b75eaefa82e1fea25e6eb03fa26a56e5

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B on the nyu-mll/glue [qqp] dataset. It achieves the following results on the evaluation set:

• Loss: 2.2841
• Data Size: 1.0
• Epoch Runtime: 1952.5781
• Accuracy: 0.8662
• F1 Macro: 0.8587
• Rouge1: 0.8662
• Rouge2: 0.0
• Rougel: 0.8664
• Rougelsum: 0.8662

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	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	10.3129	0	67.8157	0.6294	0.3881	0.6293	0.0	0.6295	0.6292
6.8779	1	11370	2.5360	0.0078	82.7738	0.6682	0.5122	0.6680	0.0	0.6682	0.6680
1.9199	2	22740	1.8470	0.0156	96.9261	0.7777	0.7632	0.7778	0.0	0.7777	0.7777
1.897	3	34110	1.6423	0.0312	126.3305	0.8044	0.7928	0.8044	0.0	0.8044	0.8043
1.6645	4	45480	1.7090	0.0625	185.6664	0.8178	0.8034	0.8177	0.0	0.8178	0.8178
1.5283	5	56850	1.4800	0.125	304.4613	0.8295	0.8165	0.8295	0.0	0.8295	0.8294
1.4045	6	68220	1.3782	0.25	538.8266	0.8446	0.8365	0.8445	0.0	0.8447	0.8446
1.2849	7	79590	1.3944	0.5	1001.1672	0.8404	0.8353	0.8404	0.0	0.8405	0.8404
1.2144	8.0	90960	1.1959	1.0	1901.3346	0.8726	0.8632	0.8726	0.0	0.8726	0.8726
0.8439	9.0	102330	1.2658	1.0	1915.2505	0.8714	0.8641	0.8714	0.0	0.8714	0.8713
0.6733	10.0	113700	1.3954	1.0	1912.7384	0.8777	0.8680	0.8777	0.0	0.8778	0.8777
0.4956	11.0	125070	1.6140	1.0	1910.8574	0.8715	0.8621	0.8715	0.0	0.8716	0.8715
0.3946	12.0	136440	2.2841	1.0	1952.5781	0.8662	0.8587	0.8662	0.0	0.8664	0.8662

Framework versions

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