train_conala_101112_1760638005

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

• Loss: 1.2132
• Num Input Tokens Seen: 2719664

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: 101112
• 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
0.7178	2.0	952	0.8225	271336
0.5327	4.0	1904	0.7195	544016
0.3083	6.0	2856	0.7400	816320
0.3385	8.0	3808	0.7788	1087144
0.3304	10.0	4760	0.8422	1358736
0.1908	12.0	5712	0.9783	1631240
0.2107	14.0	6664	1.0549	1902720
0.1336	16.0	7616	1.1561	2175464
0.1052	18.0	8568	1.1978	2447544
0.1348	20.0	9520	1.2132	2719664

Framework versions

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