bangla-meme-sarcasm-classifier-checkpoints

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4459
• Accuracy: 0.8005
• F1: 0.8182

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
• lr_scheduler_warmup_steps: 0.1
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.4709	1.0	1282	0.4734	0.7878	0.7749
0.4092	2.0	2564	0.4459	0.8005	0.8182
0.2948	3.0	3846	0.5369	0.7960	0.7846
0.2534	4.0	5128	0.7913	0.7923	0.7795
0.1447	5.0	6410	0.9074	0.7863	0.7677
0.0838	6.0	7692	1.0915	0.7882	0.7732
0.0989	7.0	8974	1.1061	0.7970	0.8020
0.0115	8.0	10256	1.6093	0.7693	0.7329
0.0615	9.0	11538	1.4467	0.7968	0.7948
0.0004	10.0	12820	1.4821	0.7964	0.7920
0.0027	11.0	14102	1.6057	0.7958	0.7926
0.0266	12.0	15384	1.6298	0.7956	0.7976
0.1105	13.0	16666	1.7722	0.7908	0.7785
0.0001	14.0	17948	1.8275	0.7925	0.7794
0.0021	15.0	19230	1.8582	0.7919	0.7783

Framework versions

• Transformers 5.16.1
• Pytorch 2.11.0+cu128
• Datasets 4.0.0
• Tokenizers 0.23.1