43bd7e74faf21e9f15deaaad494b15d5

This model is a fine-tuned version of deepseek-ai/DeepSeek-R1-Distill-Qwen-7B on the nyu-mll/glue [mrpc] dataset. It achieves the following results on the evaluation set:

• Loss: 5.8488
• Data Size: 1.0
• Epoch Runtime: 120.6807
• Accuracy: 0.7995
• F1 Macro: 0.7842
• Rouge1: 0.7995
• Rouge2: 0.0
• Rougel: 0.7989
• Rougelsum: 0.7995

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	7.4657	0	5.7654	0.6604	0.4107	0.6616	0.0	0.6604	0.6604
No log	1	114	79.7619	0.0078	6.1570	0.3349	0.2509	0.3343	0.0	0.3355	0.3349
No log	2	228	37.1240	0.0156	18.8215	0.3349	0.2509	0.3343	0.0	0.3355	0.3349
No log	3	342	5.1122	0.0312	30.8926	0.6792	0.4787	0.6792	0.0	0.6787	0.6787
0.4304	4	456	3.2691	0.0625	39.4693	0.6462	0.6327	0.6462	0.0	0.6456	0.6468
0.4304	5	570	1.8845	0.125	48.9782	0.7983	0.7760	0.7983	0.0	0.7983	0.7983
0.4304	6	684	2.8282	0.25	60.8006	0.6291	0.6291	0.6285	0.0	0.6285	0.6279
0.5711	7	798	1.5321	0.5	75.4529	0.8296	0.8123	0.8296	0.0	0.8290	0.8296
1.0265	8.0	912	1.8242	1.0	119.6966	0.8031	0.7856	0.8031	0.0	0.8037	0.8025
0.4925	9.0	1026	3.1007	1.0	117.1617	0.7930	0.7325	0.7930	0.0	0.7933	0.7925
0.3132	10.0	1140	3.5572	1.0	129.2212	0.8308	0.8116	0.8308	0.0	0.8308	0.8302
0.3243	11.0	1254	5.8488	1.0	120.6807	0.7995	0.7842	0.7995	0.0	0.7989	0.7995

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.1