Phi-3.5-MultiCap-mt-8

This model is a fine-tuned version of microsoft/Phi-3.5-mini-instruct on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6751

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.03
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss
1.005	0.5109	50	0.9799
0.7452	1.0217	100	0.7840
0.7469	1.5326	150	0.7372
0.7326	2.0434	200	0.7166
0.7264	2.5543	250	0.7045
0.7241	3.0651	300	0.6965
0.6842	3.5760	350	0.6910
0.6712	4.0868	400	0.6865
0.6824	4.5977	450	0.6831
0.6654	5.1086	500	0.6809
0.7041	5.6194	550	0.6791
0.6573	6.1303	600	0.6775
0.6606	6.6411	650	0.6762
0.6561	7.1520	700	0.6755
0.6799	7.6628	750	0.6751

Framework versions

• PEFT 0.12.0
• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1