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

Loss: 1.6162

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: 1e-06
train_batch_size: 2
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 100
training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss
1.8795	0.4	50	1.7359
1.5557	0.8	100	1.5149
1.5505	1.2	150	1.4878
1.4839	1.6	200	1.4811
1.4928	2.0	250	1.4778
1.3677	2.4	300	1.4931
1.3947	2.8	350	1.4940
1.1632	3.2	400	1.5277
1.2544	3.6	450	1.5207
1.147	4.0	500	1.5292
1.1403	4.4	550	1.5664
1.0704	4.8	600	1.5711
1.0585	5.2	650	1.6079
1.0515	5.6	700	1.6006
0.9566	6.0	750	1.6039
0.9733	6.4	800	1.6169
0.9837	6.8	850	1.6162
0.9766	7.2	900	1.6158
0.924	7.6	950	1.6164
1.0258	8.0	1000	1.6162

Framework versions

Transformers 4.44.2
Pytorch 2.0.0+cu117
Datasets 3.0.0
Tokenizers 0.19.1

RichardErkhov/tsavage68_-_IE_L3_1000steps_1e6rate_SFT-8bits

Original model description:

IE_L3_1000steps_1e6rate_SFT

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

RichardErkhov/
tsavage68_-_IE_L3_1000steps_1e6rate_SFT-8bits