tinystories-50m / config.json
Compactbot's picture
v2 config: vocab 12288, 56,902,144 params
563c223 verified
Raw
History Blame
299 Bytes
{
"architectures": ["TinyStoriesGPT"],
"d_model": 512,
"n_layers": 16,
"n_heads": 8,
"ffn_dim": 2048,
"vocab_size": 12288,
"max_seq_len": 512,
"tie_word_embeddings": true,
"norm": "rmsnorm",
"activation": "gelu",
"attn": "causal",
"dtype": "float32",
"n_params": 56902144
}