EVA / EVA_21M /config.json
yanjiehuang's picture
Upload folder using huggingface_hub
8e6c821 verified
{
"vocab_size": 114,
"hidden_size": 256,
"num_hidden_layers": 6,
"num_attention_heads": 8,
"num_key_value_heads": 8,
"intermediate_size": 768,
"max_position_embeddings": 8192,
"rms_norm_eps": "1e-6",
"initializer_range": 0.02,
"use_cache": true,
"num_experts": 8,
"num_experts_per_tok": 2,
"expert_capacity_factor": 1.5,
"router_aux_loss_coef": 0.01,
"moe_implementation": "megablocks",
"moe_world_size": 4,
"resid_dropout": 0.0,
"hidden_dropout": 0.0,
"dropout_warmup_steps": 0,
"dropout_ramp_steps": 0,
"dropout_schedule": "linear",
"attention_dropout": 0.0,
"gradient_clip_norm": 0.0,
"label_smoothing": 0.0,
"eos_loss_weight": 10.0
}