fcea33b5986adbc1724bc58cbd7e4356

This model is a fine-tuned version of facebook/opt-6.7b on the nyu-mll/glue [mrpc] dataset. It achieves the following results on the evaluation set:

• Loss: 2.2901
• Data Size: 1.0
• Epoch Runtime: 61.7390
• Accuracy: 0.6144
• F1 Macro: 0.5770
• Rouge1: 0.6138
• Rouge2: 0.0
• Rougel: 0.6138
• Rougelsum: 0.6138

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	1.5125	0	5.2067	0.3585	0.3085	0.3585	0.0	0.3591	0.3579
No log	1	114	8.1090	0.0078	6.3588	0.6651	0.3994	0.6657	0.0	0.6645	0.6651
No log	2	228	1.0724	0.0156	13.3963	0.6828	0.4794	0.6828	0.0	0.6828	0.6822
No log	3	342	2.2983	0.0312	26.7997	0.6651	0.3994	0.6657	0.0	0.6645	0.6651
0.061	4	456	0.7924	0.0625	37.5244	0.6651	0.3994	0.6657	0.0	0.6645	0.6651
0.061	5	570	0.8707	0.125	50.1045	0.3644	0.3075	0.3638	0.0	0.3644	0.3644
0.061	6	684	0.6178	0.25	60.9226	0.6840	0.5294	0.6834	0.0	0.6840	0.6840
0.1625	7	798	0.6085	0.5	47.4606	0.6840	0.5015	0.6846	0.0	0.6840	0.6840
0.5515	8.0	912	0.6200	1.0	74.5693	0.6857	0.5926	0.6857	0.0	0.6851	0.6857
0.327	9.0	1026	0.6732	1.0	82.4028	0.6515	0.5935	0.6509	0.0	0.6509	0.6515
0.159	10.0	1140	1.2080	1.0	83.5275	0.6173	0.5750	0.6179	0.0	0.6173	0.6173
0.0824	11.0	1254	2.2901	1.0	61.7390	0.6144	0.5770	0.6138	0.0	0.6138	0.6138

Framework versions

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