gemma_toxic_sequence_cls

This model is a fine-tuned version of unsloth/gemma-2-2b-it-bnb-4bit on the None dataset. It achieves the following results on the evaluation set:

• F1 Threshold 0.5: 0.8308
• Loss: 0.1107

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: 5e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 0.1
• training_steps: 500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	F1 Threshold 0.5	Validation Loss
0.5805	0.0533	25	0.0278	0.6418
0.3259	0.1066	50	0.5364	0.2144
0.2077	0.1599	75	0.7106	0.1802
0.0284	0.2132	100	0.7817	0.1314
0.0577	0.2665	125	0.7446	0.1616
0.1123	0.3198	150	0.8280	0.1025
0.0950	0.3731	175	0.8107	0.1164
0.1635	0.4264	200	0.8395	0.0897
0.0772	0.4797	225	0.8349	0.1131
0.1412	0.5330	250	0.8097	0.1102
0.1280	0.5864	275	0.8308	0.1107

Framework versions

• PEFT 0.21.1
• Transformers 5.17.0
• Pytorch 2.11.0+cu130
• Datasets 4.8.5
• Tokenizers 0.23.2