Request for Sharing Training Scripts and Datasets for Academic Purposes
#2
by
Hrithik2212
- opened
Is it possible to share the training scripts/notebooks and the data used for academic purposes
@Hrithik2212
, you can find the training script here. The model was trained in two stages. The first stage is training on a large dataset like numind/NuNER
. Then, we fine-tuned the model on the NER subset of our multi-task datasets: https://huggingface.co/datasets/knowledgator/GLINER-multi-task-synthetic-data and https://huggingface.co/datasets/urchade/pile-mistral-v0.1.
@Ihor Thank you , May I know(out of curiosity) what was the GPU infra for both experimenting and training the final model
@Hrithik2212 , it was L40 GPU.