config.json
| 1 | { |
| 2 | "architectures": [ |
| 3 | "GlmMoeDsaForCausalLM" |
| 4 | ], |
| 5 | "attention_bias": false, |
| 6 | "attention_dropout": 0.0, |
| 7 | "dtype": "bfloat16", |
| 8 | "eos_token_id": [ |
| 9 | 154820, |
| 10 | 154827, |
| 11 | 154829 |
| 12 | ], |
| 13 | "ep_size": 1, |
| 14 | "first_k_dense_replace": 3, |
| 15 | "hidden_act": "silu", |
| 16 | "head_dim": 64, |
| 17 | "hidden_size": 6144, |
| 18 | "index_head_dim": 128, |
| 19 | "index_n_heads": 32, |
| 20 | "index_topk": 2048, |
| 21 | "indexer_rope_interleave": true, |
| 22 | "initializer_range": 0.02, |
| 23 | "intermediate_size": 12288, |
| 24 | "kv_lora_rank": 512, |
| 25 | "max_position_embeddings": 202752, |
| 26 | "moe_intermediate_size": 2048, |
| 27 | "moe_layer_freq": 1, |
| 28 | "model_type": "glm_moe_dsa", |
| 29 | "n_group": 1, |
| 30 | "n_routed_experts": 256, |
| 31 | "n_shared_experts": 1, |
| 32 | "norm_topk_prob": true, |
| 33 | "num_attention_heads": 64, |
| 34 | "num_experts_per_tok": 8, |
| 35 | "num_hidden_layers": 78, |
| 36 | "num_key_value_heads": 64, |
| 37 | "num_nextn_predict_layers": 1, |
| 38 | "pad_token_id": 154820, |
| 39 | "pretraining_tp": 1, |
| 40 | "q_lora_rank": 2048, |
| 41 | "qk_head_dim": 256, |
| 42 | "qk_nope_head_dim": 192, |
| 43 | "qk_rope_head_dim": 64, |
| 44 | "rms_norm_eps": 1e-05, |
| 45 | "rope_interleave": true, |
| 46 | "rope_parameters": { |
| 47 | "rope_theta": 1000000, |
| 48 | "rope_type": "default" |
| 49 | }, |
| 50 | "routed_scaling_factor": 2.5, |
| 51 | "scoring_func": "sigmoid", |
| 52 | "tie_word_embeddings": false, |
| 53 | "topk_group": 1, |
| 54 | "topk_method": "noaux_tc", |
| 55 | "transformers_version": "5.0.2.dev0", |
| 56 | "use_cache": true, |
| 57 | "v_head_dim": 256, |
| 58 | "vocab_size": 154880, |
| 59 | "quantization_config": { |
| 60 | "activation_scheme": "dynamic", |
| 61 | "fmt": "e4m3", |
| 62 | "quant_method": "fp8", |
| 63 | "weight_block_size": [ |
| 64 | 128, |
| 65 | 128 |
| 66 | ], |
| 67 | "modules_to_not_convert": [ |
| 68 | "lm_head", |
| 69 | "model.embed_tokens", |
| 70 | "model.layers.0.input_layernorm", |
| 71 | "model.layers.0.post_attention_layernorm", |
| 72 | "model.layers.0.self_attn.indexer.k_norm", |
| 73 | "model.layers.0.self_attn.indexer.k_norm.bias", |
| 74 | "model.layers.0.self_attn.indexers_proj", |
| 75 | "model.layers.0.self_attn.kv_a_layernorm", |
| 76 | "model.layers.0.self_attn.q_a_layernorm", |
| 77 | "model.layers.1.input_layernorm", |
| 78 | "model.layers.1.post_attention_layernorm", |
| 79 | "model.layers.1.self_attn.indexer.k_norm", |
| 80 | "model.layers.1.self_attn.indexer.k_norm.bias", |
| 81 | "model.layers.1.self_attn.indexers_proj", |
| 82 | "model.layers.1.self_attn.kv_a_layernorm", |
| 83 | "model.layers.1.self_attn.q_a_layernorm", |
| 84 | "model.layers.2.input_layernorm", |
| 85 | "model.layers.2.post_attention_layernorm", |
| 86 | "model.layers.2.self_attn.indexer.k_norm", |
| 87 | "model.layers.2.self_attn.indexer.k_norm.bias", |
| 88 | "model.layers.2.self_attn.indexers_proj", |
| 89 | "model.layers.2.self_attn.kv_a_layernorm", |
| 90 | "model.layers.2.self_attn.q_a_layernorm", |
| 91 | "model.layers.3.input_layernorm", |
| 92 | "model.layers.3.mlp.gate", |
| 93 | "model.layers.3.mlp.gate.e_score_correction_bias", |
| 94 | "model.layers.3.post_attention_layernorm", |
| 95 | "model.layers.3.self_attn.indexer.k_norm", |
| 96 | "model.layers.3.self_attn.indexer.k_norm.bias", |
| 97 | "model.layers.3.self_attn.indexers_proj", |
| 98 | "model.layers.3.self_attn.kv_a_layernorm", |
| 99 | "model.layers.3.self_attn.q_a_layernorm", |
| 100 | "model.layers.4.input_layernorm", |
| 101 | "model.layers.4.mlp.gate", |
| 102 | "model.layers.4.mlp.gate.e_score_correction_bias", |
| 103 | "model.layers.4.post_attention_layernorm", |
| 104 | "model.layers.4.self_attn.indexer.k_norm", |
| 105 | "model.layers.4.self_attn.indexer.k_norm.bias", |
| 106 | "model.layers.4.self_attn.indexers_proj", |
| 107 | "model.layers.4.self_attn.kv_a_layernorm", |
| 108 | "model.layers.4.self_attn.q_a_layernorm", |
| 109 | "model.layers.5.input_layernorm", |
| 110 | "model.layers.5.mlp.gate", |
| 111 | "model.layers.5.mlp.gate.e_score_correction_bias", |
| 112 | "model.layers.5.post_attention_layernorm", |
| 113 | "model.layers.5.self_attn.indexer.k_norm", |
| 114 | "model.layers.5.self_attn.indexer.k_norm.bias", |
| 115 | "model.layers.5.self_attn.indexers_proj", |
| 116 | "model.layers.5.self_attn.kv_a_layernorm", |
| 117 | "model.layers.5.self_attn.q_a_layernorm", |
| 118 | "model.layers.6.input_layernorm", |
| 119 | "model.layers.6.mlp.gate", |
| 120 | "model.layers.6.mlp.gate.e_score_correction_bias", |
| 121 | "model.layers.6.post_attention_layernorm", |
| 122 | "model.layers.6.self_attn.indexer.k_norm", |
| 123 | "model.layers.6.self_attn.indexer.k_norm.bias", |
| 124 | "model.layers.6.self_attn.indexers_proj", |
