mlx-community/Ring-mini-linear-2.0-4bit
373
1{2 "architectures": [3 "BailingMoeLinearV2ForCausalLM"4 ],5 "attention_dropout": 0.0,6 "auto_map": {7 "AutoConfig": "configuration_bailing_moe_linear_v2.BailingMoeLinearV2Config",8 "AutoModel": "modeling_bailing_moe_linear_v2.BailingMoeLinearV2Model",9 "AutoModelForCausalLM": "modeling_bailing_moe_linear_v2.BailingMoeLinearV2ForCausalLM"10 },11 "embedding_dropout": 0.0,12 "eos_token_id": 156892,13 "first_k_dense_replace": 1,14 "group_norm_size": 4,15 "head_dim": 128,16 "hidden_act": "silu",17 "hidden_size": 2048,18 "intermediate_size": 5120,19 "layer_group_size": 5,20 "linear_silu": false,21 "max_position_embeddings": 131072,22 "model_type": "bailing_moe_linear",23 "moe_intermediate_size": 512,24 "moe_router_enable_expert_bias": true,25 "moe_shared_expert_intermediate_size": 512,26 "n_group": 8,27 "norm_topk_prob": true,28 "num_attention_heads": 16,29 "num_experts": 256,30 "num_experts_per_tok": 8,31 "num_hidden_layers": 20,32 "num_key_value_heads": 4,33 "num_nextn_predict_layers": 0,34 "num_shared_experts": 1,35 "output_dropout": 0.0,36 "pad_token_id": 156892,37 "partial_rotary_factor": 0.5,38 "quantization": {39 "group_size": 64,40 "bits": 4,41 "mode": "affine",42 "model.layers.1.mlp.gate.gate_proj": {43 "group_size": 64,44 "bits": 845 },46 "model.layers.2.mlp.gate.gate_proj": {47 "group_size": 64,48 "bits": 849 },50 "model.layers.3.mlp.gate.gate_proj": {51 "group_size": 64,52 "bits": 853 },54 "model.layers.4.mlp.gate.gate_proj": {55 "group_size": 64,56 "bits": 857 },58 "model.layers.5.mlp.gate.gate_proj": {59 "group_size": 64,60 "bits": 861 },62 "model.layers.6.mlp.gate.gate_proj": {63 "group_size": 64,64 "bits": 865 },66 "model.layers.7.mlp.gate.gate_proj": {67 "group_size": 64,68 "bits": 869 },70 "model.layers.8.mlp.gate.gate_proj": {71 "group_size": 64,72 "bits": 873 },74 "model.layers.9.mlp.gate.gate_proj": {75 "group_size": 64,76 "bits": 877 },78 "model.layers.10.mlp.gate.gate_proj": {79 "group_size": 64,80 "bits": 881 },82 "model.layers.11.mlp.gate.gate_proj": {83 "group_size": 64,84 "bits": 885 },86 "model.layers.12.mlp.gate.gate_proj": {87 "group_size": 64,88 "bits": 889 },90 "model.layers.13.mlp.gate.gate_proj": {91 "group_size": 64,92 "bits": 893 },94 "model.layers.14.mlp.gate.gate_proj": {95 "group_size": 64,96 "bits": 897 },98 "model.layers.15.mlp.gate.gate_proj": {99 "group_size": 64,100 "bits": 8101 },102 "model.layers.16.mlp.gate.gate_proj": {103 "group_size": 64,104 "bits": 8105 },106 "model.layers.17.mlp.gate.gate_proj": {107 "group_size": 64,108 "bits": 8109 },110 "model.layers.18.mlp.gate.gate_proj": {111 "group_size": 64,112 "bits": 8113 },114 "model.layers.19.mlp.gate.gate_proj": {115 "group_size": 64,116 "bits": 8117 }118 },119 "quantization_config": {120 "group_size": 64,121 "bits": 4,122 "mode": "affine",123 "model.layers.1.mlp.gate.gate_proj": {124 "group_size": 64,125 "bits": 8126 },127 "model.layers.2.mlp.gate.gate_proj": {128 "group_size": 64,129 "bits": 8130 },131 "model.layers.3.mlp.gate.gate_proj": {132 "group_size": 64,133 "bits": 8134 },135 "model.layers.4.mlp.gate.gate_proj": {136 "group_size": 64,137 "bits": 8138 },139 "model.layers.5.mlp.gate.gate_proj": {140 "group_size": 64,141 "bits": 8142 },143 "model.layers.6.mlp.gate.gate_proj": {144 "group_size": 64,145 "bits": 8146 },147 "model.layers.7.mlp.gate.gate_proj": {148 "group_size": 64,149 "bits": 8150 },151 "model.layers.8.mlp.gate.gate_proj": {152 "group_size": 64,153 "bits": 8154 },155 "model.layers.9.mlp.gate.gate_proj": {156 "group_size": 64,157 "bits": 8158 },159 "model.layers.10.mlp.gate.gate_proj": {160 "group_size": 64,161 "bits": 8162 },163 "model.layers.11.mlp.gate.gate_proj": {164 "group_size": 64,165 "bits": 8166 },167 "model.layers.12.mlp.gate.gate_proj": {168 "group_size": 64,169 "bits": 8170 },171 "model.layers.13.mlp.gate.gate_proj": {172 "group_size": 64,173 "bits": 8174 },175 "model.layers.14.mlp.gate.gate_proj": {176 "group_size": 64,177 "bits": 8178 },179 "model.layers.15.mlp.gate.gate_proj": {180 "group_size": 64,181 "bits": 8182 },183 "model.layers.16.mlp.gate.gate_proj": {184 "group_size": 64,185 "bits": 8186 },187 "model.layers.17.mlp.gate.gate_proj": {188 "group_size": 64,189 "bits": 8190 },191 "model.layers.18.mlp.gate.gate_proj": {192 "group_size": 64,193 "bits": 8194 },195 "model.layers.19.mlp.gate.gate_proj": {196 "group_size": 64,197 "bits": 8198 }199 },200 "rms_norm_eps": 1e-06,201 "rope_scaling": null,202 "rope_theta": 1000000,203 "routed_scaling_factor": 2.5,204 "router_dtype": "fp32",205 "score_function": "sigmoid",206 "tie_word_embeddings": false,207 "topk_group": 4,208 "torch_dtype": "bfloat16",209 "transformers_version": "4.56.1",210 "use_bias": false,211 "use_cache": true,212 "use_qk_norm": true,213 "use_qkv_bias": false,214 "use_rmsnorm": true,215 "vocab_size": 157184216}