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: 2.7944

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
1.782	1.25	25	1.8304
0.7787	2.5	50	1.7098
0.3977	3.75	75	1.8415
0.1875	5.0	100	1.9322
0.0907	6.25	125	2.1723
0.0304	7.5	150	2.4278
0.0195	8.75	175	2.4924
0.02	10.0	200	2.4250
0.0194	11.25	225	2.5124
0.0069	12.5	250	2.5796
0.0048	13.75	275	2.5765
0.0055	15.0	300	2.6462
0.0038	16.25	325	2.6535
0.0023	17.5	350	2.6849
0.0026	18.75	375	2.7281
0.0021	20.0	400	2.7575
0.0018	21.25	425	2.7753
0.0016	22.5	450	2.7865
0.0016	23.75	475	2.7929
0.0017	25.0	500	2.7944

Framework versions

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