I recently publish My TF2 inference kernel. However, the metric/validation part is not work yet. It doesn't give scores that matches the public LB score. I am still trying to work on it.
\n\nAlso, my way of calculating final prediction is not as optimal as prokaj's kernel or as mmmarchetti's kernel. Probably I will update it also.
", "messages": [ { "id": 701248, "postDate": "2019-12-23T09:32:57.950Z", "content": "I recently publish My TF2 inference kernel. However, the metric/validation part is not work yet. It doesn't give scores that matches the public LB score. I am still trying to work on it.
\n\nAlso, my way of calculating final prediction is not as optimal as prokaj's kernel or as mmmarchetti's kernel. Probably I will update it also.
", "rawMarkdown": "I recently publish [My TF2 inference kernel](https://www.kaggle.com/yihdarshieh/inference-use-hugging-face-models). However, **the metric/validation part is not work yet.** It doesn't give scores that matches the public LB score. I am still trying to work on it.\n\nAlso, my way of calculating final prediction is not as optimal as [prokaj's kernel](https://www.kaggle.com/prokaj/bert-joint-baseline-notebook) or as [mmmarchetti's kernel](https://www.kaggle.com/mmmarchetti/tensorflow-2-0-bert-yes-no-answers). Probably I will update it also.", "votes": 2 } ], "comments": [], "raw_markdown_by_id": { "701248": "I recently publish [My TF2 inference kernel](https://www.kaggle.com/yihdarshieh/inference-use-hugging-face-models). However, **the metric/validation part is not work yet.** It doesn't give scores that matches the public LB score. I am still trying to work on it.\n\nAlso, my way of calculating final prediction is not as optimal as [prokaj's kernel](https://www.kaggle.com/prokaj/bert-joint-baseline-notebook) or as [mmmarchetti's kernel](https://www.kaggle.com/mmmarchetti/tensorflow-2-0-bert-yes-no-answers). Probably I will update it also." } }