Edit model card

mistral-ft

This model is a fine-tuned version of TheBloke/Mistral-7B-Instruct-v0.2-GPTQ on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 1.2527

Model description

This model is a fine-tuned version of TheBloke/Mistral-7B-Instruct-v0.2-GPTQ for radiology reports conclusions generation.

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: 0.0002
  • train_batch_size: 4
  • eval_batch_size: 4
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 2
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
2.3229 0.97 27 1.8742
1.7299 1.98 55 1.6318
1.5704 2.99 83 1.4831
1.4553 4.0 111 1.4052
1.4421 4.97 138 1.3805
1.3759 5.98 166 1.3759
1.3658 6.99 194 1.3355
1.3271 8.0 222 1.2890
1.3299 8.97 249 1.2618
1.2296 9.73 270 1.2527

Framework versions

  • PEFT 0.10.0
  • Transformers 4.38.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
1
Inference Examples
Inference API (serverless) does not yet support peft models for this pipeline type.

Adapter for