train/models/gpt2/distil_config.json

{
  "emb_size": 768,
  "feedforward_size": 3072,
  "hidden_size": 768,
  "hidden_act": "gelu_fast",
  "heads_num": 12,
  "layers_num": 6,
  "max_seq_length": 1024,
  "dropout": 0.1,
  "data_processor": "lm",
  "embedding": ["word", "pos"],
  "remove_embedding_layernorm": true,
  "encoder": "transformer",
  "mask": "causal",
  "layernorm_positioning": "pre",
  "target": ["lm"],
  "tie_weights": true
}