sunrainyg/RandOpt

• highreadme#1

  Reposition README to clarify LLM post-training focus

  原因：

  当前

  # RandOpt
  **Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights**

  复制粘贴的修复

  复制

  # RandOpt
  **Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights**
  
  This repository provides the official codebase for RandOpt, a method for efficiently post-training large language models by creating diverse task experts around pretrained weights.

  待办完成忽略
• highlicense#2

  Add a LICENSE file to the repository

  原因：

  复制粘贴的修复

  复制

  Add a `LICENSE` file to the repository root, specifying the chosen open-source license (e.g., MIT, Apache-2.0).

  待办完成忽略
• mediumhomepage#3

  Set the repository homepage URL

  原因：

  复制粘贴的修复

  复制

  Set the repository homepage URL to the project page: `https://yulugan.github.io/neural-thickets/`

  待办完成忽略

本次扫描解析到的品类 GEO 通道：google/gemini-2.5-flash, deepseek/deepseek-v4-flash

• 品类问题

  How to efficiently post-train large language models for specific tasks using diverse experts?

  你：未被推荐
  AI 推荐顺序：

  1. Hugging Face PEFT
  2. bitsandbytes
  3. Fairseq
  4. Hugging Face transformers
  5. PyTorch
  6. TensorFlow
  7. Hugging Face TRL
  8. trlx

  AI 推荐了 8 个替代方案，却始终没点名 sunrainyg/RandOpt。这就是要补上的差距。

  查看 AI 完整回答
• 品类问题

  Seeking methods for black-box optimization to enhance reasoning capabilities in transformer models.

  你：未被推荐
  AI 推荐顺序：

  1. Spearmint
  2. BoTorch
  3. GPflowOpt

  AI 推荐了 3 个替代方案，却始终没点名 sunrainyg/RandOpt。这就是要补上的差距。

  查看 AI 完整回答

• Metadata completeness

  warn

  建议：

• README presence

  pass

• Compared to common alternatives in this category, what is the core differentiator of sunrainyg/RandOpt?

  pass
  ✓AI 明确点名了 sunrainyg/RandOpt

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• If a team adopts sunrainyg/RandOpt in production, what risks or prerequisites should they evaluate first?

  pass
  ✓AI 明确点名了 sunrainyg/RandOpt

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• In one sentence, what problem does the repo sunrainyg/RandOpt solve, and who is the primary audience?

  pass
  ✓AI 明确点名了 sunrainyg/RandOpt

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？