train_svamp_42_1757596060

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: 0.0600
• Num Input Tokens Seen: 704336

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: 42
• 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: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
1.8622	0.5	79	1.7169	35680
0.1542	1.0	158	0.1326	70512
0.0517	1.5	237	0.1069	105904
0.052	2.0	316	0.0929	140960
0.052	2.5	395	0.0873	176096
0.0962	3.0	474	0.0847	211424
0.0284	3.5	553	0.0809	246784
0.1498	4.0	632	0.0747	281968
0.0422	4.5	711	0.0786	317232
0.0423	5.0	790	0.0697	352368
0.0947	5.5	869	0.0642	387824
0.0595	6.0	948	0.0630	422704
0.0149	6.5	1027	0.0656	457744
0.0533	7.0	1106	0.0607	493200
0.0465	7.5	1185	0.0603	528304
0.1566	8.0	1264	0.0603	563520
0.063	8.5	1343	0.0600	599072
0.0428	9.0	1422	0.0600	634176
0.0764	9.5	1501	0.0603	669440
0.0419	10.0	1580	0.0605	704336

Framework versions

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