# Instruction Fine-tuning

Kanana-1.5-8B-Instruct-2505 is an 8B-parameter instruction fine-tuned language model developed by Kakao Corp, suitable for text generation tasks.

marin-8b-instruct is an 8B-parameter-scale instruction fine-tuned language model suitable for text generation tasks.

OLMo-2-0425-1B-Instruct is a 1-billion-parameter instruction-finetuned language model developed by AllenAI, focused on text generation tasks.

OLMo 2 1B Instruct Edition is a post-training variant of the OLMo-2-0425-1B-RLVR1 model, optimized through supervised fine-tuning, DPO training, and RLVR training to achieve state-of-the-art performance across multiple tasks.

This is the GGUF quantized version of the Josiefied-Qwen3-4B-abliterated-v1 model, suitable for local deployment and execution.

OLMo 2 1B is a post-training variant of the allenai/OLMo-2-0425-1B-RLVR1 model, undergoing supervised fine-tuning, DPO training, and RLVR training, aiming to achieve state-of-the-art performance across multiple tasks.

StableLM Zephyr 3B is a 3-billion-parameter instruction-tuned model trained on public datasets, synthetic datasets, and Direct Preference Optimization (DPO), delivering excellent performance.

A quantized version based on Gemma 2.9B, optimized using llama.cpp, suitable for running in LM Studio.

Badger Writer is a normalized Fourier task superposition model based on multiple Llama 3 8B models, specializing in text generation tasks, particularly excelling in creative writing and instruction following.

A Japanese-enhanced large language model based on the Gemma-2 architecture, significantly improving Japanese capabilities while retaining original English proficiency

The Gemma-2-Llama-Swallow series is built through continued pre-training of the gemma-2 model, significantly enhancing Japanese language processing capabilities while retaining original English proficiency.

OLMo 2 1B is the smallest model in the open language model series released by the Allen Institute for Artificial Intelligence, based on OLMo-mix-1124 pre-training and further trained with the Dolmino-mix-1124 dataset during the intermediate training phase.

VideoChat-R1-thinking_7B is a multimodal model based on Qwen2.5-VL-7B-Instruct, focusing on video-text-to-text tasks.

Multilingual E5 large instruction model, supporting text embedding and classification tasks in multiple languages with strong cross-language capabilities.

Gemma 3 is a lightweight open-source multimodal model introduced by Google, built on the same technology as Gemini, supporting text and image inputs to generate text outputs.

This model is a text generation model fine-tuned based on Qwen2.5-14B-Instruct and trained using the TRL library.

A creative writing model based on Gemma 3 12B, excelling in narration and scene construction with a novelistic style

GGUF format model file converted from allura-org/Gemma-3-Glitter-4B, optimized with imatrix quantization

Doge 320M Instruct is a lightweight language model based on dynamic masked attention, trained with supervised fine-tuning (SFT) and direct preference optimization (DPO), suitable for question-answering and dialogue tasks.

This is a quantized version of TheDrummer/Fallen-Gemma3-4B-v1 model, processed using llama.cpp, suitable for text generation tasks.

Mistral Small 3.1 Instruct 24B is a large language model based on instruction fine-tuning, focusing on text generation tasks.

An instruction-tuned variant based on Qwen2.5 Bakeneko 32B, enhanced with Chat Vector and ORPO optimization for improved instruction-following capabilities, excelling in Japanese MT-Bench.

Qwen2-1.5B-Instruct is a 1.5 billion parameter instruction fine-tuned large language model released by Alibaba Cloud, suitable for Q&A and dialogue tasks.

T3Q-qwen2.5-14b-v1.2-e2 is a post-trained version based on the Qwen/Qwen2.5-14B-Instruct-1M model, using LoRA-8-4-0.0001-cosine-32-16 configuration and trained on train_data_v1.2.

This is a quantized model based on Qwen2.5-14B-Instruct-1M, converted to GGUF format, suitable for the llama.cpp framework.

A text-only version converted from google/gemma-3-12b-it, with visual components removed, focusing on text generation tasks

Gemma 3.4B IT is a 3.4 billion parameter large language model developed by Google, focusing on the instruction-tuned version, suitable for various natural language processing tasks.

TraceBack 12b is a 4bit quantized version based on the Mistral-Nemo-Instruct architecture, focusing on instruction-following and chain-of-thought reasoning tasks.

An 8B-parameter language model based on the Llama 3.1 architecture, created by merging multiple specialized models, excelling in text generation tasks.

Kanana is a bilingual (Korean/English) language model series developed by Kakao. This 2.1B parameter version outperforms similar models in Korean while maintaining efficient computational costs.

Hiber-Multi-10B-Instruct is an advanced multilingual large language model based on Transformer architecture, supporting multiple languages with 10 billion parameters, suitable for text generation tasks.

A 14B-parameter large language model focused on instruction-following tasks, supporting text generation capabilities.

A dialogue model fine-tuned based on Llama-3-8B, supporting multiple quantization versions, suitable for tasks such as chatting, reasoning, and role-playing.

GuardReasoner 1B is a version fine-tuned via R-SFT and HS-DPO based on meta-llama/Llama-3.2-1B, focusing on classification tasks for analyzing human-AI interactions.

GuardReasoner 8B is a fine-tuned model based on meta-llama/Llama-3.1-8B, specializing in reasoning-based LLM safety protection

GGUF quantized version of Mistral-Small-24B-Instruct-2501, suitable for local deployment and text generation tasks.

DeepSeer is a vision-language model developed based on the DeepSeek-R1 model, supporting chain-of-thought reasoning and trained through dialogue templates for visual models.

Mistral-7B-Instruct-v0.3 is a large language model fine-tuned for instruction following based on Mistral-7B-v0.3, supporting function calls and extended vocabulary.

A vision-language model based on Microsoft's Phi-1.5 architecture, combined with CLIP for image processing capabilities

A multimodal large language model developed based on the paper 'Task Preference Optimization: Improving Multimodal Large Language Models through Visual Task Alignment'