RichardErkhov/ArashMarandi_-_Llama-2-7b-chat-hf-4bits

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Llama-2-7b-chat-hf - bnb 4bits

• Model creator: https://huggingface.co/ArashMarandi/
• Original model: https://huggingface.co/ArashMarandi/Llama-2-7b-chat-hf/

Original model description:

library_name: transformers license: mit base_model: EleutherAI/gpt-neo-1.3B tags: - generated_from_trainer model-index: - name: Llama-2-7b-chat-hf results: []

Llama-2-7b-chat-hf

This model is a fine-tuned version of EleutherAI/gpt-neo-1.3B on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1051

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: 2e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	5	1.3871
3.5512	2.0	10	0.1261
3.5512	3.0	15	0.1115
0.1406	4.0	20	0.1050
0.1406	5.0	25	0.1016
0.0738	6.0	30	0.0973
0.0738	7.0	35	0.0974
0.0526	8.0	40	0.0992
0.0526	9.0	45	0.1030
0.025	10.0	50	0.1045
0.025	11.0	55	0.1055
0.0224	12.0	60	0.1059
0.0224	13.0	65	0.1054
0.0134	14.0	70	0.1051
0.0134	15.0	75	0.1051

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1