k3

This model is a fine-tuned version of deepseek-ai/deepseek-coder-1.3b-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1575
• Eval/rewards/chosen: 2.5930
• Eval/logps/chosen: -116.6651
• Eval/rewards/rejected: -10.1233
• Eval/logps/rejected: -282.2972
• Eval/rewards/margins: 12.7163
• Eval/kl: 7.4003

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
• gradient_accumulation_steps: 16
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• lr_scheduler_warmup_steps: 200
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.2045	0.8552	100	0.2515	16.7171
0.0457	1.7103	200	0.2056	14.2697
0.0271	2.5655	300	0.1644	9.4286
0.0157	3.4206	400	0.1575	7.4003

Framework versions

• Transformers 4.43.0.dev0
• Pytorch 2.2.2+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1