Edit model card

Phi-1.5 Summarization (LoRA)

This model is a fine-tuned version of microsoft/phi-1_5 on the scitldr dataset. It achieves the following results on the evaluation set:

  • Loss: 2.6242

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: 0.001
  • train_batch_size: 1
  • eval_batch_size: 1
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 2

Training results

Training Loss Epoch Step Validation Loss
2.6014 0.25 500 2.6496
2.5756 0.5 1000 2.6445
2.5945 0.75 1500 2.6291
2.5251 1.0 2000 2.6133
2.3196 1.26 2500 2.6370
2.2953 1.51 3000 2.6325
2.452 1.76 3500 2.6242

Framework versions

  • PEFT 0.9.0
  • Transformers 4.38.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
7
Inference Examples
Inference API (serverless) does not yet support peft models for this pipeline type.

Model tree for pkbiswas/Phi-1_5-Summarization-LoRa

Base model

microsoft/phi-1_5
Adapter
(202)
this model

Dataset used to train pkbiswas/Phi-1_5-Summarization-LoRa