different outputs from hugging face space and local testing using diffuers

WHy i am getting very different output between tetsing in space and locally, like in space i got very good output and i even used same params locally here, even same seed, but very different outputs images = pipe(prompt, image=image, num_inference_steps=50, text_guidance_scale=7.5,image_guidance_scale=1.5, generator=generator).images