SetFit / distilbert-base-uncased__sst2__train-8-6

huggingface.co
Total runs: 13
24-hour runs: 0
7-day runs: -4
30-day runs: -68
Model's Last Updated: February 10 2022
text-classification

Introduction of distilbert-base-uncased__sst2__train-8-6

Model Details of distilbert-base-uncased__sst2__train-8-6

distilbert-base-uncased__sst2__train-8-6

This model is a fine-tuned version of distilbert-base-uncased on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5336
  • Accuracy: 0.7523
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: 2e-05
  • train_batch_size: 4
  • eval_batch_size: 4
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 50
  • mixed_precision_training: Native AMP
Training results
Training Loss Epoch Step Validation Loss Accuracy
0.7161 1.0 3 0.6941 0.5
0.6786 2.0 6 0.7039 0.25
0.6586 3.0 9 0.7090 0.25
0.6121 4.0 12 0.7183 0.25
0.5696 5.0 15 0.7266 0.25
0.522 6.0 18 0.7305 0.25
0.4899 7.0 21 0.7339 0.25
0.3985 8.0 24 0.7429 0.25
0.3758 9.0 27 0.7224 0.25
0.2876 10.0 30 0.7068 0.5
0.2498 11.0 33 0.6751 0.75
0.1921 12.0 36 0.6487 0.75
0.1491 13.0 39 0.6261 0.75
0.1276 14.0 42 0.6102 0.75
0.0996 15.0 45 0.5964 0.75
0.073 16.0 48 0.6019 0.75
0.0627 17.0 51 0.5933 0.75
0.053 18.0 54 0.5768 0.75
0.0403 19.0 57 0.5698 0.75
0.0328 20.0 60 0.5656 0.75
0.03 21.0 63 0.5634 0.75
0.025 22.0 66 0.5620 0.75
0.0209 23.0 69 0.5623 0.75
0.0214 24.0 72 0.5606 0.75
0.0191 25.0 75 0.5565 0.75
0.0173 26.0 78 0.5485 0.75
0.0175 27.0 81 0.5397 0.75
0.0132 28.0 84 0.5322 0.75
0.0138 29.0 87 0.5241 0.75
0.0128 30.0 90 0.5235 0.75
0.0126 31.0 93 0.5253 0.75
0.012 32.0 96 0.5317 0.75
0.0118 33.0 99 0.5342 0.75
0.0092 34.0 102 0.5388 0.75
0.0117 35.0 105 0.5414 0.75
0.0124 36.0 108 0.5453 0.75
0.0109 37.0 111 0.5506 0.75
0.0112 38.0 114 0.5555 0.75
0.0087 39.0 117 0.5597 0.75
0.01 40.0 120 0.5640 0.75
Framework versions
  • Transformers 4.15.0
  • Pytorch 1.10.2+cu102
  • Datasets 1.18.2
  • Tokenizers 0.10.3

Runs of SetFit distilbert-base-uncased__sst2__train-8-6 on huggingface.co

13
Total runs
0
24-hour runs
-2
3-day runs
-4
7-day runs
-68
30-day runs

More Information About distilbert-base-uncased__sst2__train-8-6 huggingface.co Model

More distilbert-base-uncased__sst2__train-8-6 license Visit here:

https://choosealicense.com/licenses/apache-2.0

distilbert-base-uncased__sst2__train-8-6 huggingface.co

distilbert-base-uncased__sst2__train-8-6 huggingface.co is an AI model on huggingface.co that provides distilbert-base-uncased__sst2__train-8-6's model effect (), which can be used instantly with this SetFit distilbert-base-uncased__sst2__train-8-6 model. huggingface.co supports a free trial of the distilbert-base-uncased__sst2__train-8-6 model, and also provides paid use of the distilbert-base-uncased__sst2__train-8-6. Support call distilbert-base-uncased__sst2__train-8-6 model through api, including Node.js, Python, http.

distilbert-base-uncased__sst2__train-8-6 huggingface.co Url

https://huggingface.co/SetFit/distilbert-base-uncased__sst2__train-8-6

SetFit distilbert-base-uncased__sst2__train-8-6 online free

distilbert-base-uncased__sst2__train-8-6 huggingface.co is an online trial and call api platform, which integrates distilbert-base-uncased__sst2__train-8-6's modeling effects, including api services, and provides a free online trial of distilbert-base-uncased__sst2__train-8-6, you can try distilbert-base-uncased__sst2__train-8-6 online for free by clicking the link below.

SetFit distilbert-base-uncased__sst2__train-8-6 online free url in huggingface.co:

https://huggingface.co/SetFit/distilbert-base-uncased__sst2__train-8-6

distilbert-base-uncased__sst2__train-8-6 install

distilbert-base-uncased__sst2__train-8-6 is an open source model from GitHub that offers a free installation service, and any user can find distilbert-base-uncased__sst2__train-8-6 on GitHub to install. At the same time, huggingface.co provides the effect of distilbert-base-uncased__sst2__train-8-6 install, users can directly use distilbert-base-uncased__sst2__train-8-6 installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

distilbert-base-uncased__sst2__train-8-6 install url in huggingface.co:

https://huggingface.co/SetFit/distilbert-base-uncased__sst2__train-8-6

Url of distilbert-base-uncased__sst2__train-8-6

distilbert-base-uncased__sst2__train-8-6 huggingface.co Url

Provider of distilbert-base-uncased__sst2__train-8-6 huggingface.co

SetFit
ORGANIZATIONS

Other API from SetFit