| 125 | "model.layers.6.self_attn.kv_a_layernorm", |
| 126 | "model.layers.6.self_attn.q_a_layernorm", |
| 127 | "model.layers.7.input_layernorm", |
| 128 | "model.layers.7.mlp.gate", |
| 129 | "model.layers.7.mlp.gate.e_score_correction_bias", |
| 130 | "model.layers.7.post_attention_layernorm", |
| 131 | "model.layers.7.self_attn.indexer.k_norm", |
| 132 | "model.layers.7.self_attn.indexer.k_norm.bias", |
| 133 | "model.layers.7.self_attn.indexers_proj", |
| 134 | "model.layers.7.self_attn.kv_a_layernorm", |
| 135 | "model.layers.7.self_attn.q_a_layernorm", |
| 136 | "model.layers.8.input_layernorm", |
| 137 | "model.layers.8.mlp.gate", |
| 138 | "model.layers.8.mlp.gate.e_score_correction_bias", |
| 139 | "model.layers.8.post_attention_layernorm", |
| 140 | "model.layers.8.self_attn.indexer.k_norm", |
| 141 | "model.layers.8.self_attn.indexer.k_norm.bias", |
| 142 | "model.layers.8.self_attn.indexers_proj", |
| 143 | "model.layers.8.self_attn.kv_a_layernorm", |
| 144 | "model.layers.8.self_attn.q_a_layernorm", |
| 145 | "model.layers.9.input_layernorm", |
| 146 | "model.layers.9.mlp.gate", |
| 147 | "model.layers.9.mlp.gate.e_score_correction_bias", |
| 148 | "model.layers.9.post_attention_layernorm", |
| 149 | "model.layers.9.self_attn.indexer.k_norm", |
| 150 | "model.layers.9.self_attn.indexer.k_norm.bias", |
| 151 | "model.layers.9.self_attn.indexers_proj", |
| 152 | "model.layers.9.self_attn.kv_a_layernorm", |
| 153 | "model.layers.9.self_attn.q_a_layernorm", |
| 154 | "model.layers.10.input_layernorm", |
| 155 | "model.layers.10.mlp.gate", |
| 156 | "model.layers.10.mlp.gate.e_score_correction_bias", |
| 157 | "model.layers.10.post_attention_layernorm", |
| 158 | "model.layers.10.self_attn.indexer.k_norm", |
| 159 | "model.layers.10.self_attn.indexer.k_norm.bias", |
| 160 | "model.layers.10.self_attn.indexers_proj", |
| 161 | "model.layers.10.self_attn.kv_a_layernorm", |
| 162 | "model.layers.10.self_attn.q_a_layernorm", |
| 163 | "model.layers.11.input_layernorm", |
| 164 | "model.layers.11.mlp.gate", |
| 165 | "model.layers.11.mlp.gate.e_score_correction_bias", |
| 166 | "model.layers.11.post_attention_layernorm", |
| 167 | "model.layers.11.self_attn.indexer.k_norm", |
| 168 | "model.layers.11.self_attn.indexer.k_norm.bias", |
| 169 | "model.layers.11.self_attn.indexers_proj", |
| 170 | "model.layers.11.self_attn.kv_a_layernorm", |
| 171 | "model.layers.11.self_attn.q_a_layernorm", |
| 172 | "model.layers.12.input_layernorm", |
| 173 | "model.layers.12.mlp.gate", |
| 174 | "model.layers.12.mlp.gate.e_score_correction_bias", |
| 175 | "model.layers.12.post_attention_layernorm", |
| 176 | "model.layers.12.self_attn.indexer.k_norm", |
| 177 | "model.layers.12.self_attn.indexer.k_norm.bias", |
| 178 | "model.layers.12.self_attn.indexers_proj", |
| 179 | "model.layers.12.self_attn.kv_a_layernorm", |
| 180 | "model.layers.12.self_attn.q_a_layernorm", |
| 181 | "model.layers.13.input_layernorm", |
| 182 | "model.layers.13.mlp.gate", |
| 183 | "model.layers.13.mlp.gate.e_score_correction_bias", |
| 184 | "model.layers.13.post_attention_layernorm", |
| 185 | "model.layers.13.self_attn.indexer.k_norm", |
| 186 | "model.layers.13.self_attn.indexer.k_norm.bias", |
| 187 | "model.layers.13.self_attn.indexers_proj", |
| 188 | "model.layers.13.self_attn.kv_a_layernorm", |
| 189 | "model.layers.13.self_attn.q_a_layernorm", |
| 190 | "model.layers.14.input_layernorm", |
| 191 | "model.layers.14.mlp.gate", |
| 192 | "model.layers.14.mlp.gate.e_score_correction_bias", |
| 193 | "model.layers.14.post_attention_layernorm", |
| 194 | "model.layers.14.self_attn.indexer.k_norm", |
| 195 | "model.layers.14.self_attn.indexer.k_norm.bias", |
| 196 | "model.layers.14.self_attn.indexers_proj", |
| 197 | "model.layers.14.self_attn.kv_a_layernorm", |
| 198 | "model.layers.14.self_attn.q_a_layernorm", |
| 199 | "model.layers.15.input_layernorm", |
| 200 | "model.layers.15.mlp.gate", |
| 201 | "model.layers.15.mlp.gate.e_score_correction_bias", |
| 202 | "model.layers.15.post_attention_layernorm", |
| 203 | "model.layers.15.self_attn.indexer.k_norm", |
| 204 | "model.layers.15.self_attn.indexer.k_norm.bias", |
| 205 | "model.layers.15.self_attn.indexers_proj", |
| 206 | "model.layers.15.self_attn.kv_a_layernorm", |
| 207 | "model.layers.15.self_attn.q_a_layernorm", |
| 208 | "model.layers.16.input_layernorm", |
| 209 | "model.layers.16.mlp.gate", |
| 210 | "model.layers.16.mlp.gate.e_score_correction_bias", |
| 211 | "model.layers.16.post_attention_layernorm", |
| 212 | "model.layers.16.self_attn.indexer.k_norm", |
| 213 | "model.layers.16.self_attn.indexer.k_norm.bias", |
| 214 | "model.layers.16.self_attn.indexers_proj", |
| 215 | "model.layers.16.self_attn.kv_a_layernorm", |
| 216 | "model.layers.16.self_attn.q_a_layernorm", |
| 217 | "model.layers.17.input_layernorm", |
| 218 | "model.layers.17.mlp.gate", |
