metadata

license: mit
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - accuracy
  - f1
model-index:
  - name: camembert-keyword-discriminator
    results: []

camembert-keyword-discriminator

This model is a fine-tuned version of camembert-base on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.2180
Precision: 0.6646
Recall: 0.7047
Accuracy: 0.9344
F1: 0.6841
Ent/precision: 0.7185
Ent/accuracy: 0.8157
Ent/f1: 0.7640
Con/precision: 0.5324
Con/accuracy: 0.4860
Con/f1: 0.5082

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: 8
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	Accuracy	F1	Ent/precision	Ent/accuracy	Ent/f1	Con/precision	Con/accuracy	Con/f1
0.2016	1.0	1875	0.1910	0.5947	0.7125	0.9243	0.6483	0.6372	0.8809	0.7395	0.4560	0.3806	0.4149
0.1454	2.0	3750	0.1632	0.6381	0.7056	0.9324	0.6701	0.6887	0.8291	0.7524	0.5064	0.4621	0.4833
0.1211	3.0	5625	0.1702	0.6703	0.6678	0.9343	0.6690	0.7120	0.7988	0.7529	0.5471	0.4094	0.4684
0.1021	4.0	7500	0.1745	0.6777	0.6708	0.9351	0.6742	0.7206	0.7956	0.7562	0.5557	0.4248	0.4815
0.0886	5.0	9375	0.1913	0.6540	0.7184	0.9340	0.6847	0.7022	0.8396	0.7648	0.5288	0.4795	0.5030
0.0781	6.0	11250	0.2021	0.6605	0.7132	0.9344	0.6858	0.7139	0.8258	0.7658	0.5293	0.4913	0.5096
0.0686	7.0	13125	0.2127	0.6539	0.7132	0.9337	0.6822	0.7170	0.8172	0.7638	0.5112	0.5083	0.5098
0.0667	8.0	15000	0.2180	0.6646	0.7047	0.9344	0.6841	0.7185	0.8157	0.7640	0.5324	0.4860	0.5082

Framework versions

Transformers 4.19.2
Pytorch 1.11.0+cu113
Datasets 2.2.2
Tokenizers 0.12.1