96529d1973d145b0ce126c03233b50f6

This model is a fine-tuned version of distilbert/distilroberta-base on the nyu-mll/glue [mnli] dataset. It achieves the following results on the evaluation set:

• Loss: 0.6210
• Data Size: 1.0
• Epoch Runtime: 490.5435
• Accuracy: 0.8114
• F1 Macro: 0.8107
• Rouge1: 0.8115
• Rouge2: 0.0
• Rougel: 0.8113
• Rougelsum: 0.8115

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	1.1023	0	4.6905	0.3545	0.1747	0.3544	0.0	0.3545	0.3543
1.0933	1	12271	0.9008	0.0078	8.6237	0.6021	0.5969	0.6019	0.0	0.6023	0.6023
0.8248	2	24542	0.7477	0.0156	12.2017	0.6857	0.6797	0.6858	0.0	0.6860	0.6857
0.7031	3	36813	0.6315	0.0312	19.8785	0.7371	0.7345	0.7369	0.0	0.7373	0.7372
0.649	4	49084	0.5877	0.0625	34.7162	0.7629	0.7623	0.7625	0.0	0.7628	0.7627
0.5638	5	61355	0.5520	0.125	65.1321	0.7786	0.7767	0.7784	0.0	0.7789	0.7786
0.5506	6	73626	0.6088	0.25	125.5367	0.7786	0.7792	0.7786	0.0	0.7786	0.7786
0.4764	7	85897	0.5373	0.5	246.3156	0.7978	0.7971	0.7978	0.0	0.7981	0.7980
0.446	8.0	98168	0.5068	1.0	490.8175	0.8110	0.8100	0.8109	0.0	0.8110	0.8110
0.3883	9.0	110439	0.5302	1.0	486.9774	0.8090	0.8086	0.8091	0.0	0.8089	0.8089
0.3427	10.0	122710	0.5458	1.0	497.9757	0.8088	0.8077	0.8087	0.0	0.8088	0.8090
0.3202	11.0	134981	0.5834	1.0	487.6945	0.8106	0.8105	0.8104	0.0	0.8105	0.8108
0.254	12.0	147252	0.6210	1.0	490.5435	0.8114	0.8107	0.8115	0.0	0.8113	0.8115

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1