| 219 | "model.layers.17.mlp.gate.e_score_correction_bias", |
| 220 | "model.layers.17.post_attention_layernorm", |
| 221 | "model.layers.17.self_attn.indexer.k_norm", |
| 222 | "model.layers.17.self_attn.indexer.k_norm.bias", |
| 223 | "model.layers.17.self_attn.indexers_proj", |
| 224 | "model.layers.17.self_attn.kv_a_layernorm", |
| 225 | "model.layers.17.self_attn.q_a_layernorm", |
| 226 | "model.layers.18.input_layernorm", |
| 227 | "model.layers.18.mlp.gate", |
| 228 | "model.layers.18.mlp.gate.e_score_correction_bias", |
| 229 | "model.layers.18.post_attention_layernorm", |
| 230 | "model.layers.18.self_attn.indexer.k_norm", |
| 231 | "model.layers.18.self_attn.indexer.k_norm.bias", |
| 232 | "model.layers.18.self_attn.indexers_proj", |
| 233 | "model.layers.18.self_attn.kv_a_layernorm", |
| 234 | "model.layers.18.self_attn.q_a_layernorm", |
| 235 | "model.layers.19.input_layernorm", |
| 236 | "model.layers.19.mlp.gate", |
| 237 | "model.layers.19.mlp.gate.e_score_correction_bias", |
| 238 | "model.layers.19.post_attention_layernorm", |
| 239 | "model.layers.19.self_attn.indexer.k_norm", |
| 240 | "model.layers.19.self_attn.indexer.k_norm.bias", |
| 241 | "model.layers.19.self_attn.indexers_proj", |
| 242 | "model.layers.19.self_attn.kv_a_layernorm", |
| 243 | "model.layers.19.self_attn.q_a_layernorm", |
| 244 | "model.layers.20.input_layernorm", |
| 245 | "model.layers.20.mlp.gate", |
| 246 | "model.layers.20.mlp.gate.e_score_correction_bias", |
| 247 | "model.layers.20.post_attention_layernorm", |
| 248 | "model.layers.20.self_attn.indexer.k_norm", |
| 249 | "model.layers.20.self_attn.indexer.k_norm.bias", |
| 250 | "model.layers.20.self_attn.indexers_proj", |
| 251 | "model.layers.20.self_attn.kv_a_layernorm", |
| 252 | "model.layers.20.self_attn.q_a_layernorm", |
| 253 | "model.layers.21.input_layernorm", |
| 254 | "model.layers.21.mlp.gate", |
| 255 | "model.layers.21.mlp.gate.e_score_correction_bias", |
| 256 | "model.layers.21.post_attention_layernorm", |
| 257 | "model.layers.21.self_attn.indexer.k_norm", |
| 258 | "model.layers.21.self_attn.indexer.k_norm.bias", |
| 259 | "model.layers.21.self_attn.indexers_proj", |
| 260 | "model.layers.21.self_attn.kv_a_layernorm", |
| 261 | "model.layers.21.self_attn.q_a_layernorm", |
| 262 | "model.layers.22.input_layernorm", |
| 263 | "model.layers.22.mlp.gate", |
| 264 | "model.layers.22.mlp.gate.e_score_correction_bias", |
| 265 | "model.layers.22.post_attention_layernorm", |
| 266 | "model.layers.22.self_attn.indexer.k_norm", |
| 267 | "model.layers.22.self_attn.indexer.k_norm.bias", |
| 268 | "model.layers.22.self_attn.indexers_proj", |
| 269 | "model.layers.22.self_attn.kv_a_layernorm", |
| 270 | "model.layers.22.self_attn.q_a_layernorm", |
| 271 | "model.layers.23.input_layernorm", |
| 272 | "model.layers.23.mlp.gate", |
| 273 | "model.layers.23.mlp.gate.e_score_correction_bias", |
| 274 | "model.layers.23.post_attention_layernorm", |
| 275 | "model.layers.23.self_attn.indexer.k_norm", |
| 276 | "model.layers.23.self_attn.indexer.k_norm.bias", |
| 277 | "model.layers.23.self_attn.indexers_proj", |
| 278 | "model.layers.23.self_attn.kv_a_layernorm", |
| 279 | "model.layers.23.self_attn.q_a_layernorm", |
| 280 | "model.layers.24.input_layernorm", |
| 281 | "model.layers.24.mlp.gate", |
| 282 | "model.layers.24.mlp.gate.e_score_correction_bias", |
| 283 | "model.layers.24.post_attention_layernorm", |
| 284 | "model.layers.24.self_attn.indexer.k_norm", |
| 285 | "model.layers.24.self_attn.indexer.k_norm.bias", |
| 286 | "model.layers.24.self_attn.indexers_proj", |
| 287 | "model.layers.24.self_attn.kv_a_layernorm", |
| 288 | "model.layers.24.self_attn.q_a_layernorm", |
| 289 | "model.layers.25.input_layernorm", |
| 290 | "model.layers.25.mlp.gate", |
| 291 | "model.layers.25.mlp.gate.e_score_correction_bias", |
| 292 | "model.layers.25.post_attention_layernorm", |
| 293 | "model.layers.25.self_attn.indexer.k_norm", |
| 294 | "model.layers.25.self_attn.indexer.k_norm.bias", |
| 295 | "model.layers.25.self_attn.indexers_proj", |
| 296 | "model.layers.25.self_attn.kv_a_layernorm", |
| 297 | "model.layers.25.self_attn.q_a_layernorm", |
| 298 | "model.layers.26.input_layernorm", |
| 299 | "model.layers.26.mlp.gate", |
| 300 | "model.layers.26.mlp.gate.e_score_correction_bias", |
| 301 | "model.layers.26.post_attention_layernorm", |
| 302 | "model.layers.26.self_attn.indexer.k_norm", |
| 303 | "model.layers.26.self_attn.indexer.k_norm.bias", |
| 304 | "model.layers.26.self_attn.indexers_proj", |
| 305 | "model.layers.26.self_attn.kv_a_layernorm", |
| 306 | "model.layers.26.self_attn.q_a_layernorm", |
| 307 | "model.layers.27.input_layernorm", |
| 308 | "model.layers.27.mlp.gate", |
| 309 | "model.layers.27.mlp.gate.e_score_correction_bias", |
| 310 | "model.layers.27.post_attention_layernorm", |
| 311 | "model.layers.27.self_attn.indexer.k_norm", |
| 312 | "model.layers.27.self_attn.indexer.k_norm.bias", |
