phi-2-IA3_2-Python

This model is a fine-tuned version of microsoft/phi-2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 4.4937

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: 0.0002
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
6.003	0.1333	100	6.1601
5.786	0.2667	200	5.8407
5.4842	0.4	300	5.5615
5.4838	0.5333	400	5.3248
5.2047	0.6667	500	5.1423
4.7877	0.8	600	5.0128
4.755	0.9333	700	4.9153
4.9015	1.0667	800	4.8451
4.8667	1.2	900	4.7826
4.3832	1.3333	1000	4.7270
4.7475	1.4667	1100	4.6790
4.3649	1.6	1200	4.6366
4.3746	1.7333	1300	4.6068
4.6436	1.8667	1400	4.5778
4.8167	2.0	1500	4.5526
4.1062	2.1333	1600	4.5333
4.6027	2.2667	1700	4.5226
3.9483	2.4	1800	4.5102
4.742	2.5333	1900	4.5021
4.5186	2.6667	2000	4.4976
4.3743	2.8	2100	4.4951
4.3757	2.9333	2200	4.4937

Framework versions

• PEFT 0.17.1
• Transformers 4.56.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.0