# Natural language generation

A creative writing model based on Gemma 3 27B, composed of a 50% mix of 27B IT and 27B PT, excelling in natural and fluent text expression

Vision-enhanced version based on soob3123/amoral-gemma3-12B, combining Gemma3-12B large language model with visual encoder for multimodal tasks

A quantized and optimized model based on the Meta-Llama-3.1 architecture, supporting multiple languages, suitable for business and research scenarios, while reducing resource requirements and maintaining high performance.

A LoRa fine-tuned model developed based on FLUX1-dev, specifically designed for generating soft pastel-style comic covers

This model is an image captioning model based on the BLIP architecture, specifically fine-tuned for high-level scene descriptions.

LaMini-T5-738M is an instruction-tuned model based on t5-large, fine-tuned on the LaMini-instruction dataset, with 738M parameters, and is one of the LaMini-LM series models.

LaMini-Flan-T5-783M is an instruction-following model fine-tuned on the LaMini-instruction dataset based on google/flan-t5-large, with 783M parameters, belonging to the LaMini-LM series.

LaMini-GPT-1.5B is a large language model fine-tuned based on the GPT-2-xl architecture, belonging to the LaMini-LM series, focusing on instruction-following tasks

A keyword-to-text generation model based on T5-small architecture that converts input keywords into coherent sentences

A T5-based model capable of converting input keywords into coherent sentences.

A T5-based model capable of converting input keywords into complete sentences.

Chinese version of T5-large model specialized in natural language conversion tasks