RichardErkhov/ICT3214-Group5_-_MD5_gpt_neo_L_v1.0.0-8bits

Quantization made by Richard Erkhov.

Github

Discord

Request more models

MD5_gpt_neo_L_v1.0.0 - bnb 8bits

• Model creator: https://huggingface.co/ICT3214-Group5/
• Original model: https://huggingface.co/ICT3214-Group5/MD5_gpt_neo_L_v1.0.0/

Original model description:

library_name: transformers license: mit base_model: EleutherAI/gpt-neo-1.3B tags: - generated_from_trainer metrics: - rouge model-index: - name: MD5_gpt_neo_L_v1.0.0 results: []

MD5_gpt_neo_L_v1.0.0

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: 0.3792
• Rouge1: 0.0
• Rouge2: 0.0
• Rougel: 0.0

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: 2
• eval_batch_size: 2
• 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: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel
No log	0.9935	77	11.6516	0.0023	0.0	0.0023
No log	2.0	155	10.0057	0.0129	0.0	0.0129
No log	2.9935	232	11.0702	0.0	0.0	0.0
No log	4.0	310	0.6714	0.0	0.0	0.0
No log	4.9677	385	0.3792	0.0	0.0	0.0

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.0+cu121
• Datasets 2.14.4
• Tokenizers 0.19.1