train_record_456_1765497008

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

• Loss: 0.7662
• Num Input Tokens Seen: 825842872

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: 1e-05
• 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
0.3162	2.0	55542	0.3295	82575768
0.2316	4.0	111084	0.3089	165157440
0.1282	6.0	166626	0.3197	247729272
0.3115	8.0	222168	0.3684	330335744
0.1834	10.0	277710	0.3845	412926800
0.1034	12.0	333252	0.4592	495527344
0.1371	14.0	388794	0.5166	578116744
0.1282	16.0	444336	0.5714	660702320
0.095	18.0	499878	0.7066	743276176
0.1437	20.0	555420	0.7662	825842872

Framework versions

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