phi-3-4k-instruct-domain-sft-1

This model is a fine-tuned version of microsoft/Phi-3-mini-4k-instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9328

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: 2e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 128
• total_train_batch_size: 512
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.8958	0.1445	10	1.0716
1.7839	0.2890	20	1.0458
1.7146	0.4335	30	1.0222
1.6457	0.5780	40	1.0029
1.591	0.7225	50	0.9872
1.552	0.8670	60	0.9740
1.5115	1.0115	70	0.9631
1.4681	1.1560	80	0.9541
1.4469	1.3005	90	0.9468
1.419	1.4450	100	0.9412
1.4033	1.5895	110	0.9371
1.3928	1.7340	120	0.9343
1.3887	1.8785	130	0.9328

Framework versions

• PEFT 0.10.0
• Transformers 4.40.1
• Pytorch 2.2.2+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1