| 313 | "model.layers.27.self_attn.indexers_proj", |
| 314 | "model.layers.27.self_attn.kv_a_layernorm", |
| 315 | "model.layers.27.self_attn.q_a_layernorm", |
| 316 | "model.layers.28.input_layernorm", |
| 317 | "model.layers.28.mlp.gate", |
| 318 | "model.layers.28.mlp.gate.e_score_correction_bias", |
| 319 | "model.layers.28.post_attention_layernorm", |
| 320 | "model.layers.28.self_attn.indexer.k_norm", |
| 321 | "model.layers.28.self_attn.indexer.k_norm.bias", |
| 322 | "model.layers.28.self_attn.indexers_proj", |
| 323 | "model.layers.28.self_attn.kv_a_layernorm", |
| 324 | "model.layers.28.self_attn.q_a_layernorm", |
| 325 | "model.layers.29.input_layernorm", |
| 326 | "model.layers.29.mlp.gate", |
| 327 | "model.layers.29.mlp.gate.e_score_correction_bias", |
| 328 | "model.layers.29.post_attention_layernorm", |
| 329 | "model.layers.29.self_attn.indexer.k_norm", |
| 330 | "model.layers.29.self_attn.indexer.k_norm.bias", |
| 331 | "model.layers.29.self_attn.indexers_proj", |
| 332 | "model.layers.29.self_attn.kv_a_layernorm", |
| 333 | "model.layers.29.self_attn.q_a_layernorm", |
| 334 | "model.layers.30.input_layernorm", |
| 335 | "model.layers.30.mlp.gate", |
| 336 | "model.layers.30.mlp.gate.e_score_correction_bias", |
| 337 | "model.layers.30.post_attention_layernorm", |
| 338 | "model.layers.30.self_attn.indexer.k_norm", |
| 339 | "model.layers.30.self_attn.indexer.k_norm.bias", |
| 340 | "model.layers.30.self_attn.indexers_proj", |
| 341 | "model.layers.30.self_attn.kv_a_layernorm", |
| 342 | "model.layers.30.self_attn.q_a_layernorm", |
| 343 | "model.layers.31.input_layernorm", |
| 344 | "model.layers.31.mlp.gate", |
| 345 | "model.layers.31.mlp.gate.e_score_correction_bias", |
| 346 | "model.layers.31.post_attention_layernorm", |
| 347 | "model.layers.31.self_attn.indexer.k_norm", |
| 348 | "model.layers.31.self_attn.indexer.k_norm.bias", |
| 349 | "model.layers.31.self_attn.indexers_proj", |
| 350 | "model.layers.31.self_attn.kv_a_layernorm", |
| 351 | "model.layers.31.self_attn.q_a_layernorm", |
| 352 | "model.layers.32.input_layernorm", |
| 353 | "model.layers.32.mlp.gate", |
| 354 | "model.layers.32.mlp.gate.e_score_correction_bias", |
| 355 | "model.layers.32.post_attention_layernorm", |
| 356 | "model.layers.32.self_attn.indexer.k_norm", |
| 357 | "model.layers.32.self_attn.indexer.k_norm.bias", |
| 358 | "model.layers.32.self_attn.indexers_proj", |
| 359 | "model.layers.32.self_attn.kv_a_layernorm", |
| 360 | "model.layers.32.self_attn.q_a_layernorm", |
| 361 | "model.layers.33.input_layernorm", |
| 362 | "model.layers.33.mlp.gate", |
| 363 | "model.layers.33.mlp.gate.e_score_correction_bias", |
| 364 | "model.layers.33.post_attention_layernorm", |
| 365 | "model.layers.33.self_attn.indexer.k_norm", |
| 366 | "model.layers.33.self_attn.indexer.k_norm.bias", |
| 367 | "model.layers.33.self_attn.indexers_proj", |
| 368 | "model.layers.33.self_attn.kv_a_layernorm", |
| 369 | "model.layers.33.self_attn.q_a_layernorm", |
| 370 | "model.layers.34.input_layernorm", |
| 371 | "model.layers.34.mlp.gate", |
| 372 | "model.layers.34.mlp.gate.e_score_correction_bias", |
| 373 | "model.layers.34.post_attention_layernorm", |
| 374 | "model.layers.34.self_attn.indexer.k_norm", |
| 375 | "model.layers.34.self_attn.indexer.k_norm.bias", |
| 376 | "model.layers.34.self_attn.indexers_proj", |
| 377 | "model.layers.34.self_attn.kv_a_layernorm", |
| 378 | "model.layers.34.self_attn.q_a_layernorm", |
| 379 | "model.layers.35.input_layernorm", |
| 380 | "model.layers.35.mlp.gate", |
| 381 | "model.layers.35.mlp.gate.e_score_correction_bias", |
| 382 | "model.layers.35.post_attention_layernorm", |
| 383 | "model.layers.35.self_attn.indexer.k_norm", |
| 384 | "model.layers.35.self_attn.indexer.k_norm.bias", |
| 385 | "model.layers.35.self_attn.indexers_proj", |
| 386 | "model.layers.35.self_attn.kv_a_layernorm", |
| 387 | "model.layers.35.self_attn.q_a_layernorm", |
| 388 | "model.layers.36.input_layernorm", |
| 389 | "model.layers.36.mlp.gate", |
| 390 | "model.layers.36.mlp.gate.e_score_correction_bias", |
| 391 | "model.layers.36.post_attention_layernorm", |
| 392 | "model.layers.36.self_attn.indexer.k_norm", |
| 393 | "model.layers.36.self_attn.indexer.k_norm.bias", |
| 394 | "model.layers.36.self_attn.indexers_proj", |
| 395 | "model.layers.36.self_attn.kv_a_layernorm", |
| 396 | "model.layers.36.self_attn.q_a_layernorm", |
| 397 | "model.layers.37.input_layernorm", |
| 398 | "model.layers.37.mlp.gate", |
| 399 | "model.layers.37.mlp.gate.e_score_correction_bias", |
| 400 | "model.layers.37.post_attention_layernorm", |
| 401 | "model.layers.37.self_attn.indexer.k_norm", |
| 402 | "model.layers.37.self_attn.indexer.k_norm.bias", |
| 403 | "model.layers.37.self_attn.indexers_proj", |
| 404 | "model.layers.37.self_attn.kv_a_layernorm", |
| 405 | "model.layers.37.self_attn.q_a_layernorm", |
