| license: llama3 | |
| metrics: | |
| - accuracy | |
| The official base model weights for "Efficient Continual Pre-training by Mitigating the Stability Gap". | |
| The model has been continually pretrained on a high-quality medical sub-corpus from the RefinedWeb dataset. |
| license: llama3 | |
| metrics: | |
| - accuracy | |
| The official base model weights for "Efficient Continual Pre-training by Mitigating the Stability Gap". | |
| The model has been continually pretrained on a high-quality medical sub-corpus from the RefinedWeb dataset. |