c5da44fd38210e9447aeb7ce6d0204b3

This model is a fine-tuned version of Qwen/Qwen2.5-1.5B on the contemmcm/cls_20newsgroups dataset. It achieves the following results on the evaluation set:

• Loss: 1.9249
• Data Size: 1.0
• Epoch Runtime: 359.3043
• Accuracy: 0.9007
• F1 Macro: 0.8990

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	24.6036	0	25.4824	0.0542	0.0177
No log	1	499	15.6952	0.0078	28.7425	0.0930	0.0664
0.1856	2	998	6.6139	0.0156	32.7025	0.5658	0.5337
0.1748	3	1497	3.6692	0.0312	39.1321	0.7427	0.7437
0.1322	4	1996	3.5274	0.0625	50.8931	0.7641	0.7552
2.7498	5	2495	2.2999	0.125	72.8615	0.8264	0.8254
1.7977	6	2994	1.8570	0.25	117.0181	0.8687	0.8664
1.3449	7	3493	1.6318	0.5	197.9701	0.8816	0.8817
1.0358	8.0	3992	1.3887	1.0	358.9648	0.8886	0.8885
0.8312	9.0	4491	1.7446	1.0	359.7488	0.8914	0.8926
0.5715	10.0	4990	1.7201	1.0	360.2899	0.9032	0.9029
0.6476	11.0	5489	1.9233	1.0	359.1780	0.8999	0.8990
0.4213	12.0	5988	1.9249	1.0	359.3043	0.9007	0.8990

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1