| 406 | "model.layers.38.input_layernorm", |
| 407 | "model.layers.38.mlp.gate", |
| 408 | "model.layers.38.mlp.gate.e_score_correction_bias", |
| 409 | "model.layers.38.post_attention_layernorm", |
| 410 | "model.layers.38.self_attn.indexer.k_norm", |
| 411 | "model.layers.38.self_attn.indexer.k_norm.bias", |
| 412 | "model.layers.38.self_attn.indexers_proj", |
| 413 | "model.layers.38.self_attn.kv_a_layernorm", |
| 414 | "model.layers.38.self_attn.q_a_layernorm", |
| 415 | "model.layers.39.input_layernorm", |
| 416 | "model.layers.39.mlp.gate", |
| 417 | "model.layers.39.mlp.gate.e_score_correction_bias", |
| 418 | "model.layers.39.post_attention_layernorm", |
| 419 | "model.layers.39.self_attn.indexer.k_norm", |
| 420 | "model.layers.39.self_attn.indexer.k_norm.bias", |
| 421 | "model.layers.39.self_attn.indexers_proj", |
| 422 | "model.layers.39.self_attn.kv_a_layernorm", |
| 423 | "model.layers.39.self_attn.q_a_layernorm", |
| 424 | "model.layers.40.input_layernorm", |
| 425 | "model.layers.40.mlp.gate", |
| 426 | "model.layers.40.mlp.gate.e_score_correction_bias", |
| 427 | "model.layers.40.post_attention_layernorm", |
| 428 | "model.layers.40.self_attn.indexer.k_norm", |
| 429 | "model.layers.40.self_attn.indexer.k_norm.bias", |
| 430 | "model.layers.40.self_attn.indexers_proj", |
| 431 | "model.layers.40.self_attn.kv_a_layernorm", |
| 432 | "model.layers.40.self_attn.q_a_layernorm", |
| 433 | "model.layers.41.input_layernorm", |
| 434 | "model.layers.41.mlp.gate", |
| 435 | "model.layers.41.mlp.gate.e_score_correction_bias", |
| 436 | "model.layers.41.post_attention_layernorm", |
| 437 | "model.layers.41.self_attn.indexer.k_norm", |
| 438 | "model.layers.41.self_attn.indexer.k_norm.bias", |
| 439 | "model.layers.41.self_attn.indexers_proj", |
| 440 | "model.layers.41.self_attn.kv_a_layernorm", |
| 441 | "model.layers.41.self_attn.q_a_layernorm", |
| 442 | "model.layers.42.input_layernorm", |
| 443 | "model.layers.42.mlp.gate", |
| 444 | "model.layers.42.mlp.gate.e_score_correction_bias", |
| 445 | "model.layers.42.post_attention_layernorm", |
| 446 | "model.layers.42.self_attn.indexer.k_norm", |
| 447 | "model.layers.42.self_attn.indexer.k_norm.bias", |
| 448 | "model.layers.42.self_attn.indexers_proj", |
| 449 | "model.layers.42.self_attn.kv_a_layernorm", |
| 450 | "model.layers.42.self_attn.q_a_layernorm", |
| 451 | "model.layers.43.input_layernorm", |
| 452 | "model.layers.43.mlp.gate", |
| 453 | "model.layers.43.mlp.gate.e_score_correction_bias", |
| 454 | "model.layers.43.post_attention_layernorm", |
| 455 | "model.layers.43.self_attn.indexer.k_norm", |
| 456 | "model.layers.43.self_attn.indexer.k_norm.bias", |
| 457 | "model.layers.43.self_attn.indexers_proj", |
| 458 | "model.layers.43.self_attn.kv_a_layernorm", |
| 459 | "model.layers.43.self_attn.q_a_layernorm", |
| 460 | "model.layers.44.input_layernorm", |
| 461 | "model.layers.44.mlp.gate", |
| 462 | "model.layers.44.mlp.gate.e_score_correction_bias", |
| 463 | "model.layers.44.post_attention_layernorm", |
| 464 | "model.layers.44.self_attn.indexer.k_norm", |
| 465 | "model.layers.44.self_attn.indexer.k_norm.bias", |
| 466 | "model.layers.44.self_attn.indexers_proj", |
| 467 | "model.layers.44.self_attn.kv_a_layernorm", |
| 468 | "model.layers.44.self_attn.q_a_layernorm", |
| 469 | "model.layers.45.input_layernorm", |
| 470 | "model.layers.45.mlp.gate", |
| 471 | "model.layers.45.mlp.gate.e_score_correction_bias", |
| 472 | "model.layers.45.post_attention_layernorm", |
| 473 | "model.layers.45.self_attn.indexer.k_norm", |
| 474 | "model.layers.45.self_attn.indexer.k_norm.bias", |
| 475 | "model.layers.45.self_attn.indexers_proj", |
| 476 | "model.layers.45.self_attn.kv_a_layernorm", |
| 477 | "model.layers.45.self_attn.q_a_layernorm", |
| 478 | "model.layers.46.input_layernorm", |
| 479 | "model.layers.46.mlp.gate", |
| 480 | "model.layers.46.mlp.gate.e_score_correction_bias", |
| 481 | "model.layers.46.post_attention_layernorm", |
| 482 | "model.layers.46.self_attn.indexer.k_norm", |
| 483 | "model.layers.46.self_attn.indexer.k_norm.bias", |
| 484 | "model.layers.46.self_attn.indexers_proj", |
| 485 | "model.layers.46.self_attn.kv_a_layernorm", |
| 486 | "model.layers.46.self_attn.q_a_layernorm", |
| 487 | "model.layers.47.input_layernorm", |
| 488 | "model.layers.47.mlp.gate", |
| 489 | "model.layers.47.mlp.gate.e_score_correction_bias", |
| 490 | "model.layers.47.post_attention_layernorm", |
| 491 | "model.layers.47.self_attn.indexer.k_norm", |
| 492 | "model.layers.47.self_attn.indexer.k_norm.bias", |
| 493 | "model.layers.47.self_attn.indexers_proj", |
| 494 | "model.layers.47.self_attn.kv_a_layernorm", |
| 495 | "model.layers.47.self_attn.q_a_layernorm", |
| 496 | "model.layers.48.input_layernorm", |
| 497 | "model.layers.48.mlp.gate", |
| 498 | "model.layers.48.mlp.gate.e_score_correction_bias", |
