KrithikV's picture
HuggingFaceUser/adapter-medmobile-MEDQA-Extended-V2
f605ec5 verified
|
raw
history blame
2.12 kB
metadata
base_model: microsoft/Phi-3-mini-4k-instruct
library_name: peft
license: mit
tags:
  - trl
  - sft
  - generated_from_trainer
model-index:
  - name: phi-3-mini-LoRA-MEDQA-Extended-V2
    results: []

phi-3-mini-LoRA-MEDQA-Extended-V2

This model is a fine-tuned version of microsoft/Phi-3-mini-4k-instruct on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6514

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: 32
  • eval_batch_size: 32
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss
0.7812 0.2205 100 0.6947
0.6795 0.4410 200 0.6694
0.6686 0.6615 300 0.6633
0.662 0.8820 400 0.6597
0.6603 1.1025 500 0.6578
0.6586 1.3230 600 0.6566
0.6538 1.5436 700 0.6552
0.6551 1.7641 800 0.6539
0.6542 1.9846 900 0.6535
0.65 2.2051 1000 0.6525
0.6522 2.4256 1100 0.6519
0.6526 2.6461 1200 0.6515
0.6501 2.8666 1300 0.6514

Framework versions

  • PEFT 0.12.0
  • Transformers 4.43.3
  • Pytorch 2.3.1+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1