llama381binstruct_summarize_short

This model is a fine-tuned version of NousResearch/Meta-Llama-3.1-8B-Instruct on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 2.7209

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: 0.0002
• train_batch_size: 1
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 30
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss
1.632	1.1905	25	1.5798
0.8162	2.3810	50	1.6864
0.4602	3.5714	75	1.8529
0.1629	4.7619	100	2.1229
0.0928	5.9524	125	2.2184
0.0498	7.1429	150	2.2687
0.042	8.3333	175	2.3336
0.0151	9.5238	200	2.4483
0.0179	10.7143	225	2.3340
0.0106	11.9048	250	2.4193
0.0059	13.0952	275	2.4428
0.0033	14.2857	300	2.5659
0.0024	15.4762	325	2.6152
0.002	16.6667	350	2.6552
0.0017	17.8571	375	2.6778
0.0018	19.0476	400	2.6956
0.0019	20.2381	425	2.7082
0.0015	21.4286	450	2.7155
0.0015	22.6190	475	2.7195
0.0015	23.8095	500	2.7209

Framework versions

• PEFT 0.12.0
• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1