CoolFace
Modelpublic

devika-tiwari/gpt2_small_expandedbabyLM_100M_43

sourceHugging Faceupdated 1mo agoView on Hugging Face
0likes779downloads
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. -->

gpt2smallexpandedbabyLM100M43

This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset. It achieves the following results on the evaluation set:

  • —Loss: 2.7602

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
  • —trainbatchsize: 256
  • —evalbatchsize: 256
  • —seed: 43
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 4000
  • —num_epochs: 20

Training results

Training LossEpochStepValidation Loss
3.56471.050833.1769
3.29872.0101662.9774
3.18313.0152492.9004
3.11334.0203322.8601
3.07125.0254152.8346
3.02736.0304982.8154
2.99867.0355812.8037
2.97228.0406642.7926
2.94339.0457472.7850
2.919810.0508302.7810
2.905111.0559132.7749
2.880812.0609962.7715
2.865313.0660792.7689
2.843914.0711622.7663
2.835515.0762452.7656
2.816716.0813282.7631
2.797717.0864112.7611
2.779818.0914942.7608
2.764419.0965772.7602
2.749820.01016602.7604

Framework versions

  • —Transformers 4.30.2
  • —Pytorch 2.11.0+cu130
  • —Datasets 4.1.1
  • —Tokenizers 0.13.3