Phi-3.5-MultiCap-tool

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: 0.4301

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: 6

Training results

Training Loss	Epoch	Step	Validation Loss
0.9379	0.2256	50	0.9483
0.5752	0.4512	100	0.5681
0.5014	0.6768	150	0.4917
0.5155	0.9024	200	0.4731
0.4501	1.1280	250	0.4633
0.4451	1.3536	300	0.4565
0.4214	1.5792	350	0.4517
0.4795	1.8049	400	0.4477
0.4448	2.0305	450	0.4447
0.4181	2.2561	500	0.4423
0.4353	2.4817	550	0.4403
0.4613	2.7073	600	0.4386
0.4573	2.9329	650	0.4371
0.4719	3.1585	700	0.4362
0.4174	3.3841	750	0.4347
0.4337	3.6097	800	0.4340
0.4478	3.8353	850	0.4332
0.4156	4.0609	900	0.4325
0.4177	4.2865	950	0.4318
0.4113	4.5121	1000	0.4315
0.4343	4.7377	1050	0.4311
0.423	4.9633	1100	0.4307
0.4492	5.1889	1150	0.4307
0.4417	5.4146	1200	0.4303
0.4485	5.6402	1250	0.4302
0.4374	5.8658	1300	0.4301

Framework versions

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