7c51113e523aa9ced607b87ca7db905a

This model is a fine-tuned version of openai-community/gpt2-large on the nyu-mll/glue [mnli] dataset. It achieves the following results on the evaluation set:

• Loss: 0.6595
• Data Size: 1.0
• Epoch Runtime: 2482.4669
• Accuracy: 0.8080
• F1 Macro: 0.8073
• Rouge1: 0.8079
• Rouge2: 0.0
• Rougel: 0.8081
• Rougelsum: 0.8081

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	1.3682	0	21.5456	0.3235	0.2054	0.3237	0.0	0.3232	0.3236
0.9336	1	12271	0.6733	0.0078	40.0026	0.7249	0.7265	0.7248	0.0	0.7249	0.7250
0.6613	2	24542	0.5809	0.0156	59.5731	0.7617	0.7596	0.7616	0.0	0.7619	0.7616
0.6297	3	36813	0.5667	0.0312	97.7997	0.7666	0.7639	0.7666	0.0	0.7666	0.7669
0.5523	4	49084	0.5254	0.0625	173.8515	0.7885	0.7882	0.7886	0.0	0.7885	0.7882
0.5121	5	61355	0.5170	0.125	327.9709	0.7898	0.7895	0.7897	0.0	0.7898	0.7897
0.4901	6	73626	0.4884	0.25	636.7369	0.8047	0.8034	0.8046	0.0	0.8047	0.8048
0.4208	7	85897	0.4804	0.5	1241.3797	0.8158	0.8145	0.8157	0.0	0.8158	0.8159
0.3833	8.0	98168	0.4867	1.0	2470.0903	0.8158	0.8157	0.8158	0.0	0.8158	0.8157
0.2843	9.0	110439	0.5317	1.0	2492.2505	0.8173	0.8163	0.8170	0.0	0.8173	0.8174
0.2198	10.0	122710	0.5789	1.0	2481.6612	0.8135	0.8126	0.8134	0.0	0.8135	0.8135
0.1957	11.0	134981	0.6595	1.0	2482.4669	0.8080	0.8073	0.8079	0.0	0.8081	0.8081

Framework versions

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