模型多源确认

Fulcrum 使用 Fireworks API 训练 Echo 模型

Congrats to the @Fulcrum_inc team on Echo! Final training runs were done entirely via Fireworks' ser...

精选理由

Fulcrum 用 Fireworks API 和 Kimi K3 训练 Echo,成本仅 4500 美元,还分享了角色训练技巧。

Fulcrum 团队完成了 Echo 模型的训练,最终训练完全通过 Fireworks 的无服务器训练 API 完成。他们使用了 Kimi K3 LoRA(rank 32)进行训练,总训练成本为 4500 美元。团队在技术文档中详细介绍了"角色训练后"的方法。

原文 · Fireworks AI

Congrats to the @Fulcrum_inc team on Echo! Final training runs were done entirely via Fireworks' ser...

Congrats to the @Fulcrum_inc team on Echo! Final training runs were done entirely via Fireworks' serverless training API, on a Kimi K3 LoRA (rank 32). Make sure to read their writeup, which includes the "persona post-training" approach and how they landed a $4.5k total training cost. Fulcrum @fulcrum_inc For our final training run and deployment we used @FireworksAI_HQ 's with Kimi K3 - it was quite convenient 🔗 View Quoted Tweet 💬 0 🔄 0 ❤️ 2 👀 471 ⚡

  • Aran Komatsuzaki (论文推介)10-01 04:59原文
  • LangChain10-01 14:49原文
  • arXiv: DeepSeek09-29 03:19原文
  • IT之家09-30 10:19原文
  • Xianbao Qian09-30 18:07原文