slim-llm-llama-7b / config.json
jp288881's picture
Upload 7 files
75b60ee verified
{
"_name_or_path": "huggyllama/llama-7b",
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 4096,
"initializer_range": 0.02,
"intermediate_size": 11008,
"max_position_embeddings": 2048,
"max_sequence_length": 2048,
"model_type": "llama",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 32,
"pad_token_id": 0,
"pretraining_tp": 1,
"quantization_config": {
"bits": 2,
"checkpoint_format": "gptq",
"damp_percent": 0.01,
"desc_act": false,
"group_size": 128,
"model_file_base_name": null,
"model_name_or_path": null,
"quant_method": "gptq",
"qweight_heights": {
"model.layers.0.mlp.down_proj": 692,
"model.layers.0.mlp.gate_proj": 280,
"model.layers.0.mlp.up_proj": 276,
"model.layers.0.self_attn.k_proj": 284,
"model.layers.0.self_attn.o_proj": 296,
"model.layers.0.self_attn.q_proj": 276,
"model.layers.0.self_attn.v_proj": 268,
"model.layers.1.mlp.down_proj": 692,
"model.layers.1.mlp.gate_proj": 288,
"model.layers.1.mlp.up_proj": 316,
"model.layers.1.self_attn.k_proj": 268,
"model.layers.1.self_attn.o_proj": 316,
"model.layers.1.self_attn.q_proj": 268,
"model.layers.1.self_attn.v_proj": 276,
"model.layers.10.mlp.down_proj": 856,
"model.layers.10.mlp.gate_proj": 280,
"model.layers.10.mlp.up_proj": 316,
"model.layers.10.self_attn.k_proj": 268,
"model.layers.10.self_attn.o_proj": 316,
"model.layers.10.self_attn.q_proj": 276,
"model.layers.10.self_attn.v_proj": 280,
"model.layers.11.mlp.down_proj": 852,
"model.layers.11.mlp.gate_proj": 272,
"model.layers.11.mlp.up_proj": 312,
"model.layers.11.self_attn.k_proj": 308,
"model.layers.11.self_attn.o_proj": 316,
"model.layers.11.self_attn.q_proj": 280,
"model.layers.11.self_attn.v_proj": 268,
"model.layers.12.mlp.down_proj": 856,
"model.layers.12.mlp.gate_proj": 276,
"model.layers.12.mlp.up_proj": 316,
"model.layers.12.self_attn.k_proj": 300,
"model.layers.12.self_attn.o_proj": 316,
"model.layers.12.self_attn.q_proj": 272,
"model.layers.12.self_attn.v_proj": 268,
"model.layers.13.mlp.down_proj": 856,
"model.layers.13.mlp.gate_proj": 308,
"model.layers.13.mlp.up_proj": 316,
"model.layers.13.self_attn.k_proj": 272,
"model.layers.13.self_attn.o_proj": 316,
"model.layers.13.self_attn.q_proj": 276,
"model.layers.13.self_attn.v_proj": 264,
"model.layers.14.mlp.down_proj": 856,
"model.layers.14.mlp.gate_proj": 272,
"model.layers.14.mlp.up_proj": 316,
"model.layers.14.self_attn.k_proj": 296,
"model.layers.14.self_attn.o_proj": 316,
"model.layers.14.self_attn.q_proj": 280,
"model.layers.14.self_attn.v_proj": 268,
"model.layers.15.mlp.down_proj": 856,
"model.layers.15.mlp.gate_proj": 292,
"model.layers.15.mlp.up_proj": 316,
"model.layers.15.self_attn.k_proj": 296,
"model.layers.15.self_attn.o_proj": 316,
"model.layers.15.self_attn.q_proj": 268,
"model.layers.15.self_attn.v_proj": 316,
"model.layers.16.mlp.down_proj": 856,
"model.layers.16.mlp.gate_proj": 272,
"model.layers.16.mlp.up_proj": 316,
"model.layers.16.self_attn.k_proj": 304,
"model.layers.16.self_attn.o_proj": 316,
"model.layers.16.self_attn.q_proj": 272,
"model.layers.16.self_attn.v_proj": 276,
"model.layers.17.mlp.down_proj": 856,
"model.layers.17.mlp.gate_proj": 268,
