llama_3_gsm8k_midset_helpful – AI Model by CharlesLi | AlphaNeural AI | AlphaNeural AI

You can deploy this model and start earning money today!

CharlesLi/
llama_3_gsm8k_midset_helpful

like

0

alignment-handbook

generated_from_trainer

meta-llama/Llama-3.1-8B-Instruct

Views

No views yet

llama_3_gsm8k_midset_helpful

This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.5614

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: 4
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
gradient_accumulation_steps: 2
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
training_steps: 200

Training results

Training Loss	Epoch	Step	Validation Loss
1.1152	0.04	5	1.0754
0.8157	0.08	10	0.7806
0.7435	0.12	15	0.6954
0.6353	0.16	20	0.6491
0.6048	0.2	25	0.6166
0.5449	0.24	30	0.6001
0.6039	0.28	35	0.5862
0.6182	0.32	40	0.5862
0.5985	0.36	45	0.5789
0.5475	0.4	50	0.5770
0.5766	0.44	55	0.5765
0.5648	0.48	60	0.5759
0.5472	0.52	65	0.5727
0.5047	0.56	70	0.5739
0.5853	0.6	75	0.5697
0.5506	0.64	80	0.5700
0.5399	0.68	85	0.5689
0.5665	0.72	90	0.5676
0.5807	0.76	95	0.5659
0.563	0.8	100	0.5647
0.5776	0.84	105	0.5625
0.5238	0.88	110	0.5622
0.5207	0.92	115	0.5607
0.6012	0.96	120	0.5605
0.5395	1.0	125	0.5603
0.4443	1.04	130	0.5597
0.4903	1.08	135	0.5613
0.4373	1.12	140	0.5649
0.4675	1.16	145	0.5657
0.5231	1.2	150	0.5642
0.4862	1.24	155	0.5634
0.4808	1.28	160	0.5613
0.4593	1.32	165	0.5618
0.5099	1.3600	170	0.5618
0.4777	1.4	175	0.5618
0.4927	1.44	180	0.5615
0.4805	1.48	185	0.5615
0.4681	1.52	190	0.5619
0.4562	1.56	195	0.5610
0.471	1.6	200	0.5614

Framework versions

PEFT 0.12.0
Transformers 4.44.2
Pytorch 2.4.1+cu121
Datasets 3.0.0
Tokenizers 0.19.1