GaetanMichelet's picture
End of training
d4418c3 verified
metadata
base_model: google/gemma-2-2b-it
datasets:
  - GaetanMichelet/chat-60_ft_task-3
  - GaetanMichelet/chat-120_ft_task-3
  - GaetanMichelet/chat-180_ft_task-3
library_name: peft
license: gemma
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
model-index:
  - name: Gemma-2-2B_task-3_180-samples_config-2_full
    results: []

Gemma-2-2B_task-3_180-samples_config-2_full

This model is a fine-tuned version of google/gemma-2-2b-it on the GaetanMichelet/chat-60_ft_task-3, the GaetanMichelet/chat-120_ft_task-3 and the GaetanMichelet/chat-180_ft_task-3 datasets. It achieves the following results on the evaluation set:

  • Loss: 0.9456

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: 1
  • eval_batch_size: 1
  • seed: 42
  • distributed_type: multi-GPU
  • gradient_accumulation_steps: 16
  • 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: 0.1
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss
1.3981 0.9412 8 1.3785
1.3072 2.0 17 1.2429
1.1985 2.9412 25 1.1442
0.9971 4.0 34 1.0312
0.9268 4.9412 42 0.9882
0.9442 6.0 51 0.9653
0.9253 6.9412 59 0.9537
0.8684 8.0 68 0.9479
0.8043 8.9412 76 0.9456
0.7924 10.0 85 0.9502
0.7535 10.9412 93 0.9591
0.694 12.0 102 0.9863
0.6881 12.9412 110 0.9994
0.6566 14.0 119 1.0534
0.5597 14.9412 127 1.1117
0.497 16.0 136 1.1691

Framework versions

  • PEFT 0.12.0
  • Transformers 4.44.0
  • Pytorch 2.1.2+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1