train_boolq_456_1765357262

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

• Loss: 0.3382
• Num Input Tokens Seen: 42758400

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: 0.001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 456
• 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
0.3414	1.0	2121	0.3479	2134176
0.5563	2.0	4242	0.3527	4262848
0.3714	3.0	6363	0.3680	6393376
0.3401	4.0	8484	0.3407	8527168
0.3356	5.0	10605	0.3411	10663616
0.3954	6.0	12726	0.3401	12804064
0.3437	7.0	14847	0.3379	14945088
0.3653	8.0	16968	0.3673	17092000
0.2817	9.0	19089	0.3553	19221568
0.3382	10.0	21210	0.3547	21363392
0.3377	11.0	23331	0.3380	23510432
0.3508	12.0	25452	0.3559	25647872
0.3379	13.0	27573	0.3386	27792544
0.3317	14.0	29694	0.3395	29926112
0.3598	15.0	31815	0.3384	32059584
0.3823	16.0	33936	0.3384	34199392
0.3576	17.0	36057	0.3374	36344896
0.3536	18.0	38178	0.3390	38482112
0.348	19.0	40299	0.3383	40624896
0.2996	20.0	42420	0.3378	42758400

Framework versions

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