LM-from-scratch / config.json
flbjcy's picture
Upload TinyStories Transformer LM (from-scratch implementation)
bfafe8d verified
Raw History Blame Contribute Delete
145 Bytes
{
"vocab_size": 10000,
"context_length": 256,
"d_model": 512,
"d_ff": 1344,
"num_layers": 4,
"num_heads": 16,
"rope_theta": 10000
}