Hugging Face's logo

cdreetz
/
codeparrot-ds2

Text Generation
Transformers
PyTorch
gpt2
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Community
3
codeparrot-ds2 / README.md
cdreetz's picture
cdreetz
Update README.md
f1423b8
preview code
|
raw
history blame
8.07 kB
metadata
license: mit
tags:
  - generated_from_trainer
model-index:
  - name: codeparrot-ds2
    results: []

codeparrot-ds2

This model is a fine-tuned version of gpt2 on a filtered version of The Stack. Filtered for Data Science related code. It achieves the following results on the evaluation set:

  • Loss: 1.0584

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: 0.0005
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 256
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 200
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss
2.2038 0.01 500 2.1062
2.0551 0.02 1000 2.0109
1.9622 0.02 1500 1.9219
1.9512 0.03 2000 1.8461
1.8817 0.04 2500 1.7903
1.8341 0.05 3000 1.7401
1.7877 0.05 3500 1.7022
1.7586 0.06 4000 1.6694
1.7271 0.07 4500 1.6457
1.7034 0.08 5000 1.6193
1.6756 0.08 5500 1.5978
1.6576 0.09 6000 1.5772
1.6377 0.1 6500 1.5611
1.6211 0.11 7000 1.5453
1.6033 0.11 7500 1.5317
1.591 0.12 8000 1.5193
1.5765 0.13 8500 1.5053
1.5661 0.14 9000 1.4966
1.5548 0.15 9500 1.4846
1.5429 0.15 10000 1.4729
1.5347 0.16 10500 1.4641
1.5215 0.17 11000 1.4557
1.5151 0.18 11500 1.4454
1.5059 0.18 12000 1.4381
1.499 0.19 12500 1.4288
1.4906 0.2 13000 1.4210
1.4849 0.21 13500 1.4143
1.4765 0.21 14000 1.4085
1.4708 0.22 14500 1.4026
1.4602 0.23 15000 1.3936
1.4533 0.24 15500 1.3896
1.4523 0.25 16000 1.3818
1.4415 0.25 16500 1.3748
1.4417 0.26 17000 1.3701
1.4311 0.27 17500 1.3645
1.4282 0.28 18000 1.3585
1.4223 0.28 18500 1.3531
1.4165 0.29 19000 1.3473
1.4105 0.3 19500 1.3419
1.3993 0.31 20000 1.3374
1.4034 0.31 20500 1.3322
1.3982 0.32 21000 1.3278
1.3951 0.33 21500 1.3225
1.3806 0.34 22000 1.3180
1.3781 0.34 22500 1.3121
1.3761 0.35 23000 1.3082
1.3662 0.36 23500 1.3038
1.3631 0.37 24000 1.2995
1.3549 0.38 24500 1.2955
1.3577 0.38 25000 1.2912
1.3498 0.39 25500 1.2851
1.3428 0.4 26000 1.2807
1.342 0.41 26500 1.2768
1.3365 0.41 27000 1.2720
1.3313 0.42 27500 1.2678
1.3309 0.43 28000 1.2629
1.3221 0.44 28500 1.2594
1.3214 0.44 29000 1.2558
1.3099 0.45 29500 1.2510
1.31 0.46 30000 1.2449
1.31 0.47 30500 1.2414
1.305 0.48 31000 1.2390
1.2975 0.48 31500 1.2358
1.2882 0.49 32000 1.2311
1.2831 0.5 32500 1.2251
1.2836 0.51 33000 1.2212
1.2817 0.51 33500 1.2178
1.2772 0.52 34000 1.2130
1.2651 0.53 34500 1.2080
1.2683 0.54 35000 1.2048
1.2581 0.54 35500 1.1999
1.263 0.55 36000 1.1972
1.255 0.56 36500 1.1924
1.2466 0.57 37000 1.1884
1.2448 0.57 37500 1.1860
1.2413 0.58 38000 1.1804
1.2362 0.59 38500 1.1782
1.2309 0.6 39000 1.1732
1.2289 0.61 39500 1.1687
1.2208 0.61 40000 1.1649
1.2225 0.62 40500 1.1605
1.2178 0.63 41000 1.1555
1.208 0.64 41500 1.1533
1.2069 0.64 42000 1.1490
1.206 0.65 42500 1.1453
1.2013 0.66 43000 1.1414
1.2003 0.67 43500 1.1374
1.1867 0.67 44000 1.1337
1.187 0.68 44500 1.1302
1.188 0.69 45000 1.1270
1.179 0.7 45500 1.1237
1.1866 0.71 46000 1.1204
1.173 0.71 46500 1.1173
1.1706 0.72 47000 1.1134
1.1645 0.73 47500 1.1099
1.1641 0.74 48000 1.1063
1.1623 0.74 48500 1.1032
1.1561 0.75 49000 1.1006
1.1531 0.76 49500 1.0977
1.1569 0.77 50000 1.0950
1.1505 0.77 50500 1.0927
1.1473 0.78 51000 1.0902
1.1428 0.79 51500 1.0870
1.1412 0.8 52000 1.0844
1.1452 0.8 52500 1.0823
1.1391 0.81 53000 1.0805
1.1329 0.82 53500 1.0783
1.1295 0.83 54000 1.0764
1.125 0.84 54500 1.0746
1.1295 0.84 55000 1.0730
1.1247 0.85 55500 1.0711
1.1225 0.86 56000 1.0696
1.1235 0.87 56500 1.0680
1.1192 0.87 57000 1.0670
1.1189 0.88 57500 1.0654
1.1196 0.89 58000 1.0646
1.1152 0.9 58500 1.0635
1.1133 0.9 59000 1.0628
1.1126 0.91 59500 1.0619
1.1142 0.92 60000 1.0610
1.1112 0.93 60500 1.0605
1.1137 0.93 61000 1.0599
1.1127 0.94 61500 1.0595
1.1111 0.95 62000 1.0592
1.1121 0.96 62500 1.0588
1.1114 0.97 63000 1.0587
1.1121 0.97 63500 1.0585
1.1078 0.98 64000 1.0584
1.1104 0.99 64500 1.0584
1.1057 1.0 65000 1.0584

Framework versions

  • Transformers 4.30.2
  • Pytorch 1.13.1
  • Datasets 2.13.1
  • Tokenizers 0.13.3