"model.layers.17.mlp.up_proj": 316,
"model.layers.17.self_attn.k_proj": 280,
"model.layers.17.self_attn.o_proj": 316,
"model.layers.17.self_attn.q_proj": 268,
"model.layers.17.self_attn.v_proj": 316,
"model.layers.18.mlp.down_proj": 856,
"model.layers.18.mlp.gate_proj": 300,
"model.layers.18.mlp.up_proj": 316,
"model.layers.18.self_attn.k_proj": 264,
"model.layers.18.self_attn.o_proj": 316,
"model.layers.18.self_attn.q_proj": 264,
"model.layers.18.self_attn.v_proj": 316,
"model.layers.19.mlp.down_proj": 856,
"model.layers.19.mlp.gate_proj": 276,
"model.layers.19.mlp.up_proj": 316,
"model.layers.19.self_attn.k_proj": 312,
"model.layers.19.self_attn.o_proj": 316,
"model.layers.19.self_attn.q_proj": 264,
"model.layers.19.self_attn.v_proj": 316,
"model.layers.2.mlp.down_proj": 856,
"model.layers.2.mlp.gate_proj": 316,
"model.layers.2.mlp.up_proj": 316,
"model.layers.2.self_attn.k_proj": 260,
"model.layers.2.self_attn.o_proj": 272,
"model.layers.2.self_attn.q_proj": 260,
"model.layers.2.self_attn.v_proj": 268,
"model.layers.20.mlp.down_proj": 856,
"model.layers.20.mlp.gate_proj": 300,
"model.layers.20.mlp.up_proj": 316,
"model.layers.20.self_attn.k_proj": 308,
"model.layers.20.self_attn.o_proj": 316,
"model.layers.20.self_attn.q_proj": 268,
"model.layers.20.self_attn.v_proj": 316,
"model.layers.21.mlp.down_proj": 856,
"model.layers.21.mlp.gate_proj": 280,
"model.layers.21.mlp.up_proj": 316,
"model.layers.21.self_attn.k_proj": 272,
"model.layers.21.self_attn.o_proj": 312,
"model.layers.21.self_attn.q_proj": 280,
"model.layers.21.self_attn.v_proj": 316,
"model.layers.22.mlp.down_proj": 856,
"model.layers.22.mlp.gate_proj": 288,
"model.layers.22.mlp.up_proj": 316,
"model.layers.22.self_attn.k_proj": 316,
"model.layers.22.self_attn.o_proj": 312,
"model.layers.22.self_attn.q_proj": 264,
"model.layers.22.self_attn.v_proj": 316,
"model.layers.23.mlp.down_proj": 856,
"model.layers.23.mlp.gate_proj": 300,
"model.layers.23.mlp.up_proj": 316,
"model.layers.23.self_attn.k_proj": 300,
"model.layers.23.self_attn.o_proj": 300,
"model.layers.23.self_attn.q_proj": 260,
"model.layers.23.self_attn.v_proj": 316,
"model.layers.24.mlp.down_proj": 856,
"model.layers.24.mlp.gate_proj": 280,
"model.layers.24.mlp.up_proj": 316,
"model.layers.24.self_attn.k_proj": 276,
"model.layers.24.self_attn.o_proj": 316,
"model.layers.24.self_attn.q_proj": 264,
"model.layers.24.self_attn.v_proj": 316,
"model.layers.25.mlp.down_proj": 856,
"model.layers.25.mlp.gate_proj": 300,
"model.layers.25.mlp.up_proj": 316,
"model.layers.25.self_attn.k_proj": 312,
"model.layers.25.self_attn.o_proj": 312,
"model.layers.25.self_attn.q_proj": 264,
"model.layers.25.self_attn.v_proj": 316,
"model.layers.26.mlp.down_proj": 856,
"model.layers.26.mlp.gate_proj": 304,
"model.layers.26.mlp.up_proj": 316,
"model.layers.26.self_attn.k_proj": 296,
"model.layers.26.self_attn.o_proj": 316,
"model.layers.26.self_attn.q_proj": 264,
"model.layers.26.self_attn.v_proj": 316,
"model.layers.27.mlp.down_proj": 848,
"model.layers.27.mlp.gate_proj": 300,
"model.layers.27.mlp.up_proj": 272,
"model.layers.27.self_attn.k_proj": 288,
"model.layers.27.self_attn.o_proj": 316,
"model.layers.27.self_attn.q_proj": 268,
