ninyx's picture
Model save
cf51f09 verified
metadata
license: apache-2.0
library_name: peft
tags:
  - trl
  - sft
  - generated_from_trainer
base_model: mistralai/Mistral-7B-Instruct-v0.3
datasets:
  - generator
metrics:
  - bleu
  - rouge
model-index:
  - name: Mistral-7B-Instruct-v0.3-advisegpt-v0.4
    results: []

Mistral-7B-Instruct-v0.3-advisegpt-v0.4

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

  • Loss: 0.0776
  • Bleu: {'bleu': 0.9592766854579555, 'precisions': [0.9778672968005702, 0.9629777800504739, 0.952562376464522, 0.9440303244645156], 'brevity_penalty': 1.0, 'length_ratio': 1.0002070868729431, 'translation_length': 666525, 'reference_length': 666387}
  • Rouge: {'rouge1': 0.9765393241338379, 'rouge2': 0.960274899679536, 'rougeL': 0.9752854409851488, 'rougeLsum': 0.9763366883065228}
  • Exact Match: {'exact_match': 0.0}

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: 2e-05
  • train_batch_size: 1
  • eval_batch_size: 1
  • seed: 42
  • gradient_accumulation_steps: 15
  • total_train_batch_size: 15
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • num_epochs: 3
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Bleu Rouge Exact Match
0.0592 0.9998 2664 0.0792 {'bleu': 0.957140829496306, 'precisions': [0.9770110285842899, 0.9611535701983837, 0.9499650178830994, 0.9408134298916666], 'brevity_penalty': 1.0, 'length_ratio': 1.0000945396593872, 'translation_length': 666450, 'reference_length': 666387} {'rouge1': 0.9756420869808171, 'rouge2': 0.958253583847128, 'rougeL': 0.9741670140375769, 'rougeLsum': 0.9753898276329086} {'exact_match': 0.0}
0.0518 2.0000 5329 0.0776 {'bleu': 0.9592766854579555, 'precisions': [0.9778672968005702, 0.9629777800504739, 0.952562376464522, 0.9440303244645156], 'brevity_penalty': 1.0, 'length_ratio': 1.0002070868729431, 'translation_length': 666525, 'reference_length': 666387} {'rouge1': 0.9765393241338379, 'rouge2': 0.960274899679536, 'rougeL': 0.9752854409851488, 'rougeLsum': 0.9763366883065228} {'exact_match': 0.0}
0.0439 2.9994 7992 0.0830 {'bleu': 0.9593680325138967, 'precisions': [0.97789654044549, 0.9630261327317164, 0.9526617494511856, 0.9442157972615742], 'brevity_penalty': 1.0, 'length_ratio': 1.0001725723941193, 'translation_length': 666502, 'reference_length': 666387} {'rouge1': 0.9766709553577743, 'rouge2': 0.9604006931620985, 'rougeL': 0.9753845279467352, 'rougeLsum': 0.9764641972952484} {'exact_match': 0.0}

Framework versions

  • PEFT 0.10.0
  • Transformers 4.40.2
  • Pytorch 2.2.0
  • Datasets 2.19.1
  • Tokenizers 0.19.1