lemexp-task1-v2-template_full_notypes-deepseek-coder-1.3b-base-ddp-8lr-v2

This model is a fine-tuned version of deepseek-ai/deepseek-coder-1.3b-base on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1580

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: 0.0008
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 8
  • total_train_batch_size: 16
  • total_eval_batch_size: 16
  • optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • num_epochs: 12
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
0.323 0.2 3094 0.3202
0.298 0.4 6188 0.3000
0.2894 0.6 9282 0.2873
0.2822 0.8 12376 0.2833
0.277 1.0 15470 0.2830
0.2735 1.2 18564 0.2703
0.2697 1.4 21658 0.2622
0.2644 1.6 24752 0.2595
0.2639 1.8 27846 0.2525
0.259 2.0 30940 0.2543
0.2525 2.2 34034 0.2585
0.2527 2.4 37128 0.2484
0.2479 2.6 40222 0.2459
0.2474 2.8 43316 0.2459
0.2446 3.0 46410 0.2534
0.2406 3.2 49504 0.2390
0.2406 3.4 52598 0.2351
0.236 3.6 55692 0.2347
0.2342 3.8 58786 0.2295
0.235 4.0 61880 0.2346
0.2275 4.2 64974 0.2235
0.2234 4.4 68068 0.2277
0.2231 4.6 71162 0.2263
0.2181 4.8 74256 0.2214
0.2177 5.0 77350 0.2195
0.2153 5.2 80444 0.2148
0.2134 5.4 83538 0.2133
0.2115 5.6 86632 0.2122
0.2102 5.8 89726 0.2129
0.2063 6.0 92820 0.2095
0.2021 6.2 95914 0.2089
0.2007 6.4 99008 0.2052
0.2002 6.6 102102 0.2038
0.2011 6.8 105196 0.1991
0.1965 7.0 108290 0.1989
0.1892 7.2 111384 0.1965
0.1871 7.4 114478 0.1933
0.1891 7.6 117572 0.1976
0.1866 7.8 120666 0.1919
0.1856 8.0 123760 0.1932
0.1757 8.2 126854 0.1914
0.1758 8.4 129948 0.1854
0.1739 8.6 133042 0.1827
0.1772 8.8 136136 0.1812
0.1746 9.0 139230 0.1789
0.1653 9.2 142324 0.1767
0.165 9.4 145418 0.1739
0.1644 9.6 148512 0.1730
0.163 9.8 151606 0.1720
0.1587 10.0 154700 0.1699
0.1536 10.2 157794 0.1684
0.1508 10.4 160888 0.1662
0.1516 10.6 163982 0.1665
0.1494 10.8 167076 0.1640
0.1494 11.0 170170 0.1621
0.1419 11.2 173264 0.1627
0.1388 11.4 176358 0.1603
0.1384 11.6 179452 0.1588
0.1376 11.8 182546 0.1583
0.1387 12.0 185640 0.1580

Framework versions

  • PEFT 0.14.0
  • Transformers 4.47.0
  • Pytorch 2.5.1+cu124
  • Datasets 3.2.0
  • Tokenizers 0.21.0
Downloads last month
2
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for yalhessi/lemexp-task1-v2-template_full_notypes-deepseek-coder-1.3b-base-ddp-8lr-v2

Adapter
(168)
this model