ud-hungarian.master-spec 4.5 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214
  1. component {
  2. name: "rl_rnn"
  3. transition_system {
  4. registered_name: "shift-only"
  5. parameters {
  6. key: "left-to-right"
  7. value: "false"
  8. }
  9. parameters {
  10. key: "parser_skip_deterministic"
  11. value: "false"
  12. }
  13. }
  14. resource {
  15. name: "char-ngram-map"
  16. part {
  17. file_pattern: "TOPDIR/ud-hungarian.char-ngram-map"
  18. file_format: "text"
  19. record_format: ""
  20. }
  21. }
  22. resource {
  23. name: "word-map"
  24. part {
  25. file_pattern: "TOPDIR/ud-hungarian.word-map"
  26. file_format: "text"
  27. record_format: ""
  28. }
  29. }
  30. resource {
  31. name: "label-map"
  32. part {
  33. file_pattern: "TOPDIR/ud-hungarian.label-map"
  34. file_format: "text"
  35. record_format: ""
  36. }
  37. }
  38. fixed_feature {
  39. name: "char_ngram"
  40. fml: "input.token.char-ngram"
  41. embedding_dim: 16
  42. vocabulary_size: 9943
  43. size: 1
  44. }
  45. fixed_feature {
  46. name: "other"
  47. fml: "input.token {digit hyphen punctuation-amount quote }"
  48. embedding_dim: 8
  49. vocabulary_size: 5
  50. size: 4
  51. }
  52. fixed_feature {
  53. name: "words"
  54. fml: "input.word"
  55. embedding_dim: 64
  56. vocabulary_size: 11090
  57. size: 1
  58. }
  59. network_unit {
  60. registered_name: "wrapped_units.LayerNormBasicLSTMNetwork"
  61. parameters {
  62. key: "hidden_layer_sizes"
  63. value: "256"
  64. }
  65. }
  66. component_builder {
  67. registered_name: 'DynamicComponentBuilder'
  68. }
  69. backend {
  70. registered_name: "SyntaxNetComponent"
  71. }
  72. num_actions: 1
  73. attention_component: ""
  74. }
  75. component {
  76. name: "tagger"
  77. transition_system {
  78. registered_name: "tagger"
  79. parameters {
  80. key: "join_category_to_pos"
  81. value: "true"
  82. }
  83. parameters {
  84. key: "parser_skip_deterministic"
  85. value: "false"
  86. }
  87. }
  88. resource {
  89. name: "tag-map"
  90. part {
  91. file_pattern: "TOPDIR/ud-hungarian.tag-map"
  92. file_format: "text"
  93. record_format: ""
  94. }
  95. }
  96. resource {
  97. name: "label-map"
  98. part {
  99. file_pattern: "TOPDIR/ud-hungarian.label-map"
  100. file_format: "text"
  101. record_format: ""
  102. }
  103. }
  104. fixed_feature {
  105. name: "action"
  106. fml: "last-action"
  107. embedding_dim: 32
  108. vocabulary_size: 100
  109. size: 1
  110. }
  111. linked_feature {
  112. name: "encoder"
  113. fml: "input.focus"
  114. embedding_dim: 64
  115. size: 1
  116. source_component: "rl_rnn"
  117. source_translator: "reverse-token"
  118. source_layer: "state_h_0"
  119. }
  120. network_unit {
  121. registered_name: "wrapped_units.LayerNormBasicLSTMNetwork"
  122. parameters {
  123. key: "hidden_layer_sizes"
  124. value: "256"
  125. }
  126. }
  127. component_builder {
  128. registered_name: 'DynamicComponentBuilder'
  129. }
  130. backend {
  131. registered_name: "SyntaxNetComponent"
  132. }
  133. num_actions: 642
  134. attention_component: ""
  135. }
  136. component {
  137. name: "parser"
  138. transition_system {
  139. registered_name: "arc-standard"
  140. parameters {
  141. key: "parser_skip_deterministic"
  142. value: "false"
  143. }
  144. }
  145. resource {
  146. name: "label-map"
  147. part {
  148. file_pattern: "TOPDIR/ud-hungarian.label-map"
  149. file_format: "text"
  150. record_format: ""
  151. }
  152. }
  153. fixed_feature {
  154. name: "action"
  155. fml: "last-action"
  156. embedding_dim: 32
  157. vocabulary_size: 100
  158. size: 1
  159. }
  160. fixed_feature {
  161. name: "labels"
  162. fml: "stack.child(1).label stack.child(1).sibling(-1).label stack.child(-1).label stack.child(-1).sibling(1).label stack(1).child(1).label stack(1).child(1).sibling(-1).label stack(1).child(-1).label stack(1).child(-1).sibling(1).label stack.child(2).label stack.child(-2).label stack(1).child(2).label stack(1).child(-2).label"
  163. embedding_dim: 16
  164. vocabulary_size: 57
  165. size: 12
  166. }
  167. linked_feature {
  168. name: "encoder"
  169. fml: "input.focus"
  170. embedding_dim: 64
  171. size: 1
  172. source_component: "rl_rnn"
  173. source_translator: "reverse-token"
  174. source_layer: "state_h_0"
  175. }
  176. linked_feature {
  177. name: "parser-rnn"
  178. fml: "stack.focus stack(1).focus"
  179. embedding_dim: 64
  180. size: 2
  181. source_component: "parser"
  182. source_translator: "shift-reduce-step"
  183. source_layer: "layer_0"
  184. }
  185. linked_feature {
  186. name: "tagger"
  187. fml: "input.focus stack.focus stack(1).focus"
  188. embedding_dim: 64
  189. size: 3
  190. source_component: "tagger"
  191. source_translator: "identity"
  192. source_layer: "state_h_0"
  193. }
  194. network_unit {
  195. registered_name: 'FeedForwardNetwork'
  196. parameters {
  197. key: "hidden_layer_sizes"
  198. value: "256,256"
  199. }
  200. parameters {
  201. key: "layer_norm_hidden"
  202. value: "True"
  203. }
  204. }
  205. component_builder {
  206. registered_name: 'DynamicComponentBuilder'
  207. }
  208. backend {
  209. registered_name: "SyntaxNetComponent"
  210. }
  211. num_actions: 109
  212. attention_component: ""
  213. }