gpt2_m020_tiny-stories_1024

This model is a fine-tuned version of on the roneneldan/TinyStories dataset. It achieves the following results on the evaluation set:

• Loss: 1.2217
• Accuracy: 0.6760

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
• 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
2.9424	0.0525	1000	2.4920	0.4375
1.9972	0.1050	2000	1.8161	0.5644
1.7424	0.1575	3000	1.6275	0.5979
1.6233	0.2100	4000	1.5269	0.6162
1.5467	0.2625	5000	1.4625	0.6284
1.4918	0.3150	6000	1.4187	0.6369
1.4537	0.3675	7000	1.3823	0.6438
1.4228	0.4200	8000	1.3529	0.6496
1.3987	0.4725	9000	1.3283	0.6543
1.378	0.5250	10000	1.3099	0.6578
1.3567	0.5775	11000	1.2938	0.6611
1.3436	0.6300	12000	1.2784	0.6643
1.328	0.6825	13000	1.2664	0.6666
1.3157	0.7350	14000	1.2546	0.6690
1.3052	0.7875	15000	1.2462	0.6707
1.2966	0.8400	16000	1.2370	0.6726
1.2943	0.8925	17000	1.2296	0.6742
1.2837	0.9450	18000	1.2244	0.6754
1.2825	0.9975	19000	1.2217	0.6760

Framework versions

• Transformers 4.42.3
• Pytorch 2.2.2+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1