codeparrot-ds

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

• Loss: 5.1291

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: 0.0005
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
9.5166	0.2332	10	7.9560
7.161	0.4665	20	6.8948
6.7973	0.6997	30	6.7108
6.4652	0.9329	40	6.3878
6.1192	1.1662	50	6.1154
5.842	1.3994	60	5.8976
5.6262	1.6327	70	5.7493
5.4633	1.8659	80	5.6221
5.3212	2.0991	90	5.5376
5.1513	2.3324	100	5.4584
5.118	2.5656	110	5.3924
4.9714	2.7988	120	5.3301
4.9133	3.0321	130	5.2827
4.7702	3.2653	140	5.2460
4.7302	3.4985	150	5.2081
4.6988	3.7318	160	5.1740
4.6927	3.9650	170	5.1537
4.6044	4.1983	180	5.1442
4.5763	4.4315	190	5.1361
4.5913	4.6647	200	5.1298
4.5759	4.8980	210	5.1291

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1