0d045a5d6d34f93e10e6dcee53b5587f

This model is a fine-tuned version of deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B on the nyu-mll/glue dataset. It achieves the following results on the evaluation set:

• Loss: 1.7476
• Data Size: 1.0
• Epoch Runtime: 2443.0497
• Accuracy: 0.8976
• F1 Macro: 0.8904
• Rouge1: 0.8976
• Rouge2: 0.0
• Rougel: 0.8977
• Rougelsum: 0.8976

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	5.5591	0	76.6514	0.6224	0.3926	0.6223	0.0	0.6225	0.6222
2.5896	1	11370	1.8115	0.0078	96.3145	0.8022	0.7686	0.8021	0.0	0.8023	0.8021
1.4939	2	22740	1.3846	0.0156	115.4389	0.8494	0.8417	0.8494	0.0	0.8494	0.8494
1.4501	3	34110	1.3332	0.0312	155.1173	0.8548	0.8440	0.8548	0.0	0.8548	0.8548
1.253	4	45480	1.3384	0.0625	227.7592	0.8633	0.8497	0.8633	0.0	0.8633	0.8633
1.1751	5	56850	1.1482	0.125	371.8913	0.8742	0.8660	0.8743	0.0	0.8743	0.8741
1.026	6	68220	1.0992	0.25	667.6549	0.8786	0.8720	0.8787	0.0	0.8786	0.8786
0.9374	7	79590	1.0103	0.5	1252.1296	0.8927	0.8869	0.8928	0.0	0.8927	0.8927
0.8321	8.0	90960	1.0477	1.0	2447.5549	0.8984	0.8917	0.8984	0.0	0.8984	0.8985
0.4763	9.0	102330	1.4150	1.0	2455.3630	0.8963	0.8902	0.8962	0.0	0.8964	0.8962
0.4159	10.0	113700	1.5539	1.0	2474.4821	0.8957	0.8875	0.8957	0.0	0.8957	0.8956
0.3559	11.0	125070	1.7476	1.0	2443.0497	0.8976	0.8904	0.8976	0.0	0.8977	0.8976

Framework versions

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