felix
116 posts
We’re actively addressing the service-side issues raised in the comments and have increased their priority, including but not limited to: - Usage limits - Throughput - 429 errors due to capacity constraints - Availability on the BigModel platform Regarding reports of usage being consumed faster than expected, maintaining a rolling 1M-token context can consume a significant amount of quota. If this applies to you, we recommend manually lowering the Compact Size, for example to 256K, and setting Thinking Effort to a lower level, such as “low.” We’ll automatically map this setting to the closest equivalent across different harnesses.
我最近发现一个AI工具叫做WeChat，里面好多智能体，聊天都是免费的，识图能力还可以。 就是推理速度有点慢，经常会重复你的输入。
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models. images.nvidia.com/pdf/Open-Weigh…