NEWS 新闻NEWS 新闻

【OpenAI Sora模型生成视频效率待提升:1分钟视频渲染超1小时】全球知名人工智能研究机构OpenAI近日推出了一款创新的文本生成视频模型——Sora。该模型凭借其将用户提示词转化为逼真视频的能力,引发了业界广泛关注。然而,根据Reddit社区用户的实际体验反馈,Sora在生成视频方面的效率尚存瓶颈。据报道,使用Sora生成短短1分钟的视频,需要耗时超过1小时进行渲染,这无疑限制了其在实际应用中的便捷性。

尽管Sora的视频生成技术在视觉效果上展现出高水准,但目前公众的体验范围仍被限制在研究人员预先挑选的示例上,无法自由输入自定义提示词进行创作。最长的公开演示视频长度仅为17秒,这在一定程度上反映出Sora在处理更长和更复杂内容时的挑战。

OpenAI的这一技术创新虽然在人工智能领域迈出了重要一步,但目前的效率问题也暴露出技术在实际落地时的局限。未来,OpenAI可能需要进一步优化算法,提升Sora的渲染速度,以满足用户对于实时性和效率的需求。对此,OpenAI尚未发表官方评论,但该消息已在科技和媒体圈内引起了热议,人们期待着Sora在性能优化后的表现。

英语如下:

**News Title:** “OpenAI’s New Model Sora Sparks Debate: Renders 1-Minute Videos in Over 1 Hour”

**Keywords:** OpenAI Sora, video generation, rendering time

**News Content:**

OpenAI, a renowned global AI research institution, recently unveiled its innovative text-to-video model called Sora. The model, capable of transforming user prompts into realistic videos, has attracted considerable industry attention. However, feedback from Reddit users indicates that Sora faces efficiency challenges in video generation. According to reports, the rendering of a mere 1-minute video with Sora takes over an hour, which significantly hampers its practical usability.

Despite Sora’s high visual standards in video creation, the public’s experience is currently limited to pre-selected examples curated by researchers, with no option to input custom prompts for personal creativity. The longest publicly demonstrated video is only 17 seconds, suggesting difficulties that Sora encounters when handling longer and more complex content.

While this technological innovation from OpenAI represents a significant step forward in the AI domain, the current efficiency issues expose limitations in practical implementation. Going forward, OpenAI might need to optimize its algorithms to enhance Sora’s rendering speed to meet users’ demands for real-time and efficiency. OpenAI has yet to issue an official comment on this matter, but the news has sparked heated discussions within the tech and media circles. There is anticipation for Sora’s improved performance following optimization.

【来源】https://www.ithome.com/0/751/364.htm

Views: 1

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注