Hugging Face's logo

ke-lly
/
47015772_1

Text Generation
Transformers
Safetensors
gpt2
trl
sft
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Community
47015772_1 / README.md
ke-lly's picture
ke-lly
Model save
c3d99e8 verified
preview code
|
raw
history blame contribute delete
No virus
6.41 kB
metadata
license: mit
base_model: openai-community/gpt2
tags:
  - trl
  - sft
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: '47015772_1'
    results: []

47015772_1

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

  • Loss: 0.6501
  • Accuracy: 0.0001

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: 1.41e-05
  • train_batch_size: 32
  • eval_batch_size: 4
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 2
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 256
  • total_eval_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • training_steps: 2000

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.1019 0.03 25 0.9747 0.0001
0.8784 0.06 50 0.8178 0.0001
0.8308 0.09 75 0.7727 0.0002
0.7975 0.11 100 0.7508 0.0001
0.787 0.14 125 0.7370 0.0001
0.7752 0.17 150 0.7279 0.0001
0.7703 0.2 175 0.7210 0.0001
0.7584 0.23 200 0.7157 0.0001
0.7662 0.26 225 0.7108 0.0001
0.7431 0.28 250 0.7073 0.0001
0.7437 0.31 275 0.7038 0.0001
0.7378 0.34 300 0.7006 0.0001
0.7252 0.37 325 0.6974 0.0001
0.7242 0.4 350 0.6948 0.0001
0.7284 0.43 375 0.6922 0.0001
0.7187 0.46 400 0.6898 0.0001
0.7169 0.48 425 0.6878 0.0001
0.723 0.51 450 0.6858 0.0001
0.7189 0.54 475 0.6834 0.0001
0.7142 0.57 500 0.6818 0.0001
0.7071 0.6 525 0.6798 0.0001
0.7131 0.63 550 0.6782 0.0001
0.7107 0.66 575 0.6768 0.0001
0.702 0.68 600 0.6761 0.0001
0.6937 0.71 625 0.6742 0.0001
0.6984 0.74 650 0.6732 0.0001
0.7023 0.77 675 0.6720 0.0001
0.7002 0.8 700 0.6708 0.0001
0.7057 0.83 725 0.6697 0.0001
0.6991 0.85 750 0.6687 0.0001
0.6951 0.88 775 0.6674 0.0001
0.7101 0.91 800 0.6668 0.0001
0.6931 0.94 825 0.6661 0.0001
0.6866 0.97 850 0.6652 0.0001
0.702 1.0 875 0.6643 0.0001
0.6996 1.03 900 0.6637 0.0001
0.7077 1.05 925 0.6629 0.0001
0.6955 1.08 950 0.6623 0.0001
0.6947 1.11 975 0.6619 0.0001
0.6933 1.14 1000 0.6612 0.0001
0.6855 1.17 1025 0.6601 0.0001
0.7004 1.2 1050 0.6597 0.0001
0.6934 1.22 1075 0.6594 0.0001
0.6863 1.25 1100 0.6585 0.0001
0.6886 1.28 1125 0.6580 0.0001
0.6851 1.31 1150 0.6579 0.0001
0.683 1.34 1175 0.6574 0.0001
0.703 1.37 1200 0.6570 0.0001
0.6792 1.4 1225 0.6564 0.0001
0.6849 1.42 1250 0.6558 0.0001
0.6856 1.45 1275 0.6556 0.0001
0.6856 1.48 1300 0.6551 0.0001
0.6856 1.51 1325 0.6550 0.0001
0.6857 1.54 1350 0.6543 0.0001
0.6856 1.57 1375 0.6539 0.0001
0.689 1.59 1400 0.6538 0.0001
0.6892 1.62 1425 0.6534 0.0001
0.6823 1.65 1450 0.6532 0.0001
0.6828 1.68 1475 0.6529 0.0001
0.6864 1.71 1500 0.6528 0.0001
0.6886 1.74 1525 0.6523 0.0001
0.6642 1.77 1550 0.6521 0.0001
0.6849 1.79 1575 0.6519 0.0001
0.6834 1.82 1600 0.6516 0.0001
0.6839 1.85 1625 0.6515 0.0001
0.6856 1.88 1650 0.6514 0.0001
0.6725 1.91 1675 0.6511 0.0001
0.6813 1.94 1700 0.6509 0.0001
0.6832 1.97 1725 0.6508 0.0001
0.6739 1.99 1750 0.6508 0.0001
0.6716 2.02 1775 0.6506 0.0001
0.6798 2.05 1800 0.6505 0.0001
0.6758 2.08 1825 0.6503 0.0001
0.6791 2.11 1850 0.6503 0.0001
0.6735 2.14 1875 0.6503 0.0001
0.6887 2.16 1900 0.6502 0.0001
0.686 2.19 1925 0.6502 0.0001
0.682 2.22 1950 0.6501 0.0001
0.675 2.25 1975 0.6501 0.0001
0.6792 2.28 2000 0.6501 0.0001

Framework versions

  • Transformers 4.37.0
  • Pytorch 2.0.0+cu118
  • Datasets 2.16.1
  • Tokenizers 0.15.1