1e32a848c8a2030638ce0347bcd356b6

This model is a fine-tuned version of google-bert/bert-large-uncased-whole-word-masking on the nyu-mll/glue [qnli] dataset. It achieves the following results on the evaluation set:

• Loss: 0.6982
• Data Size: 0.5
• Epoch Runtime: 159.0188
• Accuracy: 0.4943
• F1 Macro: 0.3308
• Rouge1: 0.4947
• Rouge2: 0.0
• Rougel: 0.4943
• Rougelsum: 0.4944

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	0.8001	0	4.6541	0.4945	0.3318	0.4950	0.0	0.4945	0.4945
No log	1	3273	0.5296	0.0078	7.3577	0.7423	0.7359	0.7426	0.0	0.7425	0.7419
0.0101	2	6546	0.4348	0.0156	10.8851	0.8110	0.8093	0.8108	0.0	0.8108	0.8107
0.5002	3	9819	0.4245	0.0312	14.8009	0.8300	0.8290	0.8301	0.0	0.8296	0.8298
0.4668	4	13092	0.4600	0.0625	24.7220	0.8210	0.8196	0.8211	0.0	0.8213	0.8208
0.7101	5	16365	0.6945	0.125	43.7211	0.5057	0.3359	0.5053	0.0	0.5057	0.5056
0.7041	6	19638	0.6962	0.25	81.4890	0.4943	0.3308	0.4947	0.0	0.4943	0.4944
0.7076	7	22911	0.6982	0.5	159.0188	0.4943	0.3308	0.4947	0.0	0.4943	0.4944

Framework versions

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