| 499 | "model.layers.48.post_attention_layernorm", |
| 500 | "model.layers.48.self_attn.indexer.k_norm", |
| 501 | "model.layers.48.self_attn.indexer.k_norm.bias", |
| 502 | "model.layers.48.self_attn.indexers_proj", |
| 503 | "model.layers.48.self_attn.kv_a_layernorm", |
| 504 | "model.layers.48.self_attn.q_a_layernorm", |
| 505 | "model.layers.49.input_layernorm", |
| 506 | "model.layers.49.mlp.gate", |
| 507 | "model.layers.49.mlp.gate.e_score_correction_bias", |
| 508 | "model.layers.49.post_attention_layernorm", |
| 509 | "model.layers.49.self_attn.indexer.k_norm", |
| 510 | "model.layers.49.self_attn.indexer.k_norm.bias", |
| 511 | "model.layers.49.self_attn.indexers_proj", |
| 512 | "model.layers.49.self_attn.kv_a_layernorm", |
| 513 | "model.layers.49.self_attn.q_a_layernorm", |
| 514 | "model.layers.50.input_layernorm", |
| 515 | "model.layers.50.mlp.gate", |
| 516 | "model.layers.50.mlp.gate.e_score_correction_bias", |
| 517 | "model.layers.50.post_attention_layernorm", |
| 518 | "model.layers.50.self_attn.indexer.k_norm", |
| 519 | "model.layers.50.self_attn.indexer.k_norm.bias", |
| 520 | "model.layers.50.self_attn.indexers_proj", |
| 521 | "model.layers.50.self_attn.kv_a_layernorm", |
| 522 | "model.layers.50.self_attn.q_a_layernorm", |
| 523 | "model.layers.51.input_layernorm", |
| 524 | "model.layers.51.mlp.gate", |
| 525 | "model.layers.51.mlp.gate.e_score_correction_bias", |
| 526 | "model.layers.51.post_attention_layernorm", |
| 527 | "model.layers.51.self_attn.indexer.k_norm", |
| 528 | "model.layers.51.self_attn.indexer.k_norm.bias", |
| 529 | "model.layers.51.self_attn.indexers_proj", |
| 530 | "model.layers.51.self_attn.kv_a_layernorm", |
| 531 | "model.layers.51.self_attn.q_a_layernorm", |
| 532 | "model.layers.52.input_layernorm", |
| 533 | "model.layers.52.mlp.gate", |
| 534 | "model.layers.52.mlp.gate.e_score_correction_bias", |
| 535 | "model.layers.52.post_attention_layernorm", |
| 536 | "model.layers.52.self_attn.indexer.k_norm", |
| 537 | "model.layers.52.self_attn.indexer.k_norm.bias", |
| 538 | "model.layers.52.self_attn.indexers_proj", |
| 539 | "model.layers.52.self_attn.kv_a_layernorm", |
| 540 | "model.layers.52.self_attn.q_a_layernorm", |
| 541 | "model.layers.53.input_layernorm", |
| 542 | "model.layers.53.mlp.gate", |
| 543 | "model.layers.53.mlp.gate.e_score_correction_bias", |
| 544 | "model.layers.53.post_attention_layernorm", |
| 545 | "model.layers.53.self_attn.indexer.k_norm", |
| 546 | "model.layers.53.self_attn.indexer.k_norm.bias", |
| 547 | "model.layers.53.self_attn.indexers_proj", |
| 548 | "model.layers.53.self_attn.kv_a_layernorm", |
| 549 | "model.layers.53.self_attn.q_a_layernorm", |
| 550 | "model.layers.54.input_layernorm", |
| 551 | "model.layers.54.mlp.gate", |
| 552 | "model.layers.54.mlp.gate.e_score_correction_bias", |
| 553 | "model.layers.54.post_attention_layernorm", |
| 554 | "model.layers.54.self_attn.indexer.k_norm", |
| 555 | "model.layers.54.self_attn.indexer.k_norm.bias", |
| 556 | "model.layers.54.self_attn.indexers_proj", |
| 557 | "model.layers.54.self_attn.kv_a_layernorm", |
| 558 | "model.layers.54.self_attn.q_a_layernorm", |
| 559 | "model.layers.55.input_layernorm", |
| 560 | "model.layers.55.mlp.gate", |
| 561 | "model.layers.55.mlp.gate.e_score_correction_bias", |
| 562 | "model.layers.55.post_attention_layernorm", |
| 563 | "model.layers.55.self_attn.indexer.k_norm", |
| 564 | "model.layers.55.self_attn.indexer.k_norm.bias", |
| 565 | "model.layers.55.self_attn.indexers_proj", |
| 566 | "model.layers.55.self_attn.kv_a_layernorm", |
| 567 | "model.layers.55.self_attn.q_a_layernorm", |
| 568 | "model.layers.56.input_layernorm", |
| 569 | "model.layers.56.mlp.gate", |
| 570 | "model.layers.56.mlp.gate.e_score_correction_bias", |
| 571 | "model.layers.56.post_attention_layernorm", |
| 572 | "model.layers.56.self_attn.indexer.k_norm", |
| 573 | "model.layers.56.self_attn.indexer.k_norm.bias", |
| 574 | "model.layers.56.self_attn.indexers_proj", |
| 575 | "model.layers.56.self_attn.kv_a_layernorm", |
| 576 | "model.layers.56.self_attn.q_a_layernorm", |
| 577 | "model.layers.57.input_layernorm", |
| 578 | "model.layers.57.mlp.gate", |
| 579 | "model.layers.57.mlp.gate.e_score_correction_bias", |
| 580 | "model.layers.57.post_attention_layernorm", |
| 581 | "model.layers.57.self_attn.indexer.k_norm", |
| 582 | "model.layers.57.self_attn.indexer.k_norm.bias", |
| 583 | "model.layers.57.self_attn.indexers_proj", |
| 584 | "model.layers.57.self_attn.kv_a_layernorm", |
| 585 | "model.layers.57.self_attn.q_a_layernorm", |
| 586 | "model.layers.58.input_layernorm", |
| 587 | "model.layers.58.mlp.gate", |
| 588 | "model.layers.58.mlp.gate.e_score_correction_bias", |
| 589 | "model.layers.58.post_attention_layernorm", |
| 590 | "model.layers.58.self_attn.indexer.k_norm", |
| 591 | "model.layers.58.self_attn.indexer.k_norm.bias", |
| 592 | "model.layers.58.self_attn.indexers_proj", |
