train_svamp_1754652179

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: 1.5635
• Num Input Tokens Seen: 705184

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: 123
• 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
0.2917	0.5	79	0.2648	35776
0.1311	1.0	158	0.6598	70672
0.0572	1.5	237	0.1467	105904
0.1331	2.0	316	0.1527	141328
0.1217	2.5	395	0.0990	176752
0.0748	3.0	474	0.1005	211808
0.0843	3.5	553	0.0869	247104
0.1127	4.0	632	0.0806	282048
0.0631	4.5	711	0.0864	317248
0.0826	5.0	790	0.0809	352592
0.0446	5.5	869	0.0923	388176
0.0774	6.0	948	0.0772	423184
0.0554	6.5	1027	0.1098	458640
0.046	7.0	1106	0.0971	493440
0.1462	7.5	1185	0.1082	528768
0.0111	8.0	1264	0.1204	563872
0.0084	8.5	1343	0.1212	599232
0.0727	9.0	1422	0.1270	634544
0.0517	9.5	1501	0.1277	670064
0.0049	10.0	1580	0.1257	705184

Framework versions

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