74bf17130fefacfe9228f56d6e374046

This model is a fine-tuned version of facebook/opt-6.7b on the contemmcm/hate-speech-and-offensive-language dataset. It achieves the following results on the evaluation set:

• Loss: 0.7212
• Data Size: 1.0
• Epoch Runtime: 265.1607
• Accuracy: 0.8791
• F1 Macro: 0.7034

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	2.2128	0	13.0154	0.1343	0.1359
No log	1	619	0.7801	0.0078	15.1143	0.7695	0.3096
No log	2	1238	0.4857	0.0156	28.0280	0.8711	0.6139
0.0207	3	1857	0.8112	0.0312	41.0913	0.7906	0.5016
0.0207	4	2476	0.3923	0.0625	58.7204	0.8801	0.5813
0.409	5	3095	0.4174	0.125	84.0083	0.8496	0.6809
0.0311	6	3714	0.3040	0.25	91.7113	0.9016	0.6255
0.2984	7	4333	0.2914	0.5	153.3875	0.9006	0.6861
0.2593	8.0	4952	0.3033	1.0	280.1670	0.9069	0.6260
0.1486	9.0	5571	0.3423	1.0	265.6656	0.8872	0.7309
0.0971	10.0	6190	0.5307	1.0	269.9368	0.8949	0.7203
0.063	11.0	6809	0.7212	1.0	265.1607	0.8791	0.7034

Framework versions

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