I’m experimenting with GPT-Neo variants, and I wonder whether these models have different checkpoints? I.e., independently trained versions with different weights. If so, how can I access those?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Availability of checkpoints earlier in training? | 0 | 212 | April 1, 2021 | |
| Is there any reason why GPT-Neo would behave differently (fundamentally) from GPT2? | 0 | 423 | January 15, 2023 | |
| Learning rate and checkpoints | 0 | 433 | March 29, 2022 | |
| How to see checkpoint lineage? | 0 | 122 | March 6, 2024 | |
| Not getting a good model at first try | 0 | 362 | April 14, 2022 |