train_cb_1757340168

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: 1.5804
• 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
0.2601	0.5088	29	0.2370	18048
0.3128	1.0175	58	0.4424	36928
1.7629	1.5263	87	0.5561	54176
0.3378	2.0351	116	0.2020	73136
0.3303	2.5439	145	0.1614	91216
0.0453	3.0526	174	0.2379	110696
0.116	3.5614	203	0.1001	129448
0.1798	4.0702	232	0.0368	147176
0.2801	4.5789	261	0.1194	164424
0.0358	5.0877	290	0.0774	183416
0.2572	5.5965	319	0.0889	203256
0.0011	6.1053	348	0.1109	220912
0.0696	6.6140	377	0.0931	240336
0.0251	7.1228	406	0.0881	257848
0.006	7.6316	435	0.0805	276824
0.0004	8.1404	464	0.1005	294504
0.0008	8.6491	493	0.0857	313576
0.0015	9.1579	522	0.0954	332256
0.0009	9.6667	551	0.0996	350336

Framework versions

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