speedai-125M-config.json

{
  "architectures": [
    "GPTModel"
  ],
  "hidden_size": 768,
  "intermediate_size": 3072,
  "num_attention_heads": 12,
  "num_hidden_layers": 12,
  "vocab_size": 50257,
  "max_position_embeddings": 2048,
  "bos_token_id": 0,
  "eos_token_id": 1,
  "pad_token_id": 2,
  "tie_word_embeddings": true,
  "initializer_range": 0.02,
  "layer_norm_epsilon": 1e-05,
  "torch_dtype": "float32",
  "use_cache": true,
  "model_type": "gpt2",
  "attn_pdrop": 0.1,
  "embd_pdrop": 0.1,
  "resid_pdrop": 0.1
}