Once deciding on a few constant labels, can one freeze the current weights of the models, to ensure a quicker inference time?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Is it possible to disassemble a zero-shot model? | 0 | 449 | March 3, 2022 | |
| Class prediction in a zero/few-shot setting at inference time | 0 | 399 | July 27, 2022 | |
| How to freeze BERT weights | 0 | 954 | October 28, 2021 | |
| Having multiple candidate labels in a zero shot classification model | 3 | 561 | May 8, 2024 | |
| Freezing weights of new tokens in the input embedding | 2 | 583 | September 25, 2024 |