train_svamp_1754507512

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: 0.0719
• 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
0.2129	0.5	79	0.1318	35776
0.0783	1.0	158	0.0855	70672
0.021	1.5	237	0.0906	105904
0.067	2.0	316	0.0719	141328
0.0552	2.5	395	0.0803	176752
0.0169	3.0	474	0.0922	211808
0.0035	3.5	553	0.0882	247104
0.0329	4.0	632	0.0805	282048
0.0009	4.5	711	0.1044	317248
0.0186	5.0	790	0.0958	352592
0.0012	5.5	869	0.1174	388176
0.0132	6.0	948	0.1097	423184
0.0001	6.5	1027	0.1172	458640
0.0	7.0	1106	0.1209	493440
0.0019	7.5	1185	0.1226	528768
0.0001	8.0	1264	0.1217	563872
0.0	8.5	1343	0.1231	599232
0.0003	9.0	1422	0.1228	634544
0.0005	9.5	1501	0.1250	670064
0.0	10.0	1580	0.1213	705184

Framework versions

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