sindhu2003's picture
sindhu2003/legal-document-summarization
1685079 verified
metadata
base_model: NousResearch/Llama-2-7b-hf
library_name: peft
tags:
  - trl
  - sft
  - generated_from_trainer
model-index:
  - name: llama2-docsum-adapter
    results: []

Visualize in Weights & Biases

llama2-docsum-adapter

This model is a fine-tuned version of NousResearch/Llama-2-7b-hf on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.2717

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: 4
  • eval_batch_size: 8
  • 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: cosine
  • lr_scheduler_warmup_ratio: 0.05
  • num_epochs: 2
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
1.9261 0.4 8 1.6258
1.3702 0.8 16 1.3632
1.2744 1.2 24 1.3249
1.0866 1.6 32 1.2929
1.1325 2.0 40 1.2717

Framework versions

  • PEFT 0.12.0
  • Transformers 4.42.3
  • Pytorch 2.1.2
  • Datasets 2.20.0
  • Tokenizers 0.19.1