c7d295d3a4978b56db945fca0341dc19

This model is a fine-tuned version of Qwen/Qwen2.5-3B on the nyu-mll/glue [qnli] dataset. It achieves the following results on the evaluation set:

• Loss: 2.9725
• Data Size: 1.0
• Epoch Runtime: 933.5313
• Accuracy: 0.6138
• F1 Macro: 0.6123
• Rouge1: 0.6140
• Rouge2: 0.0
• Rougel: 0.6136
• Rougelsum: 0.6140

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	10.9829	0	14.5472	0.5138	0.3688	0.5136	0.0	0.5136	0.5136
No log	1	3273	3.1694	0.0078	21.4254	0.5066	0.3402	0.5062	0.0	0.5062	0.5066
0.1122	2	6546	2.8628	0.0156	31.8997	0.5121	0.3677	0.5119	0.0	0.5118	0.5121
2.9617	3	9819	2.7658	0.0312	48.9986	0.5233	0.3994	0.5231	0.0	0.5232	0.5230
2.7901	4	13092	2.8225	0.0625	80.3174	0.5193	0.4430	0.5195	0.0	0.5197	0.5195
2.8098	5	16365	2.7665	0.125	135.9528	0.5357	0.4651	0.5360	0.0	0.5358	0.5355
2.993	6	19638	2.7590	0.25	245.5732	0.5472	0.5031	0.5476	0.0	0.5476	0.5475
2.6067	7	22911	2.6310	0.5	482.6382	0.6103	0.6103	0.6105	0.0	0.6103	0.6101
2.6151	8.0	26184	2.7310	1.0	929.2936	0.6079	0.5906	0.6079	0.0	0.6079	0.6079
2.384	9.0	29457	2.6490	1.0	929.0926	0.6075	0.6059	0.6077	0.0	0.6077	0.6075
2.1036	10.0	32730	2.6881	1.0	954.7908	0.6171	0.6127	0.6172	0.0	0.6171	0.6171
1.806	11.0	36003	2.9725	1.0	933.5313	0.6138	0.6123	0.6140	0.0	0.6136	0.6140

Framework versions

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