graphcore-rahult's picture
update model card README.md
78263b2
|
raw
history blame
1.49 kB
metadata
license: mit
tags:
  - generated_from_trainer
model-index:
  - name: gpt2-finetuned-wikitext2
    results: []

gpt2-finetuned-wikitext2

This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 3.6758

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: 2e-05
  • train_batch_size: 1
  • eval_batch_size: 1
  • seed: 42
  • distributed_type: IPU
  • gradient_accumulation_steps: 16
  • total_train_batch_size: 128
  • total_eval_batch_size: 20
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 3.0
  • training precision: Mixed Precision

Training results

Training Loss Epoch Step Validation Loss
4.1344 1.0 145 3.7656
3.8846 2.0 290 3.6953
3.7496 3.0 435 3.6758

Framework versions

  • Transformers 4.20.1
  • Pytorch 1.10.0+cpu
  • Datasets 2.7.1
  • Tokenizers 0.12.1