44 lines
1.5 KiB
Markdown
44 lines
1.5 KiB
Markdown
|
---
|
||
|
|
license: mit
|
||
|
|
language:
|
||
|
|
- en
|
||
|
|
base_model:
|
||
|
|
- microsoft/phi-4
|
||
|
|
---
|
||
|
|
# Model Card for Microsoft-phi-4-Instruct-AutoRound-GPTQ-4bit
|
||
|
|
|
||
|
|
## Model Overview
|
||
|
|
|
||
|
|
**Model Name**: Microsoft-phi-4-Instruct-AutoRound-GPTQ-4bit
|
||
|
|
**Model Type**: Instruction-tuned, Quantized GPT-4-based language model
|
||
|
|
**Quantization**: GPTQ 4-bit
|
||
|
|
**Author**: Satwik11
|
||
|
|
**Hosted on**: Hugging Face
|
||
|
|
|
||
|
|
## Description
|
||
|
|
|
||
|
|
This model is a quantized version of the Microsoft phi-4 Instruct model, designed to deliver high performance while maintaining computational efficiency. By leveraging the GPTQ 4-bit quantization method, it enables deployment in environments with limited resources while retaining a high degree of accuracy.
|
||
|
|
|
||
|
|
The model is fine-tuned for instruction-following tasks, making it ideal for applications in conversational AI, question answering, and general-purpose text generation.
|
||
|
|
|
||
|
|
## Key Features
|
||
|
|
|
||
|
|
- **Instruction-tuned**: Fine-tuned to follow human-like instructions effectively.
|
||
|
|
- **Quantized for Efficiency**: Uses GPTQ 4-bit quantization to reduce memory requirements and inference latency.
|
||
|
|
- **Pre-trained Base**: Built on the Microsoft phi-4 framework, ensuring state-of-the-art performance on NLP tasks.
|
||
|
|
|
||
|
|
## Use Cases
|
||
|
|
|
||
|
|
- Chatbots and virtual assistants.
|
||
|
|
- Summarization and content generation.
|
||
|
|
- Research and educational applications.
|
||
|
|
- Semantic search and knowledge retrieval.
|
||
|
|
|
||
|
|
## Model Details
|
||
|
|
|
||
|
|
### Architecture
|
||
|
|
|
||
|
|
- **Base Model**: Microsoft phi-4
|
||
|
|
- **Quantization Technique**: GPTQ (4-bit)
|
||
|
|
- **Language**: English
|
||
|
|
- **Training Objective**: Instruction-following fine-tuning
|