cf13a4ccd9ae9ddc5526005e1a2cab42

This model is a fine-tuned version of facebook/opt-350m on the nyu-mll/glue [mnli] dataset. It achieves the following results on the evaluation set:

• Loss: 0.7295
• Data Size: 1.0
• Epoch Runtime: 1649.8046
• Accuracy: 0.7302
• F1 Macro: 0.7308
• Rouge1: 0.7301
• Rouge2: 0.0
• Rougel: 0.7301
• Rougelsum: 0.7302

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	1.8038	0	14.1334	0.3374	0.3131	0.3377	0.0	0.3372	0.3375
1.191	1	12271	0.8747	0.0078	26.4842	0.6401	0.6392	0.6400	0.0	0.6403	0.6405
0.792	2	24542	0.7253	0.0156	38.6624	0.6966	0.6964	0.6968	0.0	0.6967	0.6967
0.7442	3	36813	0.6713	0.0312	65.2249	0.7199	0.7184	0.7197	0.0	0.7201	0.7200
0.705	4	49084	0.6726	0.0625	111.5077	0.7264	0.7230	0.7261	0.0	0.7263	0.7265
0.6424	5	61355	0.6747	0.125	209.9674	0.7237	0.7216	0.7234	0.0	0.7239	0.7237
0.6683	6	73626	0.6516	0.25	422.0297	0.7427	0.7428	0.7426	0.0	0.7430	0.7427
0.634	7	85897	0.6696	0.5	832.5325	0.7258	0.7251	0.7256	0.0	0.7260	0.7258
0.5935	8.0	98168	0.6280	1.0	1650.7072	0.7528	0.7518	0.7529	0.0	0.7527	0.7529
0.5173	9.0	110439	0.6343	1.0	1651.3985	0.7395	0.7365	0.7396	0.0	0.7394	0.7393
0.532	10.0	122710	0.6600	1.0	1646.3381	0.7400	0.7402	0.7397	0.0	0.7401	0.7401
0.4848	11.0	134981	0.6879	1.0	1654.6315	0.7400	0.7384	0.7399	0.0	0.7403	0.7401
0.443	12.0	147252	0.7295	1.0	1649.8046	0.7302	0.7308	0.7301	0.0	0.7301	0.7302

Framework versions

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