train_cb_456_1760637755

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

• Loss: 0.2095
• Num Input Tokens Seen: 721856

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: 0.001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 456
• 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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.3273	1.0	57	0.3546	36072
0.2521	2.0	114	0.1955	72896
0.0475	3.0	171	0.1200	109080
0.0829	4.0	228	0.1297	145936
0.1064	5.0	285	0.0915	182024
0.0661	6.0	342	0.1748	218672
0.0432	7.0	399	0.2276	254232
0.0668	8.0	456	0.0810	290912
0.0025	9.0	513	0.1416	326432
0.0031	10.0	570	0.1528	362240
0.0008	11.0	627	0.1870	397880
0.0003	12.0	684	0.1899	433352
0.0003	13.0	741	0.1922	469568
0.0003	14.0	798	0.1941	505048
0.0003	15.0	855	0.1952	541088
0.0003	16.0	912	0.1958	577512
0.0002	17.0	969	0.1953	614128
0.0002	18.0	1026	0.1983	649608
0.0002	19.0	1083	0.1933	685200
0.0002	20.0	1140	0.1910	721856

Framework versions

• PEFT 0.17.1
• Transformers 4.51.3
• Pytorch 2.9.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4