262f26c2dc910c584afa3872e7b1e598

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

• Loss: 2.6860
• Data Size: 1.0
• Epoch Runtime: 8855.0263
• Accuracy: 0.8574
• F1 Macro: 0.8494
• Rouge1: 0.8573
• Rouge2: 0.0
• Rougel: 0.8575
• Rougelsum: 0.8573

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	20.2133	0	106.7802	0.3700	0.2732	0.3701	0.0	0.3700	0.3701
6.4484	1	11370	2.0827	0.0078	174.9975	0.7434	0.7048	0.7434	0.0	0.7432	0.7432
1.7827	2	22740	1.5796	0.0156	252.1326	0.8161	0.7950	0.8161	0.0	0.8162	0.8162
1.702	3	34110	1.5190	0.0312	395.4948	0.8247	0.8137	0.8248	0.0	0.8248	0.8248
1.4131	4	45480	1.4583	0.0625	660.1275	0.8414	0.8265	0.8414	0.0	0.8413	0.8415
1.3822	5	56850	1.2959	0.125	1209.8396	0.8571	0.8478	0.8571	0.0	0.8572	0.8571
1.2604	6	68220	1.2785	0.25	2284.3232	0.8574	0.8504	0.8574	0.0	0.8575	0.8573
1.0892	7	79590	1.2494	0.5	4507.5064	0.8602	0.8542	0.8602	0.0	0.8602	0.8602
1.023	8.0	90960	1.1721	1.0	8839.8082	0.8826	0.8742	0.8827	0.0	0.8826	0.8825
0.5925	9.0	102330	1.3181	1.0	8835.5966	0.8781	0.8669	0.8781	0.0	0.8782	0.8781
0.3926	10.0	113700	2.2544	1.0	8805.7276	0.8721	0.8595	0.8720	0.0	0.8722	0.8721
0.4668	11.0	125070	2.0069	1.0	8808.7701	0.8692	0.8569	0.8692	0.0	0.8693	0.8692
0.3441	12.0	136440	2.6860	1.0	8855.0263	0.8574	0.8494	0.8573	0.0	0.8575	0.8573

Framework versions

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