camembert_ccnet_classification_tools_NEFTune_fr_V2

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

• Loss: 0.5108
• Accuracy: 0.9062
• Learning Rate: 0.0001

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: 0.0001
• train_batch_size: 24
• eval_batch_size: 192
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 60

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Rate
1.802	1.0	15	1.3063	0.7708	0.0001
0.9616	2.0	30	0.7143	0.8438	0.0001
0.4359	3.0	45	0.3769	0.9271	0.0001
0.2292	4.0	60	0.3546	0.9167	0.0001
0.1448	5.0	75	0.2678	0.9479	0.0001
0.095	6.0	90	0.4425	0.9062	9e-05
0.0762	7.0	105	0.3686	0.9062	0.0001
0.0817	8.0	120	0.4784	0.9062	0.0001
0.0506	9.0	135	0.4753	0.8958	0.0001
0.0245	10.0	150	0.3736	0.9167	0.0001
0.0347	11.0	165	0.5036	0.9062	0.0001
0.0141	12.0	180	0.4478	0.9167	8e-05
0.0196	13.0	195	0.4295	0.9167	0.0001
0.009	14.0	210	0.3942	0.9167	0.0001
0.0076	15.0	225	0.5108	0.9062	0.0001

Framework versions

• Transformers 4.34.0
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.14.1