Views
No views yet
microsoft/deberta-v3-base trained on the QASC dataset.1import torch
2from transformers import AutoModelForSequenceClassification, AutoTokenizer
3
4model_id = "yophis/DRM-DeBERTa-v3-Base-qasc"
5
6# Load the tokenizer
7tokenizer = AutoTokenizer.from_pretrained(model_id)
8tokenizer.pad_token = tokenizer.eos_token
9
10# Load the model
11model = AutoModelForSequenceClassification.from_pretrained(model_id, device_map="auto")
12model.config.pad_token_id = model.config.eos_token_id
13
14# Input template
15input_text = "Question: {formatted_question} Context: {combinedfact}"1@article{chaichana2025decom,
2 title={Decom-Renorm-Merge: Model Merging on the Right Space Improves Multitasking},
3 author={Chaichana, Yuatyong and Trachu, Thanapat and Limkonchotiwat, Peerat and Preechakul, Konpat and Khandhawit, Tirasan and Chuangsuwanich, Ekapol},
4 journal={arXiv preprint arXiv:2505.23117},
5 year={2025}
6}