c103c51f28b49fef7c43ecc9e6982489

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

• Loss: 1.5111
• Data Size: 1.0
• Epoch Runtime: 2799.0246
• Accuracy: 0.875
• F1 Macro: 0.8749
• Rouge1: 0.8752
• Rouge2: 0.0
• Rougel: 0.875
• Rougelsum: 0.875

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	12.2989	0	17.3046	0.5108	0.3609	0.5103	0.0	0.5105	0.5107
No log	1	3273	1.9887	0.0078	37.5470	0.7798	0.7763	0.7800	0.0	0.7798	0.7798
0.0584	2	6546	1.4484	0.0156	67.6180	0.8526	0.8526	0.8528	0.0	0.8529	0.8526
1.8512	3	9819	1.3563	0.0312	112.2425	0.8638	0.8634	0.8638	0.0	0.8640	0.8638
1.6293	4	13092	1.2922	0.0625	205.0164	0.8684	0.8683	0.8686	0.0	0.8686	0.8682
1.3397	5	16365	1.2003	0.125	383.4920	0.8812	0.8812	0.8811	0.0	0.8814	0.8814
1.3307	6	19638	1.3905	0.25	730.1136	0.8562	0.8553	0.8561	0.0	0.8563	0.8562
1.2215	7	22911	1.4402	0.5	1404.8970	0.8669	0.8667	0.8669	0.0	0.8671	0.8671
1.1027	8.0	26184	1.5401	1.0	2810.6302	0.8458	0.8450	0.8458	0.0	0.8461	0.8458
0.6854	9.0	29457	1.5111	1.0	2799.0246	0.875	0.8749	0.8752	0.0	0.875	0.875

Framework versions

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