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

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: 1.4709

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
• lr_scheduler_warmup_steps: 50
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
1.3811	0.0680	50	1.5852
1.3546	0.1359	100	1.5141
1.3012	0.2039	150	1.5571
1.3035	0.2718	200	1.5795
1.3397	0.3398	250	1.5959
1.3092	0.4077	300	1.5243
1.3247	0.4757	350	1.5147
1.3035	0.5436	400	1.4985
1.347	0.6116	450	1.5187
1.2792	0.6795	500	1.5179
1.2825	0.7475	550	1.5464
1.3046	0.8154	600	1.4881
1.2838	0.8834	650	1.4943
1.2639	0.9513	700	1.5134
1.2119	1.0193	750	1.5082
0.9573	1.0873	800	1.4866
0.9336	1.1552	850	1.4962
0.9816	1.2232	900	1.4968
0.9159	1.2911	950	1.4953
0.9604	1.3591	1000	1.5154
0.9581	1.4270	1050	1.4999
0.9403	1.4950	1100	1.5007
0.9735	1.5629	1150	1.4754
0.9606	1.6309	1200	1.4703
0.9313	1.6988	1250	1.4776
0.924	1.7668	1300	1.4747
0.9283	1.8347	1350	1.4719
0.928	1.9027	1400	1.4701
0.9281	1.9706	1450	1.4709

Framework versions

• Transformers 4.41.0.dev0
• Pytorch 1.13.1+cu117
• Datasets 2.19.0
• Tokenizers 0.19.1