loqira/Qwen3.5-0.8B-GEC-KAZ-RUS-ENG

Set Ollama num_ctx to 2048 to match the training context window

Document 2048-token training context and recommend chunking long input

Fix stop-token rendering in the model-details table

Add Qwen3.5-0.8B GEC model: Q4_0 GGUF weights, model card, and Ollama Modelfile

initial commit