train_svamp_101112_1760637999

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

• Loss: 0.6033
• Num Input Tokens Seen: 1272000

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
0.6119	2.0	280	0.7101	127328
0.4197	4.0	560	0.6222	254624
0.5296	6.0	840	0.6052	381760
0.2225	8.0	1120	0.4053	509056
0.2433	10.0	1400	0.4110	636096
0.1727	12.0	1680	0.4295	763200
0.1309	14.0	1960	0.5017	890752
0.0895	16.0	2240	0.5760	1017696
0.1119	18.0	2520	0.6015	1145152
0.0136	20.0	2800	0.6033	1272000

Framework versions

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