train_cb_789_1760637868

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: 0.0390
• Num Input Tokens Seen: 711112

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: 789
• 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
0.4751	1.0	57	1.2149	35448
0.3845	2.0	114	1.1475	70496
0.2278	3.0	171	0.1985	106416
0.2558	4.0	228	0.0931	142480
0.1601	5.0	285	0.1481	177224
0.0544	6.0	342	0.0972	212000
0.0212	7.0	399	0.0559	248272
0.0503	8.0	456	0.0554	284248
0.0561	9.0	513	0.0994	319488
0.0031	10.0	570	0.0718	354472
0.0013	11.0	627	0.1540	389408
0.0013	12.0	684	0.1397	425328
0.0005	13.0	741	0.1118	461216
0.0003	14.0	798	0.1103	496704
0.0002	15.0	855	0.1187	532504
0.0003	16.0	912	0.1145	567952
0.0002	17.0	969	0.1075	603760
0.0004	18.0	1026	0.1133	639784
0.0002	19.0	1083	0.1141	675800
0.0002	20.0	1140	0.1113	711112

Framework versions

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