Mistral-7B-v0.1_AviationQA

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.1 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 10.2218
• Bleu: 0.3549
• Rougel: 0.4821
• F1: 0.0005
• Perplexity: 30611.9414

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: 3
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• lr_scheduler_warmup_ratio: 0.03
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Rougel	F1	Perplexity
10.2565	1.0	25	10.2347	0.2507	0.3703	0.0009	29825.1777
10.1919	2.0	50	10.2232	0.3097	0.4539	0.0008	30276.8613
10.1739	3.0	75	10.2218	0.3549	0.4821	0.0005	30611.9414

Framework versions

• PEFT 0.14.0
• Transformers 4.49.0
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0