train/models/sbert/large_config.json

{
  "stream_0": {
    "vocab_size": 21128,
    "emb_size": 1024,
    "feedforward_size": 4096,
    "hidden_size": 1024,
    "hidden_act": "gelu",
    "heads_num": 16,
    "layers_num": 24,
    "max_seq_length": 512,
    "dropout": 0.1,
    "embedding": ["word", "pos", "seg"],
    "encoder": "transformer",
    "mask": "fully_visible"
  },

  "stream_1": {
    "vocab_size": 21128,
    "emb_size": 1024,
    "feedforward_size": 4096,
    "hidden_size": 1024,
    "hidden_act": "gelu",
    "heads_num": 16,
    "layers_num": 24,
    "max_seq_length": 512,
    "dropout": 0.1,
    "embedding": ["word", "pos", "seg"],
    "encoder": "transformer",
    "mask": "fully_visible"
  },

  "embedding": "dual",
  "encoder": "dual",
  "pooling": "first",
  "tie_weights": true
}