Eve-2-MoE-250M / config.json
anthonym21's picture
Upload folder using huggingface_hub
9f12aaa
{
"architecture": "Eve-2-MoE",
"vocab_size": 50304,
"n_layer": 12,
"n_embd": 512,
"n_head": 8,
"head_dim": 64,
"block_size": 2048,
"num_experts": 8,
"top_k": 2,
"expert_intermediate_size": 1408,
"shared_expert_intermediate_size": 1408,
"rope_theta": 10000.0
}