Downloads · 30 days
0
parrot_en_es.13B.GPTQ
eduardem/parrot_en_es.13B.GPTQ
parrot_en_es.13B.GPTQ is a machine learning model from eduardem. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as apache-2.0.
This model is a variant of the original parrotenes model, utilizing 128-Act Order Quantization. It serves as an experiment to study the effects of quantization on language models.
Access
Public
Updated Sep 4, 2023
Repo size
—
Likes
0
Public
Repo makeup
Click a slice to open those files.
Other1.5 KB · 55%
At a glance
- Library
- transformers
- License
- apache-2.0
- Access
- Public
- Created
- Sep 4, 2023
- Updated
- Sep 4, 2023
- SHA
- a4f359f4
Model card
From the Hugging Face model README
GPTQ 4 bits with 128-Group size and Act Order Quantization
Introduction
This model is a variant of the original parrot_en_es model, utilizing 128-Act Order Quantization. It serves as an experiment to study the effects of quantization on language models.
Warning: Catastrophic Loss Through Quantization
Please be advised that the quantization process has resulted in catastrophic loss of functionality. The model is no longer capable of acting as a translator. It has issues with context understanding and exhibits hallucinatory behavior.
Original Model
For detailed usage instructions and prompt formatting, please refer to the original model card.
Usage
Given the current limitations, this quantized model is not recommended for translation tasks. However, it may still be of interest for research purposes, specifically in studying the effects of quantization on neural networks.
Contributing
If you have insights or suggestions on how to improve the model or mitigate the issues caused by quantization, please feel free to open an issue or submit a pull request.
- Library
- transformers
- License
- apache-2.0
- Languages
- en, es
- Created
- Sep 4, 2023
- Updated
- Sep 4, 2023