gm-qlora-float16-idefics2-8b-indiana

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: 0.3401

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 10
• total_train_batch_size: 20
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 50
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.4248	0.1408	41	0.7825
0.7397	0.2817	82	0.6691
0.6407	0.4225	123	0.6101
0.606	0.5634	164	0.5650
0.575	0.7042	205	0.5273
0.5301	0.8451	246	0.4916
0.4834	0.9859	287	0.4677
0.3982	1.1268	328	0.4448
0.366	1.2676	369	0.4266
0.3493	1.4085	410	0.4068
0.3478	1.5493	451	0.3865
0.3173	1.6901	492	0.3681
0.3036	1.8310	533	0.3515
0.2917	1.9718	574	0.3401

Framework versions

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