gm-lora-bfloat16-idefics2-8b-xrayvqa-finetuned-medir

This model is a fine-tuned version of HuggingFaceM4/idefics2-8b on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.6893

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 10
• total_train_batch_size: 80
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.0712	0.1529	50	1.4681
1.1045	0.3057	100	1.4921
1.1032	0.4586	150	1.4891
1.1182	0.6114	200	1.4860
1.3049	0.7643	250	1.4402
1.4376	0.9172	300	1.4256
1.2084	1.0700	350	1.5379
0.8495	1.2229	400	1.5216
0.8827	1.3757	450	1.5241
0.8995	1.5286	500	1.5431
0.8939	1.6814	550	1.5198
0.9077	1.8343	600	1.5106
0.9128	1.9872	650	1.5176
0.5954	2.1400	700	1.6831
0.569	2.2929	750	1.6783
0.5578	2.4457	800	1.6861
0.5624	2.5986	850	1.6859
0.5754	2.7515	900	1.6907
0.5582	2.9043	950	1.6893

Framework versions

• Transformers 4.41.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1