f4712aefdb8a50d3880444a6d4a145b7

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

• Loss: 2.8378
• Data Size: 0.125
• Epoch Runtime: 461.5158
• Accuracy: 0.7020
• F1 Macro: 0.6939
• Rouge1: 0.7021
• Rouge2: 0.0
• Rougel: 0.7023
• Rougelsum: 0.7023

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	11.5448	0	24.9959	0.3241	0.2119	0.3242	0.0	0.3238	0.3243
4.3487	1	12271	2.4961	0.0078	51.0444	0.7479	0.7412	0.7475	0.0	0.7482	0.7477
2.68	2	24542	2.5649	0.0156	82.9646	0.7548	0.7555	0.7549	0.0	0.7551	0.7548
2.9192	3	36813	3.9912	0.0312	137.1969	0.4733	0.4487	0.4732	0.0	0.4739	0.4734
4.0407	4	49084	2.9462	0.0625	242.9109	0.6916	0.6914	0.6916	0.0	0.6915	0.6916
3.0393	5	61355	2.8378	0.125	461.5158	0.7020	0.6939	0.7021	0.0	0.7023	0.7023

Framework versions

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