a2daee8cde1a18f649abe542a2f6695f

This model is a fine-tuned version of meta-llama/Llama-3.2-3B on the contemmcm/cls_mmlu dataset. It achieves the following results on the evaluation set:

• Loss: 6.4439
• Data Size: 1.0
• Epoch Runtime: 196.7443
• Accuracy: 0.2640
• F1 Macro: 0.2303

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	10.4197	0	4.9812	0.2440	0.2043
No log	1	438	11.6142	0.0078	6.4001	0.2640	0.1927
No log	2	876	7.3212	0.0156	12.7673	0.2527	0.1008
No log	3	1314	5.8588	0.0312	18.8025	0.2527	0.1008
No log	4	1752	5.8624	0.0625	29.0361	0.2540	0.1246
0.3875	5	2190	5.7226	0.125	42.1571	0.2547	0.1052
0.758	6	2628	5.8211	0.25	67.7670	0.2566	0.1844
5.6509	7	3066	5.6819	0.5	111.9252	0.2434	0.1015
5.6199	8.0	3504	5.5830	1.0	199.7878	0.2527	0.1008
5.6166	9.0	3942	5.7825	1.0	198.9046	0.2533	0.1011
5.4415	10.0	4380	5.8796	1.0	194.0456	0.2646	0.1614
5.2183	11.0	4818	5.9236	1.0	194.4793	0.2892	0.1927
4.5755	12.0	5256	6.4439	1.0	196.7443	0.2640	0.2303

Framework versions

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