microsoft/encoder-decoder-slm
Efficient encoder-decoder architecture for small language models (≤1B parameters) with cross-architecture knowledge distillation and vision-language capabilities
This project helps machine learning engineers and researchers deploy more efficient and performant small language models (SLMs) for specific tasks like question answering or summarization. It takes in either text or text and images, and outputs generated text. The end user is a machine learning practitioner who needs to optimize the performance and efficiency of language models under 1 billion parameters, especially for deployment on edge devices.
No commits in the last 6 months.
Use this if you are developing or deploying small language models (under 1 billion parameters) and need to maximize their performance, throughput, and memory efficiency, particularly for applications on resource-constrained devices.
Not ideal if you are working with very large language models (over 1 billion parameters) or if your primary goal is to compete with the latest state-of-the-art SLMs without architectural optimization concerns.
Stars
32
Forks
3
Language
Python
License
MIT
Category
Last pushed
Feb 07, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/microsoft/encoder-decoder-slm"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
AI-Hypercomputer/maxtext
A simple, performant and scalable Jax LLM!
rasbt/reasoning-from-scratch
Implement a reasoning LLM in PyTorch from scratch, step by step
mindspore-lab/mindnlp
MindSpore + 🤗Huggingface: Run any Transformers/Diffusers model on MindSpore with seamless...
mosaicml/llm-foundry
LLM training code for Databricks foundation models
rickiepark/llm-from-scratch
<밑바닥부터 만들면서 공부하는 LLM>(길벗, 2025)의 코드 저장소