b7679abaacc43d0db38b31fc29eff0ef

This model is a fine-tuned version of facebook/opt-350m on the contemmcm/hate-speech-and-offensive-language dataset. It achieves the following results on the evaluation set:

• Loss: 0.5992
• Data Size: 1.0
• Epoch Runtime: 96.7110
• Accuracy: 0.8994
• F1 Macro: 0.6939

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: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro
No log	0	0	1.3285	0	7.7873	0.3815	0.2867
No log	1	619	0.7768	0.0078	8.5595	0.7764	0.3410
No log	2	1238	0.4000	0.0156	9.6629	0.8697	0.5999
0.0143	3	1857	0.5389	0.0312	11.2834	0.8896	0.5911
0.0143	4	2476	0.3635	0.0625	14.3902	0.9056	0.6111
0.3242	5	3095	0.2940	0.125	20.8818	0.9054	0.6901
0.0273	6	3714	0.3748	0.25	31.1421	0.8951	0.6004
0.322	7	4333	0.2981	0.5	52.3742	0.8951	0.7477
0.2758	8.0	4952	0.2666	1.0	94.4678	0.9079	0.6299
0.2075	9.0	5571	0.3418	1.0	92.5680	0.8922	0.7437
0.1809	10.0	6190	0.3359	1.0	94.3792	0.9087	0.7705
0.1761	11.0	6809	0.3783	1.0	93.8172	0.8870	0.7256
0.124	12.0	7428	0.5992	1.0	96.7110	0.8994	0.6939

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1