train_qqp_1753094140

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the qqp dataset. It achieves the following results on the evaluation set:

• Loss: 0.0982
• Num Input Tokens Seen: 250787112

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: 4
• eval_batch_size: 4
• seed: 123
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.0894	0.5	40933	0.1297	12552544
0.2547	1.0	81866	0.1199	25087944
0.0632	1.5	122799	0.1031	37621672
0.0726	2.0	163732	0.1010	50164864
0.1418	2.5	204665	0.1057	62700096
0.0879	3.0	245598	0.1007	75242048
0.1207	3.5	286531	0.1086	87769248
0.2482	4.0	327464	0.0982	100320328
0.1084	4.5	368397	0.1174	112855464
0.1453	5.0	409330	0.1066	125387608
0.0404	5.5	450263	0.1204	137931704
0.063	6.0	491196	0.1151	150463800
0.0053	6.5	532129	0.1377	163003576
0.0088	7.0	573062	0.1222	175543400
0.0026	7.5	613995	0.1412	188096200
0.0033	8.0	654928	0.1379	200622552
0.0026	8.5	695861	0.1420	213151000
0.1073	9.0	736794	0.1450	225701792
0.0012	9.5	777727	0.1450	238243968
0.0753	10.0	818660	0.1456	250787112

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.7.1+cu126
• Datasets 3.6.0
• Tokenizers 0.21.1