sft-llava-1.5-7b-hf3

This model is a fine-tuned version of llava-hf/llava-1.5-7b-hf on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 13.1181
• Bleu: 0.0
• Rouge1: 0.0651
• Rouge2: 0.0043
• Rougel: 0.0508
• Bertscore Precision: 0.6243
• Bertscore Recall: 0.7482
• Bertscore F1: 0.6806

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.03
• num_epochs: 5.0

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Rouge1	Rouge2	Rougel	Bertscore Precision	Bertscore Recall	Bertscore F1
6.903	0.3101	200	22.1793	0.0	0.0440	0.0	0.0441	0.6243	0.7482	0.6806
6.585	0.6202	400	27.3559	0.0	0.0546	0.0043	0.0425	0.6243	0.7482	0.6806
6.5197	0.9302	600	26.1987	0.0	0.0546	0.0043	0.0425	0.6243	0.7482	0.6806
6.2662	1.2403	800	21.1666	0.0	0.0633	0.0043	0.0520	0.6243	0.7482	0.6806
6.0303	1.5504	1000	21.0359	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.7602	1.8605	1200	19.0201	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.6359	2.1705	1400	18.6311	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.5176	2.4806	1600	17.9442	0.0	0.0649	0.0043	0.0496	0.6243	0.7482	0.6806
5.4608	2.7907	1800	16.6921	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.2881	3.1008	2000	15.3415	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.2429	3.4109	2200	14.8475	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.1929	3.7209	2400	14.2828	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.1259	4.0310	2600	13.8075	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.0379	4.3411	2800	13.4751	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.1071	4.6512	3000	13.2275	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806
5.1082	4.9612	3200	13.1181	0.0	0.0651	0.0043	0.0508	0.6243	0.7482	0.6806

Framework versions

• Transformers 4.45.2
• Pytorch 2.2.0a0+81ea7a4
• Datasets 3.0.1
• Tokenizers 0.20.1