Edit model card

multilingual_minilm-amazon-massive-intent

This model is a fine-tuned version of microsoft/Multilingual-MiniLM-L12-H384 on the MASSIVE1.1 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.8941
  • Accuracy: 0.8234
  • F1: 0.8234

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: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 15

Training results

Training Loss Epoch Step Validation Loss Accuracy F1
3.7961 1.0 720 3.1657 0.3404 0.3404
3.1859 2.0 1440 2.4835 0.4343 0.4343
2.3104 3.0 2160 2.0474 0.5652 0.5652
2.0071 4.0 2880 1.7190 0.6503 0.6503
1.5595 5.0 3600 1.4873 0.6990 0.6990
1.3664 6.0 4320 1.3088 0.7354 0.7354
1.1272 7.0 5040 1.1964 0.7521 0.7521
1.0128 8.0 5760 1.1115 0.7718 0.7718
0.9405 9.0 6480 1.0598 0.7841 0.7841
0.7758 10.0 7200 1.0003 0.7944 0.7944
0.7457 11.0 7920 0.9599 0.8037 0.8037
0.6605 12.0 8640 0.9175 0.8165 0.8165
0.6135 13.0 9360 0.9148 0.8190 0.8190
0.5698 14.0 10080 0.8976 0.8229 0.8229
0.5578 15.0 10800 0.8941 0.8234 0.8234

Framework versions

  • Transformers 4.25.1
  • Pytorch 1.13.0+cu116
  • Datasets 2.7.1
  • Tokenizers 0.13.2
Downloads last month
14
Safetensors
Model size
118M params
Tensor type
I64
·
F32
·

Finetuned from

Dataset used to train cartesinus/multilingual_minilm-amazon-massive-intent

Evaluation results