codeparrot-ds-model

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

• Loss: 0.9517

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: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 5000
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
2.4181	0.1533	10000	1.5849
1.5361	0.3065	20000	1.4070
1.4242	0.4598	30000	1.3356
1.3721	0.6131	40000	1.2948
1.3327	0.7664	50000	1.2638
1.3053	0.9196	60000	1.2403
1.2751	1.0729	70000	1.2146
1.2491	1.2262	80000	1.1876
1.2258	1.3795	90000	1.1642
1.1997	1.5327	100000	1.1374
1.1724	1.6860	110000	1.1113
1.1444	1.8393	120000	1.0826
1.1157	1.9926	130000	1.0551
1.0762	2.1458	140000	1.0298
1.0513	2.2991	150000	1.0042
1.0275	2.4524	160000	0.9827
1.0067	2.6056	170000	0.9662
0.9906	2.7589	180000	0.9556
0.9827	2.9122	190000	0.9517

Framework versions

• Transformers 4.48.3
• Pytorch 2.3.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0