Later this month, Carnegie Mellon's CMDragons small-size
robotic
soccer team will be competing again at RoboCup, to be held in Singapore.
CMDragons has tended to find their edge in their software as opposed to their
hardware. Their latest software
advantage will be their new "physics-based planning", using physics to decide how to
move and turn with the ball in order to maintain control. Previous control strategies
simply planned where the robot should move to and shoot from, assuming a ball placed
at the front center of the dribbler bar would stay there. The goal of Robocup is to create
a humanoid robotic soccer team to compete against human players in 2050. Manuela
Veloso, the professor who leads the Carnegie Mellon robotic soccer lab, "believe[s] that
the physics-based planning algorithm is a particularly noteworthy accomplishment" that
will take the effort one step closer to the collective goal.
mikhail-miguel on 09 Jul 23OpenAI Playground is a cutting-edge web-based tool designed to facilitate the construction and evaluation of predictive language models. Unlike ChatGPT, which excels at understanding context and can produce high-quality content even with vague prompts, OpenAI Playground requires specific prompts and settings for better-quality outputs. The tool boasts many purpose-specific preset prompts, ensuring seamless customisation through adjustable parameters such as the model, response length, and temperature. OpenAI Playground also accommodates diverse base language models, allowing users to handpick the one that perfectly resonates with their preferred conversational style. OpenAI Playground's Pros: Preset options for different tasks, such as text-to-command and translation Suited for specific use cases where preset models and settings are needed OpenAI Playground's Cons: Requires specific prompts and settings for better-quality outputs It can be challenging to use if you're not familiar with LLMs OpenAI Playground's Pricing: OpenAI Playground offers a comprehensive pricing structure based on the number of tokens used. 1 word = 1 token For example, Chat, model gpt-3.5-turbo costs $0.002 for 1K tokens, while InstructGPT, model Davinci costs $0.0200 for 1K tokens.
