b74fb0e7168e28e2083fe6bba2016bc0

This model is a fine-tuned version of Qwen/Qwen2.5-1.5B on the nyu-mll/glue [qqp] dataset. It achieves the following results on the evaluation set:

• Loss: 1.8957
• Data Size: 1.0
• Epoch Runtime: 2438.1968
• Accuracy: 0.8753
• F1 Macro: 0.8655
• Rouge1: 0.8753
• Rouge2: 0.0
• Rougel: 0.8753
• Rougelsum: 0.8753

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	4.3786	0	78.3524	0.4785	0.4703	0.4787	0.0	0.4785	0.4785
3.728	1	11370	2.3613	0.0078	98.4689	0.7962	0.7906	0.7963	0.0	0.7963	0.7962
1.9838	2	22740	1.5804	0.0156	118.1508	0.8261	0.8108	0.8262	0.0	0.8260	0.8262
1.67	3	34110	1.4803	0.0312	158.1568	0.8263	0.8173	0.8263	0.0	0.8263	0.8262
1.3885	4	45480	1.7937	0.0625	233.7576	0.8330	0.8151	0.8330	0.0	0.8329	0.8331
1.5099	5	56850	1.4786	0.125	381.0814	0.8277	0.8092	0.8276	0.0	0.8276	0.8276
1.2663	6	68220	1.3431	0.25	679.0830	0.8482	0.8400	0.8481	0.0	0.8481	0.8482
1.2094	7	79590	1.2296	0.5	1277.5813	0.8650	0.8577	0.8650	0.0	0.8651	0.8649
1.1332	8.0	90960	1.2045	1.0	2482.9203	0.8748	0.8667	0.8748	0.0	0.8748	0.8748
0.7703	9.0	102330	1.2571	1.0	2473.7513	0.8776	0.8703	0.8776	0.0	0.8776	0.8776
0.5973	10.0	113700	1.3832	1.0	2437.7396	0.8799	0.8706	0.8799	0.0	0.8799	0.8799
0.4768	11.0	125070	1.6196	1.0	2435.5383	0.8786	0.8693	0.8787	0.0	0.8787	0.8786
0.3897	12.0	136440	1.8957	1.0	2438.1968	0.8753	0.8655	0.8753	0.0	0.8753	0.8753

Framework versions

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