metadata

base_model: avichr/heBERT
tags:
  - generated_from_trainer
datasets:
  - imvladikon/parashoot
model-index:
  - name: hebert_parashoot
    results: []
language:
  - he
metrics:
  - f1
  - exact_match
pipeline_tag: question-answering

hebert_parashoot

This model is a fine-tuned version of avichr/heBERT on the imvladikon/parashoot dataset.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 3e-05
train_batch_size: 4
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 5.0

Training results

results:

{
    "epoch": 5.0,
    "eval_exact_match": 18.099547511312217,
    "eval_f1": 36.8601893452485,
    "eval_runtime": 6.7527,
    "eval_samples": 249,
    "eval_samples_per_second": 36.874,
    "eval_steps_per_second": 4.739
}

(which reflects with results from the https://arxiv.org/pdf/2109.11314.pdf : F1: 36.7, EM: 18.2)

Framework versions

Transformers 4.31.0
Pytorch 2.0.1+cu118
Datasets 2.14.2
Tokenizers 0.13.3