| 593 | "model.layers.58.self_attn.kv_a_layernorm", |
| 594 | "model.layers.58.self_attn.q_a_layernorm", |
| 595 | "model.layers.59.input_layernorm", |
| 596 | "model.layers.59.mlp.gate", |
| 597 | "model.layers.59.mlp.gate.e_score_correction_bias", |
| 598 | "model.layers.59.post_attention_layernorm", |
| 599 | "model.layers.59.self_attn.indexer.k_norm", |
| 600 | "model.layers.59.self_attn.indexer.k_norm.bias", |
| 601 | "model.layers.59.self_attn.indexers_proj", |
| 602 | "model.layers.59.self_attn.kv_a_layernorm", |
| 603 | "model.layers.59.self_attn.q_a_layernorm", |
| 604 | "model.layers.60.input_layernorm", |
| 605 | "model.layers.60.mlp.gate", |
| 606 | "model.layers.60.mlp.gate.e_score_correction_bias", |
| 607 | "model.layers.60.post_attention_layernorm", |
| 608 | "model.layers.60.self_attn.indexer.k_norm", |
| 609 | "model.layers.60.self_attn.indexer.k_norm.bias", |
| 610 | "model.layers.60.self_attn.indexers_proj", |
| 611 | "model.layers.60.self_attn.kv_a_layernorm", |
| 612 | "model.layers.60.self_attn.q_a_layernorm", |
| 613 | "model.layers.61.input_layernorm", |
| 614 | "model.layers.61.mlp.gate", |
| 615 | "model.layers.61.mlp.gate.e_score_correction_bias", |
| 616 | "model.layers.61.post_attention_layernorm", |
| 617 | "model.layers.61.self_attn.indexer.k_norm", |
| 618 | "model.layers.61.self_attn.indexer.k_norm.bias", |
| 619 | "model.layers.61.self_attn.indexers_proj", |
| 620 | "model.layers.61.self_attn.kv_a_layernorm", |
| 621 | "model.layers.61.self_attn.q_a_layernorm", |
| 622 | "model.layers.62.input_layernorm", |
| 623 | "model.layers.62.mlp.gate", |
| 624 | "model.layers.62.mlp.gate.e_score_correction_bias", |
| 625 | "model.layers.62.post_attention_layernorm", |
| 626 | "model.layers.62.self_attn.indexer.k_norm", |
| 627 | "model.layers.62.self_attn.indexer.k_norm.bias", |
| 628 | "model.layers.62.self_attn.indexers_proj", |
| 629 | "model.layers.62.self_attn.kv_a_layernorm", |
| 630 | "model.layers.62.self_attn.q_a_layernorm", |
| 631 | "model.layers.63.input_layernorm", |
| 632 | "model.layers.63.mlp.gate", |
| 633 | "model.layers.63.mlp.gate.e_score_correction_bias", |
| 634 | "model.layers.63.post_attention_layernorm", |
| 635 | "model.layers.63.self_attn.indexer.k_norm", |
| 636 | "model.layers.63.self_attn.indexer.k_norm.bias", |
| 637 | "model.layers.63.self_attn.indexers_proj", |
| 638 | "model.layers.63.self_attn.kv_a_layernorm", |
| 639 | "model.layers.63.self_attn.q_a_layernorm", |
| 640 | "model.layers.64.input_layernorm", |
| 641 | "model.layers.64.mlp.gate", |
| 642 | "model.layers.64.mlp.gate.e_score_correction_bias", |
| 643 | "model.layers.64.post_attention_layernorm", |
| 644 | "model.layers.64.self_attn.indexer.k_norm", |
| 645 | "model.layers.64.self_attn.indexer.k_norm.bias", |
| 646 | "model.layers.64.self_attn.indexers_proj", |
| 647 | "model.layers.64.self_attn.kv_a_layernorm", |
| 648 | "model.layers.64.self_attn.q_a_layernorm", |
| 649 | "model.layers.65.input_layernorm", |
| 650 | "model.layers.65.mlp.gate", |
| 651 | "model.layers.65.mlp.gate.e_score_correction_bias", |
| 652 | "model.layers.65.post_attention_layernorm", |
| 653 | "model.layers.65.self_attn.indexer.k_norm", |
| 654 | "model.layers.65.self_attn.indexer.k_norm.bias", |
| 655 | "model.layers.65.self_attn.indexers_proj", |
| 656 | "model.layers.65.self_attn.kv_a_layernorm", |
| 657 | "model.layers.65.self_attn.q_a_layernorm", |
| 658 | "model.layers.66.input_layernorm", |
| 659 | "model.layers.66.mlp.gate", |
| 660 | "model.layers.66.mlp.gate.e_score_correction_bias", |
| 661 | "model.layers.66.post_attention_layernorm", |
| 662 | "model.layers.66.self_attn.indexer.k_norm", |
| 663 | "model.layers.66.self_attn.indexer.k_norm.bias", |
| 664 | "model.layers.66.self_attn.indexers_proj", |
| 665 | "model.layers.66.self_attn.kv_a_layernorm", |
| 666 | "model.layers.66.self_attn.q_a_layernorm", |
| 667 | "model.layers.67.input_layernorm", |
| 668 | "model.layers.67.mlp.gate", |
| 669 | "model.layers.67.mlp.gate.e_score_correction_bias", |
| 670 | "model.layers.67.post_attention_layernorm", |
| 671 | "model.layers.67.self_attn.indexer.k_norm", |
| 672 | "model.layers.67.self_attn.indexer.k_norm.bias", |
| 673 | "model.layers.67.self_attn.indexers_proj", |
| 674 | "model.layers.67.self_attn.kv_a_layernorm", |
| 675 | "model.layers.67.self_attn.q_a_layernorm", |
| 676 | "model.layers.68.input_layernorm", |
| 677 | "model.layers.68.mlp.gate", |
| 678 | "model.layers.68.mlp.gate.e_score_correction_bias", |
| 679 | "model.layers.68.post_attention_layernorm", |
| 680 | "model.layers.68.self_attn.indexer.k_norm", |
| 681 | "model.layers.68.self_attn.indexer.k_norm.bias", |
| 682 | "model.layers.68.self_attn.indexers_proj", |
| 683 | "model.layers.68.self_attn.kv_a_layernorm", |
| 684 | "model.layers.68.self_attn.q_a_layernorm", |
| 685 | "model.layers.69.input_layernorm", |
