train_cb_1757340195

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: 0.1492
• Num Input Tokens Seen: 367864

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
1.2233	0.5088	29	1.1734	20064
1.1998	1.0175	58	1.1734	37832
0.6419	1.5263	87	0.5481	57288
0.2289	2.0351	116	0.1922	74520
0.571	2.5439	145	0.1745	93080
0.1435	3.0526	174	0.1630	111928
0.3009	3.5614	203	0.1576	131160
0.169	4.0702	232	0.1597	150056
0.0563	4.5789	261	0.1598	167208
0.0956	5.0877	290	0.1542	186160
0.0739	5.5965	319	0.1533	206000
0.1738	6.1053	348	0.1532	224064
0.1546	6.6140	377	0.1549	243840
0.0495	7.1228	406	0.1545	261504
0.1625	7.6316	435	0.1503	280352
0.0503	8.1404	464	0.1513	299344
0.3406	8.6491	493	0.1492	318672
0.0851	9.1579	522	0.1530	337480
0.1179	9.6667	551	0.1566	356456

Framework versions

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