"model.layers.27.self_attn.v_proj": 316,
"model.layers.28.mlp.down_proj": 856,
"model.layers.28.mlp.gate_proj": 300,
"model.layers.28.mlp.up_proj": 272,
"model.layers.28.self_attn.k_proj": 308,
"model.layers.28.self_attn.o_proj": 316,
"model.layers.28.self_attn.q_proj": 268,
"model.layers.28.self_attn.v_proj": 316,
"model.layers.29.mlp.down_proj": 840,
"model.layers.29.mlp.gate_proj": 260,
"model.layers.29.mlp.up_proj": 272,
"model.layers.29.self_attn.k_proj": 316,
"model.layers.29.self_attn.o_proj": 312,
"model.layers.29.self_attn.q_proj": 268,
"model.layers.29.self_attn.v_proj": 316,
"model.layers.3.mlp.down_proj": 856,
"model.layers.3.mlp.gate_proj": 316,
"model.layers.3.mlp.up_proj": 316,
"model.layers.3.self_attn.k_proj": 260,
"model.layers.3.self_attn.o_proj": 272,
"model.layers.3.self_attn.q_proj": 268,
"model.layers.3.self_attn.v_proj": 268,
"model.layers.30.mlp.down_proj": 760,
"model.layers.30.mlp.gate_proj": 264,
"model.layers.30.mlp.up_proj": 260,
"model.layers.30.self_attn.k_proj": 288,
"model.layers.30.self_attn.o_proj": 316,
"model.layers.30.self_attn.q_proj": 268,
"model.layers.30.self_attn.v_proj": 316,
"model.layers.31.mlp.down_proj": 792,
"model.layers.31.mlp.gate_proj": 300,
"model.layers.31.mlp.up_proj": 264,
"model.layers.31.self_attn.k_proj": 304,
"model.layers.31.self_attn.o_proj": 308,
"model.layers.31.self_attn.q_proj": 268,
"model.layers.31.self_attn.v_proj": 264,
"model.layers.4.mlp.down_proj": 744,
"model.layers.4.mlp.gate_proj": 316,
"model.layers.4.mlp.up_proj": 316,
"model.layers.4.self_attn.k_proj": 292,
"model.layers.4.self_attn.o_proj": 316,
"model.layers.4.self_attn.q_proj": 272,
"model.layers.4.self_attn.v_proj": 276,
"model.layers.5.mlp.down_proj": 856,
"model.layers.5.mlp.gate_proj": 316,
"model.layers.5.mlp.up_proj": 316,
"model.layers.5.self_attn.k_proj": 260,
"model.layers.5.self_attn.o_proj": 316,
"model.layers.5.self_attn.q_proj": 272,
"model.layers.5.self_attn.v_proj": 276,
"model.layers.6.mlp.down_proj": 788,
"model.layers.6.mlp.gate_proj": 312,
"model.layers.6.mlp.up_proj": 316,
"model.layers.6.self_attn.k_proj": 284,
"model.layers.6.self_attn.o_proj": 312,
"model.layers.6.self_attn.q_proj": 264,
"model.layers.6.self_attn.v_proj": 276,
"model.layers.7.mlp.down_proj": 848,
"model.layers.7.mlp.gate_proj": 276,
"model.layers.7.mlp.up_proj": 316,
"model.layers.7.self_attn.k_proj": 268,
"model.layers.7.self_attn.o_proj": 316,
"model.layers.7.self_attn.q_proj": 268,
"model.layers.7.self_attn.v_proj": 276,
"model.layers.8.mlp.down_proj": 832,
"model.layers.8.mlp.gate_proj": 312,
"model.layers.8.mlp.up_proj": 316,
"model.layers.8.self_attn.k_proj": 272,
"model.layers.8.self_attn.o_proj": 296,
"model.layers.8.self_attn.q_proj": 268,
"model.layers.8.self_attn.v_proj": 276,
"model.layers.9.mlp.down_proj": 756,
"model.layers.9.mlp.gate_proj": 292,
"model.layers.9.mlp.up_proj": 296,
"model.layers.9.self_attn.k_proj": 264,
"model.layers.9.self_attn.o_proj": 316,
"model.layers.9.self_attn.q_proj": 260,
"model.layers.9.self_attn.v_proj": 276
},
"static_groups": false,
"sym": true,
"true_sequential": true
},
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.35.0",
"use_cache": true,
"vocab_size": 32000
}