{ | |
"producer": { | |
"name": "ammo", | |
"version": "0.7.3" | |
}, | |
"architecture": "LlamaForCausalLM", | |
"dtype": "float16", | |
"num_hidden_layers": 48, | |
"num_attention_heads": 32, | |
"num_key_value_heads": 8, | |
"hidden_size": 4096, | |
"norm_epsilon": 1e-05, | |
"vocab_size": 32000, | |
"max_position_embeddings": 4096, | |
"hidden_act": "silu", | |
"use_parallel_embedding": true, | |
"embedding_sharding_dim": 0, | |
"quantization": { | |
"quant_algo": null, | |
"kv_cache_quant_algo": "FP8" | |
}, | |
"mapping": { | |
"world_size": 1, | |
"tp_size": 1, | |
"pp_size": 1 | |
}, | |
"head_size": 128, | |
"intermediate_size": 14336, | |
"position_embedding_type": "rope_gpt_neox" | |
} |