{
  "architectures": [
    "NorbertForMaskedLM"
  ],
  "attention_probs_dropout_prob": 0.1,
  "hidden_dropout_prob": 0.1,
  "hidden_size": 1024,
  "intermediate_size": 2730,
  "layer_norm_eps": 1e-07,
  "max_position_embeddings": 512,
  "num_attention_heads": 16,
  "num_hidden_layers": 24,
  "position_bucket_size": 32,
  "torch_dtype": "float32",
  "transformers_version": "4.23.1",
  "vocab_size": 50000
}