歡迎 Llama 3：Meta 全新開源大型語言模型正式登陸 Hugging Face
Original: Welcome Llama 3 - Meta's new open LLM
Meta officially released Llama 3, the next generation of its open-source large language models, on April 18, 2024. The initial release…
Meta 正式推出全新一代開源大型語言模型 Llama 3，首波釋出 8B 與 70B 兩種參數版本。此版本採用 128k 詞彙量的新分詞器、支援 8K 上下文長度，並在超過 15 兆 Token 的超大資料集上進行訓練，程式碼與推理能力皆有飛躍性提升。Hugging Face 生態系已在第一時間全面整合，支援 Llama 3 的部署、微調與推論，為開源 AI 社群注入強大動力。
Meta officially released Llama 3, the next generation of its open-source large language models, on April 18, 2024. The initial release includes two parameter scales — 8B (8 billion) and 70B (70 billion) — both available in pretrained (Base) and instruction-tuned (Instruct) versions.
Free shows the 3-line summary; Pro unlocks the full deep summary (~300 words) so you never have to click through.
See Pro plans →Want the original English / full article?
Read on Hugging Face Blog →Summaries are AI-generated; the original article is authoritative.