|
--- |
|
license: mit |
|
base_model: microsoft/deberta-v3-base |
|
tags: |
|
- generated_from_trainer |
|
- calibration |
|
- uncertainty |
|
model-index: |
|
- name: apricot_clustering_coqa_deberta-v3-base_for_vicuna-7b-v1.5 |
|
results: [] |
|
datasets: |
|
- stanfordnlp/coqa |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# apricot_clustering_coqa_deberta-v3-base_for_vicuna-7b-v1.5 |
|
|
|
This model is fine-tuned for black-box LLM calibration as part of the 🍑 Apricot paper ["Calibrating Large Language Models Using Their Generations Only"](https://github.com/parameterlab/apricot) (ACL 2024). |
|
|
|
## Model description |
|
|
|
This model is a fine-tuned version of [microsoft/deberta-v3-base](https://huggingface.co/microsoft/deberta-v3-base) to predict the calibration score for the [lmsys/vicuna-7b-v1.5](https://huggingface.co/lmsys/vicuna-7b-v1.5) model on the questions from the stanfordnlp/coqa dataset. It uses the clustering type of calibration target score. |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
**TODO**: update the values below |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 5e-05 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 3.0 |
|
|
|
### Framework versions |
|
|
|
- Transformers 4.32.0 |
|
- Pytorch 2.0.0+cu117 |
|
- Datasets 2.14.6 |
|
- Tokenizers 0.13.3 |
