marbert-arabic-tweet-sentiment-pro

This model is a fine-tuned version of UBC-NLP/MARBERTv2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.9228
• Accuracy: 0.7438
• F1: 0.7488

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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: 100
• num_epochs: 7
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0	58	0.9408	0.5710	0.4688
No log	2.0	116	0.7689	0.6728	0.6424
No log	3.0	174	0.7542	0.7284	0.7161
No log	4.0	232	0.7831	0.7654	0.7667
No log	5.0	290	0.8615	0.7623	0.7635
No log	6.0	348	0.9228	0.7438	0.7488

Framework versions

• Transformers 5.15.1
• Pytorch 2.11.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2