train_svamp_1757340201

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: 0.1597
• 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
2.3995	0.5	79	2.2946	35776
1.9914	1.0	158	1.7998	70672
1.4951	1.5	237	1.4120	105904
1.005	2.0	316	1.0429	141328
0.7404	2.5	395	0.7376	176752
0.557	3.0	474	0.5306	211808
0.4419	3.5	553	0.3952	247104
0.3074	4.0	632	0.3051	282048
0.2277	4.5	711	0.2493	317248
0.2723	5.0	790	0.2192	352592
0.2465	5.5	869	0.1984	388176
0.2696	6.0	948	0.1844	423184
0.2411	6.5	1027	0.1750	458640
0.2105	7.0	1106	0.1691	493440
0.3454	7.5	1185	0.1649	528768
0.188	8.0	1264	0.1626	563872
0.114	8.5	1343	0.1605	599232
0.2467	9.0	1422	0.1597	634544
0.225	9.5	1501	0.1608	670064
0.0899	10.0	1580	0.1605	705184

Framework versions

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