CoolFace
Modelpublic

progs2002/star-trek-tng-script-generator

sourceHugging Facemitupdated 3y agoView on Hugging Face
1likes27downloads
Model Card

<!-- 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. -->

data cleaning and training code

https://github.com/progs2002/StarTrekTNG-ScriptGenerator

star-trek-tng-script-generator

This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:

  • —Loss: 2.8459

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: 0.0002
  • —trainbatchsize: 1
  • —evalbatchsize: 1
  • —seed: 42
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: cosine
  • —lrschedulerwarmup_steps: 50
  • —num_epochs: 3

Training results

Training LossEpochStepValidation Loss
3.15020.135003.0233
3.05380.2610002.9728
2.99510.3815002.9437
2.98910.5120002.9125
2.92890.6425002.9159
2.90910.7730002.9008
2.89160.8935002.8752
2.81221.0240002.8881
2.52241.1545002.8896
2.52841.2850002.8667
2.51911.455002.8599
2.51191.5360002.8488
2.48081.6665002.8296
2.46011.7970002.8081
2.43311.9175002.7993
2.37162.0480002.8518
2.15282.1785002.8634
2.12762.390002.8617
2.13292.4395002.8489
2.11352.55100002.8446
2.12592.68105002.8461
2.11422.81110002.8472
2.10712.94115002.8459

Framework versions

  • —Transformers 4.35.2
  • —Pytorch 2.1.0+cu121
  • —Datasets 2.16.1
  • —Tokenizers 0.15.0