joelxiangnanchen

🤗Huggingface Model Downloader

Considering the lack of multi-threaded download support in the official huggingface-cli, and the inadequate error handling in hf_transfer, this command-line tool smartly utilizes wget or aria2 for LFS files and git clone for the rest.

Features

• ⏯️ Resume from breakpoint: You can re-run it or Ctrl+C anytime.
• 🚀 Multi-threaded Download: Utilize multiple threads to speed up the download process.
• 🚫 File Exclusion: Use --exclude or --include to skip or specify files, save time for models with duplicate formats (e.g., *.bin or *.safetensors).
• 🔐 Auth Support: For gated models that require Huggingface login, use --hf_username and --hf_token to authenticate.
• 🪞 Mirror Site Support: Set up with HF_ENDPOINT environment variable.

joelxiangnanchen

Reinforcement Learning for Language Models

Yoav Goldberg, April 2023.

Why RL?

With the release of the ChatGPT model and followup large language models (LLMs), there was a lot of discussion of the importance of "RLHF training", that is, "reinforcement learning from human feedback". I was puzzled for a while as to why RL (Reinforcement Learning) is better than learning from demonstrations (a.k.a supervised learning) for training language models. Shouldn't learning from demonstrations (or, in language model terminology "instruction fine tuning", learning to immitate human written answers) be sufficient? I came up with a theoretical argument that was somewhat convincing. But I came to realize there is an additional argumment which not only supports the case of RL training, but also requires it, in particular for models like ChatGPT. This additional argument is spelled out in (the first half of) a talk by John Schulman from OpenAI. This post pretty much