Edit model card

longformer-replicated-pos-encodings-4096-2L

This model is a fine-tuned version of data/models/longformer-replicated-pos-encodings-4096 on the c4 en dataset. It achieves the following results on the evaluation set:

  • Loss: 1.9437
  • Accuracy: 0.6366

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: 0.0001
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: constant_with_warmup
  • lr_scheduler_warmup_ratio: 1.0
  • training_steps: 6400

Training results

Training Loss Epoch Step Validation Loss Accuracy
2.1052 1.0 6400 1.9433 0.6366

Framework versions

  • Transformers 4.20.0
  • Pytorch 1.12.0+cu113
  • Datasets 2.6.1
  • Tokenizers 0.12.1
Downloads last month
9

Dataset used to train kiddothe2b/longformer-replicated-pos-encodings-4096-2L

Evaluation results