config.json

{
  "architectures": [
    "KeyeForConditionalGeneration"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "configuration_keye.KeyeConfig",
    "AutoModel": "modeling_keye.KeyeForConditionalGeneration",
    "AutoModelForCausalLM": "modeling_keye.KeyeForConditionalGeneration"
  },
  "bos_token_id": 151643,
  "eos_token_id": 151645,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "image_token_id": 151655,
  "initializer_factor": 1.0,
  "initializer_range": 0.02,
  "intermediate_size": 12288,
  "max_position_embeddings": 40960,
  "max_window_layers": 36,
  "model_type": "Keye",
  "num_attention_heads": 32,
  "num_hidden_layers": 36,
  "num_key_value_heads": 8,
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "mrope_section": [
      16,
      24,
      24
    ],
    "rope_type": "default",
    "type": "default"
  },
  "rope_theta": 1000000,
  "sliding_window": null,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.52.1",
  "use_cache": true,
  "use_sliding_window": false,
  "video_token_id": 151656,
  "vision_config": {
    "architectures": [
      "SiglipVisionModel"
    ],
    "attention_dropout": 0.0,
    "auto_map": {
      "AutoConfig": "configuration_keye.KeyeVisionConfig",
      "AutoModel": "modeling_keye.SiglipVisionModel"
    },
    "hidden_act": "gelu_pytorch_tanh",
    "hidden_size": 1152,
    "image_size": 384,
    "intermediate_size": 4304,
    "layer_norm_eps": 1e-06,
    "model_type": "Keye",
    "num_attention_heads": 16,
    "num_channels": 3,
    "num_hidden_layers": 27,
    "patch_size": 14,
    "spatial_merge_size": 2,
    "temporal_patch_size": 2,
    "tokens_per_second": 2,
    "torch_dtype": "bfloat16"
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 151936
}