phi4_instruct_20250902_0749

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

• Loss: 0.6032
• Map@3: 0.8772

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• 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
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Map@3
16.4003	0.0523	20	1.3662	0.7260
9.5268	0.1046	40	1.0851	0.7793
8.3438	0.1569	60	0.9705	0.7830
7.495	0.2092	80	0.9053	0.8070
7.4283	0.2615	100	0.8488	0.8326
6.5147	0.3138	120	0.7270	0.8505
5.6739	0.3661	140	0.7694	0.8445
5.8069	0.4184	160	0.6465	0.8692
5.8288	0.4707	180	0.6613	0.8537
4.7902	0.5230	200	0.6032	0.8827
5.0708	0.5754	220	0.5588	0.8832
4.7105	0.6277	240	0.5590	0.8880
4.6516	0.6800	260	0.5281	0.8944
4.2103	0.7323	280	0.5666	0.8824
4.9022	0.7846	300	0.5486	0.8936
3.8328	0.8369	320	0.5402	0.8959
4.2973	0.8892	340	0.5149	0.9046
3.8971	0.9415	360	0.4942	0.8943
3.9711	0.9938	380	0.5736	0.8894
3.6985	1.0445	400	0.5119	0.9000
2.971	1.0968	420	0.6032	0.8772

Framework versions

• PEFT 0.17.1
• Transformers 4.56.0
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.0