jjee2/jeevanions__llama381binstruct_summarize_short

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: 2.8862

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
1.7758	1.3158	25	1.6937
0.7261	2.6316	50	1.7634
0.3962	3.9474	75	1.8509
0.1317	5.2632	100	2.1687
0.0754	6.5789	125	2.3035
0.042	7.8947	150	2.4272
0.031	9.2105	175	2.4986
0.0143	10.5263	200	2.6026
0.0126	11.8421	225	2.5424
0.0081	13.1579	250	2.6719
0.0037	14.4737	275	2.7089
0.0034	15.7895	300	2.7364
0.0035	17.1053	325	2.7922
0.0027	18.4211	350	2.8199
0.0031	19.7368	375	2.8449
0.0022	21.0526	400	2.8645
0.0025	22.3684	425	2.8749
0.0018	23.6842	450	2.8810
0.0019	25.0	475	2.8851
0.0028	26.3158	500	2.8862

Framework versions

• PEFT 0.12.0
• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1