gpt2-jokes

This model is a fine-tuned version of gpt2 on the Fraser/short-jokes dataset. It achieves the following results on the evaluation set:

• Loss: 0.6748
• Accuracy: 0.8796

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 128
• total_eval_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.06	100	0.7285	0.8732
No log	0.12	200	0.7141	0.8747
No log	0.17	300	0.7056	0.8757
No log	0.23	400	0.6992	0.8764
0.7907	0.29	500	0.6942	0.8771
0.7907	0.35	600	0.6906	0.8777
0.7907	0.41	700	0.6873	0.8779
0.7907	0.47	800	0.6848	0.8782
0.7907	0.52	900	0.6830	0.8786
0.7105	0.58	1000	0.6809	0.8788
0.7105	0.64	1100	0.6794	0.8790
0.7105	0.7	1200	0.6780	0.8792
0.7105	0.76	1300	0.6770	0.8793
0.7105	0.81	1400	0.6760	0.8794
0.7034	0.87	1500	0.6755	0.8794
0.7034	0.93	1600	0.6750	0.8795
0.7034	0.99	1700	0.6748	0.8795

Framework versions

• Transformers 4.28.0.dev0
• Pytorch 2.0.0-rc1
• Datasets 2.10.1
• Tokenizers 0.13.2