train_qqp_1754652135

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: 0.1866
• 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
0.2031	0.5	40933	0.2528	12552544
0.2786	1.0	81866	0.2403	25087944
0.2791	1.5	122799	0.2292	37621672
0.2403	2.0	163732	0.2193	50164864
0.2049	2.5	204665	0.2120	62700096
0.1847	3.0	245598	0.2067	75242048
0.181	3.5	286531	0.2104	87769248
0.2662	4.0	327464	0.1999	100320328
0.2374	4.5	368397	0.2003	112855464
0.2511	5.0	409330	0.1927	125387608
0.1266	5.5	450263	0.1914	137931704
0.1893	6.0	491196	0.1936	150463800
0.2127	6.5	532129	0.1902	163003576
0.2026	7.0	573062	0.1895	175543400
0.1628	7.5	613995	0.1886	188096200
0.2036	8.0	654928	0.1866	200622552
0.1424	8.5	695861	0.1876	213151000
0.1818	9.0	736794	0.1872	225701792
0.1845	9.5	777727	0.1875	238243968
0.2161	10.0	818660	0.1873	250787112

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.1