0136daabd370bd556a752e67a3426ad5

This model is a fine-tuned version of facebook/opt-125m on the nyu-mll/glue [qnli] dataset. It achieves the following results on the evaluation set:

• Loss: 0.5525
• Data Size: 1.0
• Epoch Runtime: 225.7927
• Accuracy: 0.8401
• F1 Macro: 0.8398
• Rouge1: 0.8401
• Rouge2: 0.0
• Rougel: 0.8401
• Rougelsum: 0.8399

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	0.9413	0	4.8448	0.5081	0.3642	0.5081	0.0	0.5080	0.5077
No log	1	3273	0.7308	0.0078	6.5876	0.6213	0.5775	0.6217	0.0	0.6217	0.6213
0.0103	2	6546	0.5336	0.0156	8.2621	0.7369	0.7295	0.7371	0.0	0.7369	0.7366
0.5292	3	9819	0.5882	0.0312	11.8899	0.6985	0.6779	0.6983	0.0	0.6987	0.6985
0.4739	4	13092	0.5634	0.0625	18.8684	0.7357	0.7239	0.7357	0.0	0.7358	0.7357
0.3996	5	16365	0.4222	0.125	32.8643	0.8187	0.8180	0.8187	0.0	0.8189	0.8184
0.4156	6	19638	0.3900	0.25	60.8931	0.8366	0.8362	0.8362	0.0	0.8369	0.8364
0.3428	7	22911	0.3820	0.5	114.4534	0.8357	0.8343	0.8357	0.0	0.8358	0.8357
0.3384	8.0	26184	0.3598	1.0	223.5399	0.8471	0.8464	0.8471	0.0	0.8472	0.8471
0.2261	9.0	29457	0.4172	1.0	226.8864	0.8434	0.8426	0.8432	0.0	0.8436	0.8436
0.166	10.0	32730	0.4662	1.0	227.6201	0.8555	0.8555	0.8557	0.0	0.8555	0.8556
0.1317	11.0	36003	0.5882	1.0	222.9718	0.8327	0.8320	0.8325	0.0	0.8329	0.8327
0.137	12.0	39276	0.5525	1.0	225.7927	0.8401	0.8398	0.8401	0.0	0.8401	0.8399

Framework versions

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