Hugging Face's logo

HiTZ
/
alpaca-lora-7b-en-pt-es-ca-eu-gl-at

Generated from Trainer
Model card Files Community
1
alpaca-lora-7b-en-pt-es-ca-eu-gl-at / README.md
juletxara's picture
juletxara
update model card README.md
0cf4b18
preview code
|
raw
history blame
2.84 kB
metadata
license: other
tags:
  - generated_from_trainer
datasets:
  - HiTZ/alpaca_mt
model-index:
  - name: alpaca-lora-7b-en-pt-es-ca-eu-gl-at
    results: []

alpaca-lora-7b-en-pt-es-ca-eu-gl-at

This model is a fine-tuned version of decapoda-research/llama-7b-hf on the HiTZ/alpaca_mt ['en', 'pt', 'es', 'ca', 'eu', 'gl', 'at'] dataset. It achieves the following results on the evaluation set:

  • Loss: 1.0667

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: 0.0003
  • train_batch_size: 26
  • eval_batch_size: 26
  • seed: 42
  • distributed_type: multi-GPU
  • gradient_accumulation_steps: 5
  • total_train_batch_size: 130
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.03
  • num_epochs: 1
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
1.3772 0.04 100 1.3860
1.3043 0.07 200 1.2904
1.2307 0.11 300 1.2409
1.2132 0.15 400 1.2086
1.1987 0.19 500 1.1854
1.1551 0.22 600 1.1660
1.1613 0.26 700 1.1516
1.144 0.3 800 1.1407
1.1494 0.34 900 1.1297
1.1072 0.37 1000 1.1196
1.1302 0.41 1100 1.1117
1.1074 0.45 1200 1.1058
1.0846 0.48 1300 1.0995
1.086 0.52 1400 1.0935
1.0793 0.56 1500 1.0889
1.0931 0.6 1600 1.0847
1.0905 0.63 1700 1.0804
1.0793 0.67 1800 1.0775
1.0795 0.71 1900 1.0748
1.0861 0.74 2000 1.0725
1.0881 0.78 2100 1.0705
1.0673 0.82 2200 1.0691
1.0626 0.86 2300 1.0681
1.0633 0.89 2400 1.0674
1.0601 0.93 2500 1.0669
1.0849 0.97 2600 1.0667

Framework versions

  • Transformers 4.28.0.dev0
  • Pytorch 2.0.0+cu117
  • Datasets 2.10.1
  • Tokenizers 0.13.2