jeeves / config.json
nicholaswaltz's picture
Jeeves-9B: fused weights, pointer head, drafters
8622b7d verified
Raw History Blame Contribute Delete
1.47 kB
{
"vocab_size": 248320,
"hidden_size": 4096,
"intermediate_size": 12288,
"num_hidden_layers": 32,
"layer_types": [
"linear_attention",
"linear_attention",
"linear_attention",
"full_attention",
"linear_attention",
"linear_attention",
"linear_attention",
"full_attention",
"linear_attention",
"linear_attention",
"linear_attention",
"full_attention",
"linear_attention",
"linear_attention",
"linear_attention",
"full_attention",
"linear_attention",
"linear_attention",
"linear_attention",
"full_attention",
"linear_attention",
"linear_attention",
"linear_attention",
"full_attention",
"linear_attention",
"linear_attention",
"linear_attention",
"full_attention",
"linear_attention",
"linear_attention",
"linear_attention",
"full_attention"
],
"tie_word_embeddings": false,
"num_attention_heads": 16,
"num_key_value_heads": 4,
"head_dim": 256,
"attention_bias": false,
"attention_dropout": 0.0,
"linear_num_key_heads": 16,
"linear_num_value_heads": 32,
"linear_key_head_dim": 128,
"linear_value_head_dim": 128,
"linear_conv_kernel_dim": 4,
"rms_norm_eps": 1e-06,
"hidden_act": "silu",
"rope_theta": 10000000,
"partial_rotary_factor": 0.25,
"max_position_embeddings": 262144,
"pad_token_id": null,
"eos_token_id": 248044,
"bos_token_id": null,
"dtype": "bfloat16",
"initializer_range": 0.02
}