| language: en | |
| license: mit | |
| base_model: roberta-base | |
| tags: | |
| - token-classification | |
| - named-entity-recognition | |
| - resumes | |
| - cv-parser | |
| pipeline_tag: token-classification | |
| # CV Parser NER - RoBERTa Dataset 2 | |
| Fine-tuned `roberta-base` token-classification model for the WQF7007 automated CV parser dashboard. | |
| This is the best logged model trained on dataset 2 (`relabel1` committed CSV before the latest relabel script changes). | |
| - Dataset LFS oid: `89bc25de2e2a800d6ff59380e409021662428f5a578d5c17c7c899d3a0a92e24` | |
| - Source checkpoint: `checkpoints/roberta-base/checkpoint-6685` | |
| - Validation F1: `0.6752` | |
| - Test F1: `0.6854` | |
| - Labels: `JOB_TITLE`, `SKILL`, `EDUCATION` using BIO tags | |