92968cf61803cd02d245adf8e3723b73

This model is a fine-tuned version of facebook/opt-2.7b on the nyu-mll/glue [cola] dataset. It achieves the following results on the evaluation set:

• Loss: 0.9233
• Data Size: 1.0
• Epoch Runtime: 72.4859
• Accuracy: 0.6309
• F1 Macro: 0.5545
• Rouge1: 0.6299
• Rouge2: 0.0
• Rougel: 0.6304
• Rougelsum: 0.6309

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	0.7377	0	3.1043	0.6680	0.4573	0.6689	0.0	0.6680	0.6670
No log	1	267	0.7526	0.0078	4.5345	0.5771	0.5342	0.5762	0.0	0.5771	0.5771
No log	2	534	0.7467	0.0156	8.4896	0.4424	0.4417	0.4404	0.0	0.4434	0.4434
No log	3	801	0.8867	0.0312	13.9387	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
No log	4	1068	0.6943	0.0625	18.1450	0.5400	0.5059	0.5405	0.0	0.5400	0.5400
0.0443	5	1335	0.7981	0.125	26.5780	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.6449	6	1602	0.6380	0.25	34.0990	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.6254	7	1869	0.6229	0.5	47.4375	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.6234	8.0	2136	0.6131	1.0	74.3336	0.6875	0.4304	0.6875	0.0	0.6875	0.6875
0.5826	9.0	2403	0.6235	1.0	69.3182	0.6494	0.5696	0.6484	0.0	0.6494	0.6484
0.5103	10.0	2670	0.6738	1.0	74.7504	0.6729	0.5538	0.6729	0.0	0.6738	0.6729
0.3934	11.0	2937	0.7409	1.0	69.1825	0.6611	0.5464	0.6606	0.0	0.6611	0.6611
0.3562	12.0	3204	0.9233	1.0	72.4859	0.6309	0.5545	0.6299	0.0	0.6304	0.6309

Framework versions

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