SFTCodePhi-3-mini-4k-instructAPPS

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

• Loss: 0.6847

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• training_steps: 2000

Training results

Training Loss	Epoch	Step	Validation Loss
0.7559	0.6740	200	0.7106
0.6901	1.3479	400	0.6968
0.6867	2.0219	600	0.6922
0.6677	2.6959	800	0.6885
0.6594	3.3698	1000	0.6857
0.6502	4.0438	1200	0.6855
0.6419	4.7178	1400	0.6854
0.642	5.3917	1600	0.6847
0.6385	6.0657	1800	0.6847
0.6324	6.7397	2000	0.6847

Framework versions

• PEFT 0.11.0
• Transformers 4.40.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1