phi-3-mini-LoRA

This model is a fine-tuned version of microsoft/Phi-3-mini-4k-instruct on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.8588

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: 0.0001
• train_batch_size: 2
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.0606	0.1071	100	1.0032
0.9107	0.2141	200	0.9256
0.8783	0.3212	300	0.9081
0.8761	0.4283	400	0.8986
0.8651	0.5353	500	0.8920
0.864	0.6424	600	0.8875
0.8759	0.7495	700	0.8828
0.8584	0.8565	800	0.8807
0.8677	0.9636	900	0.8784
0.8507	1.0707	1000	0.8757
0.8499	1.1777	1100	0.8739
0.8446	1.2848	1200	0.8718
0.8637	1.3919	1300	0.8712
0.8238	1.4989	1400	0.8686
0.8231	1.6060	1500	0.8681
0.8361	1.7131	1600	0.8661
0.8319	1.8201	1700	0.8652
0.8166	1.9272	1800	0.8643
0.8312	2.0343	1900	0.8634
0.834	2.1413	2000	0.8625
0.8362	2.2484	2100	0.8616
0.8413	2.3555	2200	0.8611
0.8153	2.4625	2300	0.8605
0.8235	2.5696	2400	0.8607
0.7958	2.6767	2500	0.8598
0.8137	2.7837	2600	0.8593
0.8162	2.8908	2700	0.8591
0.8317	2.9979	2800	0.8588

Framework versions

• PEFT 0.12.0
• Transformers 4.43.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1