train_piqa_42_1760795575

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

• Loss: 0.7820
• Num Input Tokens Seen: 39307696

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: 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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.2246	2.0	6446	0.2323	3927648
0.23	4.0	12892	0.2315	7849792
0.2273	6.0	19338	0.2314	11772672
0.2238	8.0	25784	0.2318	15703520
0.2205	10.0	32230	0.2341	19637408
0.1803	12.0	38676	0.2616	23557280
0.2274	14.0	45122	0.3020	27497200
0.1963	16.0	51568	0.4939	31434736
0.1651	18.0	58014	0.7020	35376544
0.1214	20.0	64460	0.7820	39307696

Framework versions

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