brevity_ss-promt_e30

This model is a fine-tuned version of TheBloke/Mistral-7B-Instruct-v0.2-GPTQ on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4920

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
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
3.9164	0.8	2	3.1402
2.3902	2.0	5	2.6152
3.053	2.8	7	2.3220
1.7893	4.0	10	1.9873
2.3626	4.8	12	1.8054
1.3909	6.0	15	1.5493
1.8073	6.8	17	1.3868
1.0338	8.0	20	1.1445
1.2723	8.8	22	0.9763
0.6819	10.0	25	0.7926
0.8371	10.8	27	0.7085
0.4868	12.0	30	0.6329
0.6522	12.8	32	0.5975
0.4052	14.0	35	0.5653
0.5732	14.8	37	0.5482
0.3641	16.0	40	0.5303
0.5239	16.8	42	0.5213
0.3372	18.0	45	0.5101
0.4896	18.8	47	0.5041
0.3174	20.0	50	0.4974
0.4645	20.8	52	0.4949
0.3046	22.0	55	0.4929
0.4505	22.8	57	0.4923
0.2972	24.0	60	0.4920

Framework versions

• PEFT 0.10.0
• Transformers 4.40.2
• Pytorch 2.1.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1