config.json

{
  "architectures": [
    "PPFormulaNetForConditionalGeneration"
  ],
  "dtype": "float32",
  "is_encoder_decoder": true,
  "model_type": "pp_formulanet",
  "num_hidden_layers": 1,
  "text_config": {
    "activation_dropout": 0.0,
    "activation_function": "gelu",
    "attention_dropout": 0.0,
    "bos_token_id": 0,
    "d_model": 48,
    "decoder_attention_heads": 16,
    "decoder_ffn_dim": 16,
    "decoder_layerdrop": 0.0,
    "decoder_layers": 1,
    "decoder_start_token_id": 2,
    "dropout": 0.1,
    "encoder_attention_heads": 16,
    "encoder_layers": 12,
    "eos_token_id": 2,
    "forced_eos_token_id": 2,
    "init_std": 0.02,
    "max_position_embeddings": 2560,
    "model_type": "pp_formulanet",
    "pad_token_id": 1,
    "scale_embedding": true,
    "tie_word_embeddings": false,
    "use_cache": true,
    "vocab_size": 50000
  },
  "transformers_version": "5.16.0.dev0",
  "vision_config": {
    "attention_dropout": 0.0,
    "decoder_hidden_size": 48,
    "global_attn_indexes": [
      1,
      1,
      1,
      1
    ],
    "hidden_act": "gelu",
    "hidden_size": 48,
    "image_size": 768,
    "initializer_range": 1e-10,
    "layer_norm_eps": 1e-06,
    "mlp_dim": 1,
    "model_type": "",
    "num_attention_heads": 2,
    "num_channels": 3,
    "num_hidden_layers": 1,
    "output_channels": 16,
    "patch_size": 16,
    "post_conv_in_channels": 16,
    "post_conv_mid_channels": 16,
    "post_conv_out_channels": 16,
    "qkv_bias": true,
    "use_abs_pos": true,
    "use_rel_pos": true,
    "window_size": 14,
    "windows_size": 14
  }
}