cygu/pythia-1.4b-sampling-watermark-distill-kgw-k2-gamma0.25-delta2

Model description

Sampling-based watermark distilled Pythia 1.4B using the KGW $k=2,\gamma =0.25,\delta =2k=2,\; \backslash gamma=0.25,\; \backslash delta=2$ watermarking strategy in the paper On the Learnability of Watermarks for Language Models.

Training hyperparameters

The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 64 - eval_batch_size: 8 - seed: 42 - distributed_type: multi-GPU - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 500 - num_epochs: 1.0

Framework versions

• Transformers 4.29.2
• Pytorch 2.0.1+cu117
• Datasets 2.13.1
• Tokenizers 0.13.3