Edit model card

SciBLOOM-ft-TweetAreas-ES

This model is a fine-tuned version of bigscience/bloom-560m on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4180
  • Roc Auc: 0.8398
  • Hamming Loss: 0.0450
  • F1 Score: 0.7555
  • Accuracy: 0.4712
  • Precision: 0.8527
  • Recall: 0.7085

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
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Roc Auc Hamming Loss F1 Score Accuracy Precision Recall
0.2275 1.0 747 0.3007 0.7245 0.0797 0.5268 0.2838 0.8290 0.4840
0.1338 2.0 1494 0.2027 0.7985 0.0611 0.6307 0.3788 0.7336 0.6296
0.1244 3.0 2241 0.1917 0.7985 0.0564 0.6552 0.4070 0.7901 0.6354
0.0459 4.0 2988 0.2264 0.8247 0.0535 0.7187 0.4110 0.8199 0.6832
0.046 5.0 3735 0.2932 0.8103 0.0541 0.6862 0.4003 0.8026 0.6552
0.0305 6.0 4482 0.3364 0.8318 0.0509 0.7236 0.4378 0.8015 0.7008
0.0075 7.0 5229 0.4112 0.8326 0.0482 0.7348 0.4418 0.8164 0.6929
0.001 8.0 5976 0.3984 0.8358 0.0466 0.7507 0.4538 0.8501 0.7022
0.0 9.0 6723 0.4134 0.8448 0.0454 0.7591 0.4712 0.8447 0.7198
0.0 10.0 7470 0.4180 0.8398 0.0450 0.7555 0.4712 0.8527 0.7085

Framework versions

  • Transformers 4.43.2
  • Pytorch 2.3.1+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1
Downloads last month
4
Safetensors
Model size
559M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for alecmontero/SciBLOOM-ft-TweetAreas-ES

Finetuned
(31)
this model