9890684221bde3f904efa02194dfb2a8

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

• Loss: 3.2228
• Data Size: 1.0
• Epoch Runtime: 2524.6273
• Accuracy: 0.7819
• F1 Macro: 0.7816
• Rouge1: 0.7818
• Rouge2: 0.0
• Rougel: 0.7821
• Rougelsum: 0.7820

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	14.7774	0	19.6523	0.3180	0.2229	0.3182	0.0	0.3180	0.3183
4.7322	1	12271	3.0515	0.0078	38.7396	0.6905	0.6898	0.6906	0.0	0.6906	0.6907
2.5079	2	24542	2.3675	0.0156	60.2594	0.7611	0.7618	0.7612	0.0	0.7611	0.7612
2.3241	3	36813	2.9412	0.0312	100.5598	0.7551	0.7505	0.7547	0.0	0.7553	0.7553
2.2899	4	49084	2.3337	0.0625	177.1881	0.7674	0.7677	0.7672	0.0	0.7673	0.7673
2.1197	5	61355	2.3608	0.125	331.0613	0.7590	0.7558	0.7589	0.0	0.7588	0.7589
2.2715	6	73626	2.2753	0.25	645.7571	0.7703	0.7702	0.7702	0.0	0.7704	0.7704
2.0129	7	85897	2.2204	0.5	1259.9502	0.7744	0.7741	0.7744	0.0	0.7746	0.7745
1.8546	8.0	98168	2.1504	1.0	2495.7308	0.7907	0.7907	0.7906	0.0	0.7906	0.7907
1.397	9.0	110439	2.2389	1.0	2494.4240	0.7926	0.7918	0.7924	0.0	0.7926	0.7927
1.1675	10.0	122710	2.3874	1.0	2512.6537	0.7872	0.7865	0.7872	0.0	0.7874	0.7871
0.8895	11.0	134981	2.6973	1.0	2528.5981	0.7884	0.7874	0.7882	0.0	0.7888	0.7887
0.7409	12.0	147252	3.2228	1.0	2524.6273	0.7819	0.7816	0.7818	0.0	0.7821	0.7820

Framework versions

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