| 686 | "model.layers.69.mlp.gate", |
| 687 | "model.layers.69.mlp.gate.e_score_correction_bias", |
| 688 | "model.layers.69.post_attention_layernorm", |
| 689 | "model.layers.69.self_attn.indexer.k_norm", |
| 690 | "model.layers.69.self_attn.indexer.k_norm.bias", |
| 691 | "model.layers.69.self_attn.indexers_proj", |
| 692 | "model.layers.69.self_attn.kv_a_layernorm", |
| 693 | "model.layers.69.self_attn.q_a_layernorm", |
| 694 | "model.layers.70.input_layernorm", |
| 695 | "model.layers.70.mlp.gate", |
| 696 | "model.layers.70.mlp.gate.e_score_correction_bias", |
| 697 | "model.layers.70.post_attention_layernorm", |
| 698 | "model.layers.70.self_attn.indexer.k_norm", |
| 699 | "model.layers.70.self_attn.indexer.k_norm.bias", |
| 700 | "model.layers.70.self_attn.indexers_proj", |
| 701 | "model.layers.70.self_attn.kv_a_layernorm", |
| 702 | "model.layers.70.self_attn.q_a_layernorm", |
| 703 | "model.layers.71.input_layernorm", |
| 704 | "model.layers.71.mlp.gate", |
| 705 | "model.layers.71.mlp.gate.e_score_correction_bias", |
| 706 | "model.layers.71.post_attention_layernorm", |
| 707 | "model.layers.71.self_attn.indexer.k_norm", |
| 708 | "model.layers.71.self_attn.indexer.k_norm.bias", |
| 709 | "model.layers.71.self_attn.indexers_proj", |
| 710 | "model.layers.71.self_attn.kv_a_layernorm", |
| 711 | "model.layers.71.self_attn.q_a_layernorm", |
| 712 | "model.layers.72.input_layernorm", |
| 713 | "model.layers.72.mlp.gate", |
| 714 | "model.layers.72.mlp.gate.e_score_correction_bias", |
| 715 | "model.layers.72.post_attention_layernorm", |
| 716 | "model.layers.72.self_attn.indexer.k_norm", |
| 717 | "model.layers.72.self_attn.indexer.k_norm.bias", |
| 718 | "model.layers.72.self_attn.indexers_proj", |
| 719 | "model.layers.72.self_attn.kv_a_layernorm", |
| 720 | "model.layers.72.self_attn.q_a_layernorm", |
| 721 | "model.layers.73.input_layernorm", |
| 722 | "model.layers.73.mlp.gate", |
| 723 | "model.layers.73.mlp.gate.e_score_correction_bias", |
| 724 | "model.layers.73.post_attention_layernorm", |
| 725 | "model.layers.73.self_attn.indexer.k_norm", |
| 726 | "model.layers.73.self_attn.indexer.k_norm.bias", |
| 727 | "model.layers.73.self_attn.indexers_proj", |
| 728 | "model.layers.73.self_attn.kv_a_layernorm", |
| 729 | "model.layers.73.self_attn.q_a_layernorm", |
| 730 | "model.layers.74.input_layernorm", |
| 731 | "model.layers.74.mlp.gate", |
| 732 | "model.layers.74.mlp.gate.e_score_correction_bias", |
| 733 | "model.layers.74.post_attention_layernorm", |
| 734 | "model.layers.74.self_attn.indexer.k_norm", |
| 735 | "model.layers.74.self_attn.indexer.k_norm.bias", |
| 736 | "model.layers.74.self_attn.indexers_proj", |
| 737 | "model.layers.74.self_attn.kv_a_layernorm", |
| 738 | "model.layers.74.self_attn.q_a_layernorm", |
| 739 | "model.layers.75.input_layernorm", |
| 740 | "model.layers.75.mlp.gate", |
| 741 | "model.layers.75.mlp.gate.e_score_correction_bias", |
| 742 | "model.layers.75.post_attention_layernorm", |
| 743 | "model.layers.75.self_attn.indexer.k_norm", |
| 744 | "model.layers.75.self_attn.indexer.k_norm.bias", |
| 745 | "model.layers.75.self_attn.indexers_proj", |
| 746 | "model.layers.75.self_attn.kv_a_layernorm", |
| 747 | "model.layers.75.self_attn.q_a_layernorm", |
| 748 | "model.layers.76.input_layernorm", |
| 749 | "model.layers.76.mlp.gate", |
| 750 | "model.layers.76.mlp.gate.e_score_correction_bias", |
| 751 | "model.layers.76.post_attention_layernorm", |
| 752 | "model.layers.76.self_attn.indexer.k_norm", |
| 753 | "model.layers.76.self_attn.indexer.k_norm.bias", |
| 754 | "model.layers.76.self_attn.indexers_proj", |
| 755 | "model.layers.76.self_attn.kv_a_layernorm", |
| 756 | "model.layers.76.self_attn.q_a_layernorm", |
| 757 | "model.layers.77.input_layernorm", |
| 758 | "model.layers.77.mlp.gate", |
| 759 | "model.layers.77.mlp.gate.e_score_correction_bias", |
| 760 | "model.layers.77.post_attention_layernorm", |
| 761 | "model.layers.77.self_attn.indexer.k_norm", |
| 762 | "model.layers.77.self_attn.indexer.k_norm.bias", |
| 763 | "model.layers.77.self_attn.indexers_proj", |
| 764 | "model.layers.77.self_attn.kv_a_layernorm", |
| 765 | "model.layers.77.self_attn.q_a_layernorm", |
| 766 | "model.layers.78.eh_proj", |
| 767 | "model.layers.78.enorm", |
| 768 | "model.layers.78.hnorm", |
| 769 | "model.layers.78.input_layernorm", |
| 770 | "model.layers.78.mlp.gate", |
| 771 | "model.layers.78.mlp.gate.e_score_correction_bias", |
| 772 | "model.layers.78.post_attention_layernorm", |
| 773 | "model.layers.78.self_attn.indexer.k_norm", |
| 774 | "model.layers.78.self_attn.indexer.k_norm.bias", |
| 775 | "model.layers.78.self_attn.indexers_proj", |
| 776 | "model.layers.78.self_attn.kv_a_layernorm", |
| 777 | "model.layers.78.self_attn.q_a_layernorm", |
| 778 | "model.layers.78.shared_head.norm", |
| 779 | "model.norm" |
| 780 | ] |
| 781 | } |
| 782 | } |
| 783 | |