On this video, we shall be wanting on the paper titled “QLORA: Environment friendly Finetuning of Quantized LLMs” which introduces a brand new quantization method referred to as QLoRA that permits the coaching & fine-tuning of (33B, 13B) LLMs on Shopper GPUs by drastically decreasing the Reminiscence Wanted. We shall be wanting on the demo of the mannequin on HuggingFace after which will have a look at code examples on the right way to fine-tune it.
▬▬▬▬▬▬▬▬▬▬▬▬▬▬ CONNECT ▬▬▬▬▬▬▬▬▬▬▬
☕ Purchase me a Espresso: https://ko-fi.com/promptengineering
|🔴 Help my work on Patreon: Patreon.com/PromptEngineering
🦾 Discord: https://discord.com/invite/t4eYQRUcXB
▶️️ Subscribe: https://www.youtube.com/@engineerprompt?sub_confirmation=1
📧 Enterprise Contact: [email protected]
💼Consulting: https://calendly.com/engineerprompt/consulting-call
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
LINKS:
QLoRA Paper: https://arxiv.org/pdf/2305.14314.pdf
HuggingFace Blogpost: https://huggingface.co/blog/4bit-transformers-bitsandbytes
QLoRA utilization Pocket book: https://colab.research.google.com/drive/1ge2F1QSK8Q7h0hn3YKuBCOAS0bK8E0wf?usp=sharing
QLoRA fine-tuning Pocket book: https://colab.research.google.com/drive/1VoYNfYDKcKRQRor98Zbf2-9VQTtGJ24k?usp=sharing
Guanaco 33B Demo: https://huggingface.co/spaces/uwnlp/guanaco-playground-tgi
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
All Fascinating Movies:
Every part LangChain: https://www.youtube.com/playlist?list=PLVEEucA9MYhOu89CX8H3MBZqayTbcCTMr
Every part LLM: https://youtube.com/playlist?list=PLVEEucA9MYhNF5-zeb4Iw2Nl1OKTH-Txw
Every part Midjourney: https://youtube.com/playlist?list=PLVEEucA9MYhMdrdHZtFeEebl20LPkaSmw
AI Picture Technology: https://youtube.com/playlist?list=PLVEEucA9MYhPVgYazU5hx6emMXtargd4z
