twitter 140 temporal word gap filling
Dataset from paper "Twitter Sentiment Classification using Distant Supervision" [ver. 1.0.2]
This is a long list of all submissions, if you want to see only the best, click leaderboard.
| # | submitter | when | ver. | description | dev-0 PerplexityHashed | test-A PerplexityHashed | |
|---|---|---|---|---|---|---|---|
| 5 | kubapok | 2021-11-15 12:29 | 1.0.2 | regular roberta no ft | 65.5 | 66.2 | |
| 6 | kubapok | 2021-11-15 10:51 | 1.0.2 | regular roberta no ft | 65.5 | 763.5 | |
| 2 | p/tlen | 2021-11-15 06:07 | 1.0.2 | RoBERTa fine-tuned (without year) roberta fine-tuned roberta-base no-temporal | 33.8 | 34.6 | |
| 1 | p/tlen | 2021-11-14 14:06 | 1.0.1 | RoBERTa fine-tuned (with year) temporal roberta fine-tuned roberta-base | 32.9 | 33.6 | |
| 7 | kubapok | 2021-11-14 12:24 | 1.0.1 | same probability | 1024.0 | 1024.0 | |
| 9 | kubapok | 2021-11-13 23:46 | 1.0.1 | same probability | 1024.0 | N/A | |
| 3 | p/tlen | 2021-11-13 21:50 | 1.0.1 | RoBERTa from scratch (with year) temporal roberta roberta-base no-pretrained | 45.5 | 46.1 | |
| 4 | p/tlen | 2021-11-13 08:29 | 1.0.1 | RoBERTa from scratch (without year) roberta roberta-base no-temporal no-pretrained | 48.8 | 49.1 | |
| 8 | p/tlen | 2021-11-12 11:38 | 1.0.1 | RoBERTa fine-tuned (without year) roberta fine-tuned roberta-base no-temporal | 33.8 | N/A |