gptneo-1.3B-rm-harmless

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

• Loss: 0.5449
• Accuracy: 0.7184
• Average Pos Score: 1.1621
• Average Neg Score: 0.4114
• Average Abs Score Diff: 1.1133

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Average Pos Score	Average Neg Score	Average Abs Score Diff
0.6274	0.15	200	0.5898	0.6968	1.7881	1.2979	0.8413
0.5058	0.3	400	0.5811	0.7115	0.8394	-0.0314	1.3799
0.5699	0.45	600	0.5527	0.7167	1.2803	0.6006	1.0293
0.5533	0.6	800	0.5542	0.7171	0.6689	-0.1876	1.2930
0.5396	0.75	1000	0.5444	0.7223	1.0977	0.3601	1.0977
0.5791	0.9	1200	0.5449	0.7184	1.1621	0.4114	1.1133

Framework versions

• Transformers 4.26.1
• Pytorch 1.12.1+cu113
• Datasets 2.9.0
• Tokenizers 0.13.2