align/AlignModel/config.json

{
  "architectures": [
    "AlignModel"
  ],
  "dtype": "float32",
  "initializer_range": 0.02,
  "model_type": "align",
  "projection_dim": 64,
  "temperature_init_value": 1.0,
  "text_config": {
    "attention_probs_dropout_prob": 0.1,
    "bos_token_id": null,
    "eos_token_id": null,
    "hidden_act": "gelu",
    "hidden_dropout_prob": 0.1,
    "hidden_size": 32,
    "initializer_range": 0.02,
    "intermediate_size": 37,
    "is_decoder": false,
    "layer_norm_eps": 1e-12,
    "max_position_embeddings": 512,
    "model_type": "align_text_model",
    "num_attention_heads": 4,
    "num_hidden_layers": 2,
    "pad_token_id": 0,
    "type_vocab_size": 16,
    "vocab_size": 1024
  },
  "transformers_version": "5.15.0.dev0",
  "vision_config": {
    "batch_norm_eps": 0.001,
    "batch_norm_momentum": 0.99,
    "depth_coefficient": 3.1,
    "depth_divisor": 8,
    "depthwise_padding": [],
    "drop_connect_rate": 0.2,
    "expand_ratios": [
      1,
      6,
      6
    ],
    "hidden_act": "gelu",
    "hidden_dim": 64,
    "image_size": 600,
    "in_channels": [
      32,
      16,
      24
    ],
    "initializer_range": 0.02,
    "kernel_sizes": [
      3,
      3,
      5
    ],
    "model_type": "align_vision_model",
    "num_block_repeats": [
      1,
      1,
      2
    ],
    "num_channels": 3,
    "num_hidden_layers": 16,
    "out_channels": [
      16,
      24,
      30
    ],
    "pooling_type": "mean",
    "squeeze_expansion_ratio": 0.25,
    "strides": [
      1,
      1,
      2
    ],
    "width_coefficient": 2.0
  }
}