| # Instructions for Miðeind neural parser (experimental) | |
| The docker container can be also be located on dockerhub (`docker pull mideind/neural-parser`). To load the saved image into docker use: | |
| ``` | |
| gunzip neural-parser.gz | |
| docker load -i neural-parser.tar | |
| ``` | |
| The input to the parser is a text file (`${INPUT_FILE}`) where each line contains a tokenized sentence that will be parsed. | |
| The output file will be located in `${OUTPUT_DIR}/output.txt` . To run the parser use the following: | |
| ``` | |
| docker run --volume ${INPUT_FILE}:/data/input.txt --volume ${OUTPUT_DIR}:/data/ mideind/neural-parser | |
| ``` | |
| --- | |
| # Leiðbeiningar fyrir tauganetsþáttara Miðeindar | |
| Athugið að einnig má finna gáminn á dockerhub (`docker pull mideind/neural-parser`). Til að hlaða vistuðum gámi inn í docker skal nota: | |
| ``` | |
| gunzip neural-parser.gz | |
| docker load -i neural-parser.tar | |
| ``` | |
| Inntakið í þáttarann er texta skrá (`${INPUT_FILE}`) þar sem hver lína geymir eina tókaða málsgrein. Eftir keyrslu má finna úttakið í skránni `${OUTPUT_DIR}/output.txt` . | |
| Til að keyra þáttarann skal nota: | |
| ``` | |
| docker run --volume ${INPUT_FILE}:/data/input.txt --volume ${OUTPUT_DIR}:/data/ mideind/neural-parser | |
| ``` | |