# Multi-turn Dialogue Optimization

A high-performance roleplay and creative writing reasoning model launched by ArliAI, fine-tuned based on the 32-billion-parameter QwQ-32B foundation model, focusing on long-dialogue coherence and creative output

Typhoon2.1-Gemma3-12B is a 12-billion-parameter Thai large language model based on the Gemma3 architecture, supporting 128K context length and function calling capabilities.

GGUF format quantized version of Qwen3-8B, provided by TensorBlock, compatible with llama.cpp

Qwen-2.5-7B-base-RAG-RL is a large language model with 7B parameters trained from scratch on an unknown dataset, incorporating Retrieval-Augmented Generation (RAG) and Reinforcement Learning (RL) technologies.

This is a Bert model optimized for the COVID-19 domain, developed by the Multi2ConvAI project

Mistral 7B OpenOrca is a large language model based on the Mistral-7B architecture, trained by the OpenOrca team, focusing on text generation tasks.

This is a distilled model based on Llama-70B, converted to MLX format via mlx-lm and quantized to 4 bits.

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.

Based on Google's Gemma-3-12B model, specifically fine-tuned for Telugu through supervised learning, supporting high-quality text generation and dialogue tasks.

Gemma 3 4B Instruct is a 4-billion-parameter large language model developed by Google, focusing on text generation and comprehension tasks.

The Doge model employs dynamic masked attention mechanisms for sequence transformation and can use multi-layer perceptrons or cross-domain mixture of experts for state transitions.

A Mixture of Experts model designed to serve as the soul of AI virtual anchors, combining code generation, instruction following, and multi-turn dialogue capabilities

This model is converted from google/gemma-3-4b-it to GGUF format via llama.cpp, suitable for local deployment and inference.

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

An 8B-parameter open-source large language model released by NVIDIA, based on the Llama-3 architecture, offering multiple quantization versions

A powerful fine-tuned AI model based on HomerSlerp6-7B, integrating multiple Qwen2.5-7B series models to deliver high adaptability, deep knowledge, and engaging interaction experiences

This model is a GGUF format conversion based on mlabonne/gemma-3-27b-it-abliterated, suitable for inference tasks in the llama.cpp framework.

A large language model based on 12B parameters, supporting text generation tasks, released under the Apache-2.0 license.

Gemma 3.4B IT is an open-source large language model developed by Google. This version is the 4-bit quantized version converted to GGUF format via llama.cpp.

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 GGUF format model converted from google/gemma-3-12b-it, suitable for the llama.cpp framework.

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.

A large-scale Japanese-English hybrid MoE language model developed by Japan's National Institute of Informatics, supporting an 8x13B parameter scale with instruction fine-tuning optimization

DeepHermes 3 Preview is the latest version of Nous Research's flagship Hermes series large language models, combining reasoning and standard response modes with improved annotation, judgment, and function calling capabilities.

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.

A large language model fine-tuned based on Qwen 2.5 14B Instruct, trained using the Gamma Corpus v2-50k dataset

A Qwen 2.5 model fine-tuned on the Gamma Corpus, designed to surpass other models of the same scale

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 dialogue model fine-tuned based on Qwen 2.5 7B Instruct model, trained with Gamma Corpus v2-100k

A merged model based on Nitral-AI and LatitudeGames' two 12B parameter models, using the slerp fusion method, with powerful text generation capabilities

Hermes 3 3B is a small yet powerful new member of the Hermes series LLMs by Nous Research, focusing on aligning LLMs with users and empowering end-users with strong guidance and control capabilities.

Starcannon-Unleashed-12B-v1.0 is a merged model combining the strengths of Starcannon v3 and NemoMix Unleashed 12B, designed to enhance long-form text generation and roleplaying capabilities.

A 123B-parameter model fine-tuned based on Mistral-Large-Instruct, aiming to reproduce Claude 3's text generation quality

Ichigo-llama3s is a large language model series that supports both audio and text input, focusing on enhancing speech understanding capabilities and user interaction experience.

A language model optimized for Japanese role-playing scenarios using QLoRA technology, based on the cyberagent/calm3-22b-chat model

A text generation model fine-tuned based on Qwen1.5 32B, aiming to reproduce the text quality of Claude 3 series models

BERT-based Arabic sentence embedding model supporting semantic text similarity tasks

A Korean large language model fine-tuned based on Meta-Llama-3.1, specializing in command understanding and generation tasks

A prompt enhancement model fine-tuned based on LaMini-Flan-T5-248M, used for expanding and optimizing text prompt descriptions