config.json

{
  "_name_or_path": "misecom_model/v1",
  "architectures": [
    "MiSeComModel"
  ],
  "attention_probs_dropout_prob": 0.1,
  "bos_token_id": 0,
  "classifier_dropout": null,
  "eos_token_id": 2,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.1,
  "hidden_size": 768,
  "id2label": {
    "0": "<ma>",
    "1": "<madv>",
    "2": "<mn>",
    "3": "<mp>",
    "4": "<ms>",
    "5": "<mv>"
  },
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "<ma>": 0,
    "<madv>": 1,
    "<mn>": 2,
    "<mp>": 3,
    "<ms>": 4,
    "<mv>": 5
  },
  "layer_norm_eps": 1e-05,
  "layers_use_from_last": 2,
  "max_position_embeddings": 514,
  "method_for_layers": "mean",
  "model_type": "misecom",
  "num_attention_heads": 12,
  "num_hidden_layers": 12,
  "output_hidden_states": true,
  "pad_token_id": 1,
  "position_embedding_type": "absolute",
  "pretrained_ck": "",
  "torch_dtype": "float32",
  "transformers_version": "4.25.1",
  "type_vocab_size": 1,
  "use_cache": true,
  "vocab_size": 50265
}