pythia160m-sft-tldr / README.md
mnoukhov's picture
Model save
40c0022 verified
metadata
license: apache-2.0
base_model: EleutherAI/pythia-160m-deduped
tags:
  - trl
  - sft
  - generated_from_trainer
model-index:
  - name: pythia160m-sft-tldr
    results: []

pythia160m-sft-tldr

This model is a fine-tuned version of EleutherAI/pythia-160m-deduped on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 2.7556

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: 1e-05
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 4
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 128
  • total_eval_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • num_epochs: 2.0

Training results

Training Loss Epoch Step Validation Loss
2.8454 0.4002 365 2.8609
2.7873 0.8004 730 2.7859
2.7641 1.2007 1095 2.7616
2.7486 1.6009 1460 2.7556

Framework versions

  • Transformers 4.41.1
  • Pytorch 2.2.1+cu121
  • Datasets 2.19.0
  • Tokenizers 0.19.1