# Lightweight fine-tuning

QVikhr-3-1.7B-Instruction-noreasoning is an instruction model based on Qwen/Qwen3-1.7B, trained on the Russian dataset GrandMaster2, and is specifically designed for efficient processing of Russian and English texts.

A Russian abstract generation model fine-tuned on the gazeta dataset based on google/mt5-small

A fine-tuned Mistral 7B model trained on 250,000 Portuguese chat datasets, optimized for Portuguese chat scenarios

Mamba is an efficient language model based on the State Space Model (SSM), with the ability to model sequences with linear time complexity.

Tiny-Vicuna-1B is a lightweight model fine-tuned from TinyLLama 1.1B using the WizardVicuna dataset, designed for early-stage experimental iterations.

Tiny Vicuna 1B is a fine-tuned version based on TinyLlama using the WizardVicuna dataset, compatible with the Vicuna-v1.5 series, suitable for early experimental iterations.

A pre-trained language model based on the Llama2 architecture, suitable for English text generation tasks

KoreanLM is an open-source project dedicated to developing Korean language models, aiming to address the scarcity of Korean learning resources and inefficient tokenization processing.

KoreanLM is an open-source language model project specifically optimized for Korean, designed for Korean grammar and vocabulary characteristics, providing efficient tokenization solutions

This is an efficient micro T5 model optimized for Spanish, with a small size (<29MB), suitable for CPU usage, requires fine-tuning before use

A model fine-tuned based on Salesforce/codet5-small for generating docstrings for Python functions

An optimized and calibrated mini-T5 model suitable for text generation and transformation tasks, featuring lightweight and efficient characteristics.

This is a modified version of EleutherAI's GPT-Neo (2.7B parameter version) that supports text generation and model fine-tuning on Colab or equivalent desktop GPUs.