Spaces:
Sleeping
Sleeping
| # Based on new.prm (and by extension COLLINS.prm) | |
| # The only change from new.prm is increasing MAX_ERROR. The evaluation should be | |
| # identical to the standard setup, except that evalb won't give up early for a | |
| # parser that has just started training and does not yet produce good results. | |
| ##------------------------------------------## | |
| ## Debug mode ## | |
| ## 0: No debugging ## | |
| ## 1: print data for individual sentence ## | |
| ## 2: print detailed bracketing info ## | |
| ##------------------------------------------## | |
| DEBUG 0 | |
| ##------------------------------------------## | |
| ## MAX error ## | |
| ## Number of error to stop the process. ## | |
| ## This is useful if there could be ## | |
| ## tokanization error. ## | |
| ## The process will stop when this number## | |
| ## of errors are accumulated. ## | |
| ##------------------------------------------## | |
| MAX_ERROR 10000 | |
| ##------------------------------------------## | |
| ## Cut-off length for statistics ## | |
| ## At the end of evaluation, the ## | |
| ## statistics for the senetnces of length## | |
| ## less than or equal to this number will## | |
| ## be shown, on top of the statistics ## | |
| ## for all the sentences ## | |
| ##------------------------------------------## | |
| CUTOFF_LEN 40 | |
| ##------------------------------------------## | |
| ## unlabeled or labeled bracketing ## | |
| ## 0: unlabeled bracketing ## | |
| ## 1: labeled bracketing ## | |
| ##------------------------------------------## | |
| LABELED 1 | |
| ##------------------------------------------## | |
| ## Delete labels ## | |
| ## list of labels to be ignored. ## | |
| ## If it is a pre-terminal label, delete ## | |
| ## the word along with the brackets. ## | |
| ## If it is a non-terminal label, just ## | |
| ## delete the brackets (don't delete ## | |
| ## deildrens). ## | |
| ##------------------------------------------## | |
| DELETE_LABEL TOP | |
| DELETE_LABEL S1 | |
| DELETE_LABEL -NONE- | |
| DELETE_LABEL , | |
| DELETE_LABEL : | |
| DELETE_LABEL `` | |
| DELETE_LABEL '' | |
| DELETE_LABEL . | |
| DELETE_LABEL ? | |
| DELETE_LABEL ! | |
| ##------------------------------------------## | |
| ## Delete labels for length calculation ## | |
| ## list of labels to be ignored for ## | |
| ## length calculation purpose ## | |
| ##------------------------------------------## | |
| DELETE_LABEL_FOR_LENGTH -NONE- | |
| ##------------------------------------------## | |
| ## Labels to be considered for misquote ## | |
| ## (could be possesive or quote) ## | |
| ##------------------------------------------## | |
| QUOTE_LABEL `` | |
| QUOTE_LABEL '' | |
| QUOTE_LABEL POS | |
| ##------------------------------------------## | |
| ## These ones are less common, but ## | |
| ## are on occasion output by parsers: ## | |
| ##------------------------------------------## | |
| QUOTE_LABEL NN | |
| QUOTE_LABEL CD | |
| QUOTE_LABEL VBZ | |
| QUOTE_LABEL : | |
| ##------------------------------------------## | |
| ## Equivalent labels, words ## | |
| ## the pairs are considered equivalent ## | |
| ## This is non-directional. ## | |
| ##------------------------------------------## | |
| EQ_LABEL ADVP PRT | |
| # EQ_WORD Example example | |