train_conala_101112_1760638010

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: 0.6943
• Num Input Tokens Seen: 3060208

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: 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
2.3421	1.0	536	2.2859	153344
1.758	2.0	1072	1.1338	306640
1.1588	3.0	1608	0.9419	459376
0.6676	4.0	2144	0.8533	612008
0.7588	5.0	2680	0.8118	764936
0.8085	6.0	3216	0.7846	917624
0.7514	7.0	3752	0.7636	1070488
0.5654	8.0	4288	0.7469	1223384
0.6064	9.0	4824	0.7342	1376240
0.4365	10.0	5360	0.7238	1529640
0.8136	11.0	5896	0.7153	1682336
0.5549	12.0	6432	0.7098	1835928
0.7974	13.0	6968	0.7043	1989136
0.5278	14.0	7504	0.7004	2142632
0.7199	15.0	8040	0.6974	2295280
0.5653	16.0	8576	0.6962	2447904
0.6505	17.0	9112	0.6950	2600776
0.6361	18.0	9648	0.6947	2753536
0.8247	19.0	10184	0.6947	2906984
0.5962	20.0	10720	0.6943	3060208

Framework versions

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