Edit model card

Artigenz-Coder-DS-6.7B_components_FI_dataset_size_52_epochs_10_2024-06-12_23-06-16_3525893

This model is a fine-tuned version of Artigenz/Artigenz-Coder-DS-6.7B on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2550
  • Accuracy: 0.451
  • Chrf: 0.858
  • Bleu: 0.787
  • Sacrebleu: 0.8
  • Rouge1: 0.839
  • Rouge2: 0.735
  • Rougel: 0.823
  • Rougelsum: 0.835
  • Meteor: 0.851

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: 3407
  • distributed_type: multi-GPU
  • num_devices: 4
  • total_train_batch_size: 4
  • total_eval_batch_size: 4
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-06
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 52
  • training_steps: 520
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy Chrf Bleu Sacrebleu Rouge1 Rouge2 Rougel Rougelsum Meteor
0.004 0.83 52 0.3624 0.452 0.815 0.729 0.7 0.799 0.673 0.779 0.797 0.802
0.0169 1.65 104 0.3225 0.452 0.829 0.739 0.7 0.808 0.691 0.792 0.807 0.81
0.0277 2.48 156 0.2859 0.451 0.847 0.768 0.8 0.827 0.72 0.814 0.825 0.837
0.0088 3.3 208 0.3294 0.45 0.825 0.744 0.7 0.81 0.7 0.792 0.807 0.814
0.2337 4.13 260 0.2829 0.451 0.845 0.764 0.8 0.82 0.705 0.802 0.817 0.839
0.0027 4.95 312 0.2702 0.451 0.848 0.772 0.8 0.831 0.724 0.816 0.828 0.838
0.0154 5.78 364 0.2639 0.451 0.852 0.777 0.8 0.834 0.729 0.818 0.831 0.845
0.0337 6.6 416 0.2625 0.451 0.854 0.78 0.8 0.828 0.716 0.813 0.825 0.852
0.0147 7.43 468 0.2568 0.451 0.855 0.785 0.8 0.831 0.719 0.815 0.828 0.857
0.0042 8.25 520 0.2550 0.451 0.858 0.787 0.8 0.839 0.735 0.823 0.835 0.851

Framework versions

  • PEFT 0.7.1
  • Transformers 4.37.0
  • Pytorch 2.2.1+cu121
  • Datasets 2.19.2
  • Tokenizers 0.15.2
Downloads last month
0
Safetensors
Model size
6.74B params
Tensor type
F32
·
Inference API
Unable to determine this model’s pipeline type. Check the docs .

Model tree for vdavidr/Artigenz-Coder-DS-6.7B_components_FI_dataset_size_52_epochs_10_2024-06-12_23-06-16_3525893

Adapter
(4)
this model