train_cb_1757340167

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: 0.1335
• Num Input Tokens Seen: 361992

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: 42
• 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
0.7951	0.5088	29	0.6471	18048
0.2042	1.0175	58	0.1866	36928
0.1537	1.5263	87	0.1335	54176
0.331	2.0351	116	0.1582	73136
0.1844	2.5439	145	0.2117	91216
0.0761	3.0526	174	0.1914	110696
0.1566	3.5614	203	0.1673	129448
0.197	4.0702	232	0.1535	147176
0.2652	4.5789	261	0.1619	164424
0.0332	5.0877	290	0.1772	183416
0.465	5.5965	319	0.1760	203256
0.0383	6.1053	348	0.1951	220912
0.0998	6.6140	377	0.1752	240336
0.1873	7.1228	406	0.1756	257848
0.0519	7.6316	435	0.1726	276824
0.1781	8.1404	464	0.1769	294504
0.053	8.6491	493	0.1806	313576
0.05	9.1579	522	0.1811	332256
0.3965	9.6667	551	0.1742	350336

Framework versions

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