sentinelle-sentiment-detector

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

  • Loss: 0.4579

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: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • num_epochs: 3
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
0.6887 0.0889 500 0.6459
0.6096 0.1778 1000 0.6775
0.6015 0.2667 1500 0.6109
0.5830 0.3556 2000 0.5507
0.5903 0.4444 2500 0.5418
0.5977 0.5333 3000 0.5269
0.5592 0.6222 3500 0.5251
0.5307 0.7111 4000 0.5365
0.4979 0.8 4500 0.5297
0.4893 0.8889 5000 0.4966
0.5503 0.9778 5500 0.4785
0.4640 1.0667 6000 0.4970
0.4765 1.1556 6500 0.4951
0.4854 1.2444 7000 0.4848
0.4704 1.3333 7500 0.4708
0.4465 1.4222 8000 0.4798
0.4347 1.5111 8500 0.5201
0.4194 1.6 9000 0.4804
0.4528 1.6889 9500 0.4569
0.3792 1.7778 10000 0.4705
0.4289 1.8667 10500 0.4577
0.4470 1.9556 11000 0.4620
0.3964 2.0444 11500 0.4738
0.4094 2.1333 12000 0.4698
0.3436 2.2222 12500 0.4746
0.4208 2.3111 13000 0.4710
0.3772 2.4 13500 0.4821
0.4110 2.4889 14000 0.4608
0.3324 2.5778 14500 0.4774
0.4161 2.6667 15000 0.4563
0.4208 2.7556 15500 0.4611
0.3776 2.8444 16000 0.4609
0.4345 2.9333 16500 0.4579

Framework versions

  • Transformers 5.0.0
  • Pytorch 2.10.0+cu128
  • Datasets 4.0.0
  • Tokenizers 0.22.2
Downloads last month
11
Safetensors
Model size
0.3B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for vilfrid/sentinelle-sentiment-detector

Finetuned
(4,328)
this model