h2ogpt-16k-codellama-7b-instruct-finetuned-1-2024-03-06-16

This model is a fine-tuned version of h2oai/h2ogpt-16k-codellama-7b-instruct on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0141

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: 0.0003
• train_batch_size: 1
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 64
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.6967	1.17	20	0.4241
0.2982	2.33	40	0.2148
0.1461	3.5	60	0.1113
0.0856	4.66	80	0.0586
0.0631	5.83	100	0.0419
0.0577	6.99	120	0.0318
0.064	8.16	140	0.0464
0.0444	9.33	160	0.0308
0.0463	10.49	180	0.0284
0.0372	11.66	200	0.0276
0.0345	12.82	220	0.0253
0.0384	13.99	240	0.0230
0.0313	15.15	260	0.0201
0.0246	16.32	280	0.0168
0.0196	17.49	300	0.0141
0.0166	18.65	320	0.0140
0.0146	19.82	340	0.0141

Framework versions

• Transformers 4.38.1
• Pytorch 2.2.1+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2