phi-3-mini-QLoRA

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: 0.5761

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.1336	0.1809	100	0.6788
0.6283	0.3618	200	0.6030
0.5944	0.5427	300	0.5931
0.5953	0.7237	400	0.5879
0.5793	0.9046	500	0.5852
0.5908	1.0855	600	0.5832
0.5717	1.2664	700	0.5812
0.5748	1.4473	800	0.5802
0.5876	1.6282	900	0.5787
0.5725	1.8091	1000	0.5778
0.5749	1.9900	1100	0.5772
0.5646	2.1710	1200	0.5769
0.5806	2.3519	1300	0.5764
0.5679	2.5328	1400	0.5762
0.5683	2.7137	1500	0.5761
0.5715	2.8946	1600	0.5761

Framework versions

• PEFT 0.11.1
• Transformers 4.41.2
• Pytorch 2.1.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1