_norm": false,
      "momentum": 0.9,
      "no_custom_validation": false,
      "no_ddp_bb": true,
      "no_knn": false,
      "no_prefetcher": false,
      "no_resume_opt": false,
      "no_save_checkpoint": false,
      "no_val": false,
      "num_classes": null,
      "on_demand_workload_inspector": false,
      "one_logger_app_tag": "",
      "one_logger_is_baseline": false,
      "one_logger_run_name": "",
      "onelogger": null,
      "opt_betas": null,
      "opt_eps": null,
      "overfit": false,
      "patience_epochs": 10,
      "perf_test_no_aug": false,
      "perf_test_no_decode": false,
      "perf_test_no_io": false,
      "perf_test_only_dataloader": false,
      "perf_test_simple_aug": false,
      "pin_mem": false,
      "prefetcher": true,
      "pretrained": false,
      "processed_neck_outputs": null,
      "profile_train_exit_after_profiling": false,
      "profile_train_export_chrome_trace": true,
      "profile_train_export_csv": false,
      "profile_train_iterations": 0,
      "qradio": false,
      "qradio_max_tokens": 512,
      "qradio_min_tokens": 32,
      "qradio_patch_token_mask_initial_ratio": 0.95,
      "qradio_progressive_2d": false,
      "qradio_quantizer": null,
      "qradio_ramp_alpha": 1.5,
      "rank": 0,
      "ratio": [
        0.75,
        1.3333333333333333
      ],
      "recount": 1,
      "recovery_interval": 0,
      "register_multiple": 10,
      "remode": "pixel",
      "reprob": 0.0,
      "reset_loss_state": true,
      "resplit": false,
      "sample_tracking": false,
      "save_images": false,
      "scale": [
        0.5,
        1.0
      ],
      "sched": "cosine",
      "seed": 42,
      "shift_equivariance": false,
      "smoothing": 0.1,
      "source_tracking": false,
      "spectral_heads": false,
      "spectral_reparam": false,
      "spectral_weight_decay": null,
      "split_bn": false,
      "start_epoch": null,
      "std": null,
      "stream_teachers": false,
      "student_intermediate_indices": null,
      "student_load_skip_state_dict_keys_regex": null,
      "student_reinit_model_layers_regex": null,
      "student_strict_load_ignore_mismatched_shape_keys_regex": null,
      "student_strict_load_ignore_missing_keys_regex": null,
      "student_strict_load_ignore_unexpected_keys_regex": null,
      "student_strict_load_state_dict": false,
      "sync_bn": false,
      "sync_resolutions_across_ranks": true,
      "synchronize_step": false,
      "teachers": [
        {
          "model": "siglip2-g-384",
          "name": "siglip2-g",
          "spatial_mlp_version": "attn",
          "type": "siglip2",
          "use_summary": true
        },
        {
          "model": "dinov3_vit7b16",
          "name": "dino_v3_7b",
          "type": "dino_v3",
          "use_summary": true
        },
        {
          "model": "default",
          "name": "sam3",
          "type": "sam3",
          "use_summary": false
        }
      ],
      "timing_warmup_iters": 20,
      "tokenizer_kwargs": {},
      "tokenizer_type": null,
      "tome": null,
      "torchcompile": null,
      "torchscript": false,
      "train_interpolation": "random",
      "train_split": "train",
      "tta": 0,
      "untie_neck_weights": false,
      "use_coco": false,
      "use_multi_epochs_loader": false,
      "val_ema_only": false,
      "val_split": "val",
      "vflip": 0.0,
      "vitdet_version": 1,
      "wandb_entity": "",
      "wandb_id": "",
      "wandb_job_type": "",
      "wandb_name": "",
      "wandb_project": "",
      "wandb_tags": null,
      "warmup_lr": 1e-05,
      "warmup_prefix": false,
      "worker_seeding": "all",
      "workers": 8,
      "workload_inspector_analyze_nsys_traces": false,
      "workload_inspector_baseline_start_iter": 1500,
      "workload_inspector_major_slowdown_p95_factor": 10.0,
      "workload_inspector_minor_slowdown_p95_factor": 3.0,
      "workload_inspector_no_slowdown_check": false,
      "workload_inspector_simulate_slowdown_num_times": 1,
      "workload_inspector_simulate_slowdown_start_iter": null,
      "world_size": 256
    },
    "auto_map": {
      "AutoConfig": "cradio_config.RADIOConfig",
      "AutoModel": "cradio_model.RADIOModel"
    },
    "dtype": "bfloat16",
    "feature_normalizer_config": null,
    "inter_feature_normalizer_config": null,
    "max_resolution": 2048,
    "model_type": "radio",
    "patch_size": 16,
    "preferred_resolution": [
      512,
      512
    ],
    "use_flash_attn": false,
    "version": "c-radio_v4-h",
    "vitdet_window_size": null
  },
  "vit_hidden_size": 1280
}
