laitimes

Meta open source nuclear bomb level update: LLama open source second-generation model - 34b parameter model close to ChatGPT, 70b exceeds ChatGPT - full crushing:

author:AI Lab

Meta Open Source Nuclear Bomb Update: LLama Open Source Second Generation Model

- The 34 b parameter model is close to ChatGPT, and 70 b exceeds ChatGPT

- Full crushing: 70 B models are open source best in terms of code, reasoning, mathematics, etc

- Context 2k upgrade 4k

- The amount of training data has grown from ten trillion to twenty trillion tokens

Model parameters, weights, etc. are publicly available and can be used for commercial purposes, and can be used after applying for weights. #Artificial Intelligence # # Big Model # #Meta#

Meta open source nuclear bomb level update: LLama open source second-generation model - 34b parameter model close to ChatGPT, 70b exceeds ChatGPT - full crushing:
Meta open source nuclear bomb level update: LLama open source second-generation model - 34b parameter model close to ChatGPT, 70b exceeds ChatGPT - full crushing:
Meta open source nuclear bomb level update: LLama open source second-generation model - 34b parameter model close to ChatGPT, 70b exceeds ChatGPT - full crushing:
Meta open source nuclear bomb level update: LLama open source second-generation model - 34b parameter model close to ChatGPT, 70b exceeds ChatGPT - full crushing:
Meta open source nuclear bomb level update: LLama open source second-generation model - 34b parameter model close to ChatGPT, 70b exceeds ChatGPT - full crushing:
Meta open source nuclear bomb level update: LLama open source second-generation model - 34b parameter model close to ChatGPT, 70b exceeds ChatGPT - full crushing:

Read on