Skip to content

P
Paddle

BaiXuePrincess / Paddle
与 Fork 源项目一致

Fork自 PaddlePaddle / Paddle

通知 1
Star 1
Fork 0
P
Paddle
切换分支/标签
查找文件 普通视图历史永久链接Permalink
arguments_cn.md 12.2 KB
Newer Older
L
add translation for cmd_parameter  
lifu 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 
# 参数概述

虽然Paddle看起来包含了众多参数，但是大部分参数是为开发者提供的，或者已经在集群提交环境中自动设置，因此用户并不需要关心它们。在此，根据这些参数的使用场合，我们将它们划分为不同的类别。例如，`通用`类别中的参数可用于所有场合。某些参数只可用于特定的层中，而有些参数需要在集群多机训练中使用等。

<html>
<table border="2" frame="border">
<thead>
<tr>
<th scope="col" class="left"></th>
<th scope="col" class="left">参数</th>
<th scope="col" class="left">本地训练</th>
<th scope="col" class="left">集群训练</th>
<th scope="col" class="left">本地测试</th>
<th scope="col" class="left">集群测试</th>
</tr>
</thead>

<tbody>
<tr>
<td class="left" rowspan="9">通用</td>
<td class="left">job</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">use_gpu</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">local</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">config</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">config_args</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">num_passes</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">trainer_count</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">version</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">show_layer_stat</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left" rowspan="15">训练</td><td class="left">dot_period</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">test_period</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">saving_period</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">show_parameter_stats_period</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">init_model_path</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left"></td>
</tr>

<tr>
<td class="left">load_missing_parameter_strategy</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">saving_period_by_batches</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">use_old_updater</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">enable_grad_share</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">grad_share_block_num</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">log_error_clipping</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">log_clipping</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">save_only_one</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">start_pass</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">训练/测试</td><td class="left">save_dir</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left" rowspan = "2">训练过程中测试</td><td class="left">test_period</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">average_test_period</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left" rowspan = "5">测试</td><td class="left">model_list</td>
<td class="left"></td><td class="left"></td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">test_wait</td>
<td class="left"></td><td class="left"></td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">test_pass</td>
<td class="left"></td><td class="left"></td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">predict_output_dir</td>
<td class="left"></td><td class="left"></td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">distribute_test</td>
<td class="left"></td><td class="left"></td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">Auc/正负对验证(PnpairValidation)</td><td class="left">predict_file</td>
<td class="left"></td><td class="left"></td><td class="left"></td>√<td class="left">√</td>
</tr>

<tr>
<td class="left" rowspan = "6">GPU</td><td class="left">gpu_id</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">parallel_nn</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">allow_only_one_model_on_one_gpu</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">cudnn_dir</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">cuda_dir</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">cudnn_conv_workspace_limit_in_mb</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left" rowspan = "4">递归神经网络(RNN)</td>
<td class="left">beam_size</td>
<td class="left"></td><td class="left"></td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">rnn_use_batch</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
<td class="left">prev_batch_state</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">diy_beam_search_prob_so</td>
<td class="left"></td><td class="left"></td><td class="left">√</td><td class="left">√</td>
</tr>

<tr>
C
Update arguments_cn.md  
CrossLee1 已提交
232 
<td class="left" rowspan = "16">参数服务器(PServer)</td><td class="left">start_pserver</td>
L
add translation for cmd_parameter  
lifu 已提交
233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left">√</td>
</tr>

<tr>
<td class="left">pservers</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left">√</td>
</tr>

<tr>
<td class="left">port</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left">√</td>
</tr>

<tr>
<td class="left">port_num</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left">√</td>
</tr>

<tr>
<td class="left">ports_num_for_sparse</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left">√</td>
</tr>

<tr>
<td class="left">nics</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left">√</td>
</tr>

<tr>
<td class="left">rdma_tcp</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left">√</td>
</tr>

<tr>
<td class="left">small_messages</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">loadsave_parameters_in_pserver</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left">√</td>
</tr>

<tr>
<td class="left">log_period_server</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">pserver_num_threads</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">sock_send_buf_size</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">sock_recv_buf_size</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">num_gradient_servers</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">parameter_block_size</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">parameter_block_size_for_sparse</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left" rowspan = "3">异步随机梯度下降(Async SGD)</td><td class="left">async_count</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">async_lagged_ratio_min</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">async_lagged_ratio_default</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left" rowspan = "8">性能调优(Performance Tuning)</td><td class="left">log_barrier_abstract</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">log_barrier_lowest_nodes</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">log_barrier_show_log</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">check_sparse_distribution_batches</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">check_sparse_distribution_ratio</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">check_sparse_distribution_unbalance_degree</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">check_sparse_distribution_in_pserver</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">show_check_sparse_distribution_log</td>
<td class="left"></td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
C
Update arguments_cn.md  
CrossLee1 已提交
367 
<td class="left">数据提供器(Data Provider)</td><td class="left">memory_threshold_on_load_data</td>
L
add translation for cmd_parameter  
lifu 已提交
368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left" rowspan = "2">随机数</td><td class="left">seed</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">thread_local_rand_use_global_seed</td>
<td class="left">√</td><td class="left">√</td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">单元测试</td><td class="left">checkgrad_eps</td>
<td class="left"></td><td class="left"></td><td class="left"></td><td class="left"></td>
</tr>

<tr>
<td class="left">矩阵/向量</td><td class="left">enable_parallel_vector</td>
<td class="left">√</td><td class="left">√</td><td class="left">√</td><td class="left">√</td>
</tr>

</tbody>

</table>
</html>