mistral-viggo-finetune

This model is a fine-tuned version of mistralai/Mistral-7B-v0.3 on the viggo dataset. It achieves the following results on the evaluation set:

• Loss: 0.2520

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: 2.5e-05
• train_batch_size: 4
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 5
• training_steps: 100

Training results

Training Loss	Epoch	Step	Validation Loss
0.297	0.1567	50	0.2849
0.2569	0.3135	100	0.2520

Framework versions

• PEFT 0.13.1.dev0
• Transformers 4.46.0.dev0
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0