# Text generation optimization

This is a fine-tuned version based on the Qwen-14B model, trained using the GRPO method, suitable for text generation tasks.

A 500-million-parameter text generation model based on the Llama architecture, specifically designed for story creation.

Muse-12B is a text generation model with 12B parameters, developed by LatitudeGames, aiming to provide high-quality text generation capabilities.

MagTie-v1-12B is a 12B-parameter language model merged using the DARE TIES algorithm, combining the strengths of multiple pre-trained models

Mistral-Small-24B-Instruct-2501 is a 24B-parameter instruction-finetuned large language model supporting multilingual text generation tasks.

Llama-xLAM-2-70b-fc-r is a large language model released by Salesforce, based on the Llama 2 architecture with 70 billion parameters.

This is a multi-model fusion version based on the Gemma-2-9B series models, merged using the DARE TIES method, integrating the strengths of multiple Gemma variants.

A text generation model merged from Gemma-2B and Gemma-9B models using the SLERP method

This is a result of merging multiple 12B-parameter-scale language models using the mergekit tool through the model inventory method

A 70B-parameter large language model based on the Llama3 architecture, specially optimized

The GGUF format version of Llama-3.2-1B-Instruct, providing broader support and better performance.

This is the result of llama.cpp imatrix quantization based on the nvidia/Mistral-NeMo-Minitron-8B-Base model, providing more options for model usage and deployment.

Llamacpp imatrix quantized version of WizardLM-2-7B-abliterated, offering multiple quantization options for different hardware configurations.

LOLA is an ultra-large-scale multilingual large model based on the sparse Mixture-of-Experts (MoE) Transformer architecture, supporting over 160 languages, with competitive advantages in natural language generation and understanding tasks.

FuseLLM-7B is a unified model that integrates knowledge from multiple open-source large language models, combining the capabilities of LLMs with different architectures through knowledge fusion technology.

This is a high-density merged model based on the Yi-34B-200K foundation model, integrating multiple homologous models through the DARE Ties method, featuring 200K long-context processing capability.

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

An experimental model for correcting spelling errors and punctuation in English and German

FLAN-T5 XXL is a variant of Google's T5 model, fine-tuned on over 1,000 additional tasks, supports multiple languages, and outperforms the original T5 model.

This reward model is trained to predict which generated answer human evaluators would prefer for a given question.

A Dutch language model based on the medium-scale GPT-2 version, with only the vocabulary embedding layer retrained for Dutch adaptation.