|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- squad |
|
model-index: |
|
- name: bert-base-uncased-squad-v1-jpqd-ov-int8 |
|
results: [] |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# bert-base-uncased-squad-v1-jpqd-ov-int8 |
|
|
|
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the squad dataset. |
|
It was compressed using [NNCF](https://github.com/openvinotoolkit/nncf) with [Optimum |
|
Intel](https://github.com/huggingface/optimum-intel#openvino) following the [JPQD question-answering |
|
example](https://github.com/huggingface/optimum-intel/tree/main/examples/openvino/question-answering#joint-pruning-quantization-and-distillation-jpqd-for-bert-on-squad10). |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 3e-05 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 128 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 8.0 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
``` |
|
***** eval metrics ***** |
|
epoch = 8.0 |
|
eval_exact_match = 83.141 |
|
eval_f1 = 89.5906 |
|
eval_samples = 10784 |
|
``` |
|
|
|
### Framework versions |
|
|
|
- Transformers 4.26.1 |
|
- Pytorch 1.13.1+cu117 |
|
- Datasets 2.8.0 |
|
- Tokenizers 0.13.2 |
|
|
