Changes / Feeds / deepset/roberta-base-squad2
deepset/roberta-base-squad2
4 changes to a tracked key since 2020-01-22, newest first. Head as last read: adc3b06f79f797d1c575d5479d6f5efe54a9e3b4.
GPU memory to fine-tune it, card by card: roberta-base-squad2 on /m/.
| Date | Advisory | What changed, and what it does |
|---|---|---|
| 2020-11-11 | x_NXMC-2020-0025 | tokenizer model_max_length unset to 512; truncation defaults change (2020-11-11).596edb02b58a -> deedc3e42208 |
| 2020-04-24 | x_NXMC-2020-0024 breaking | EOS declared as 2 (was unset); generation that relied on the library default now stops on this token; pad token unset to 1; batched training masks a different id (2020-04-24).784fa08ee01e -> 4aa045039343 |
| 2020-03-23 | x_NXMC-2020-0023 breaking | pad token 0 to unset; batched training masks a different id (2020-03-23).8f39c9f238c2 -> 784fa08ee01e |
| 2020-01-31 | x_NXMC-2020-0022 | architectures declared as ["RobertaForQuestionAnswering"] (was unset) (2020-01-31).0eb280af9613 -> 8f39c9f238c2 |
Pin it, and let a bot move the pin
from_pretrained("deepset/roberta-base-squad2", revision="adc3b06f79f797d1c575d5479d6f5efe54a9e3b4")With the Renovate preset, a pull request moves this pin when the head moves, and links back here.
[](https://www.neurarch.com/changes-feeds/models/deepset/roberta-base-squad2.html)