If you are working on the NeurIPS LLM Efficiency Challenge, Dolly 15k support is now merged: https://
github.com/Lightning-AI/l
it-gpt/pull/430
… python scripts/prepare_dolly.py
python finetune/lora.py –data_dir data/dolly/
LLMS
-
Dolly 15k Support Merged for NeurIPS LLM Efficiency Challenge
By
–
-
LangChain Chat Updates New Pull Request Integration
By
–
always was: https://
github.com/langchain-ai/c
hat-langchain/pull/110
… -
Balancing Long-Term Vision with Near-Term Action in Data Strategy
By
–
Balancing long-term vision with near-term action with Vercel’s VP of Data https://
spoti.fi/3OsxsgH #AI #MachineLearning #DeepLearning #LLMs #DataScience -
The Sin of Indistinguishable from Magic in LLM Hype
By
–
BTW my diagnosis of the LLM hype is the "indistinguishable from magic" sin. LLMs did something surprising and for which we humans had no obvious mental model, so LLMs quickly became capable, in our minds, of anything imaginable.
https://technologyreview.com/2017/10/06/241837/the-seven-deadly-sins-of-ai-predictions/
… -
Transformer Pioneer Launches Adaptable AI Models Research Lab
By
–
A Godfather of Transformers Launches Lab Developing More Adaptable AI Models @theinformation https://
theinformation.com/articles/a-god
father-of-transformers-launches-lab-developing-more-adaptable-ai-models-databricks-and-openai-in-dealmaking-mode
… -
LLM Hype Cycle Peaks as Tech Cycles Grow Larger
By
–
Tech hype cycles are getting larger in amplitude and frequency. Seems to me LLM hype has just passed peak. Yet another case of "Looking for 愛 in all the wrong places"? [[Yes, it is a really bad kanji/romaji pun…]]
-
Platypus: Open-Source LLM Achieves Top Leaderboard Performance
By
–
Platypus, a new open-source LLM at the top of the leaderboard: https://
arxiv.org/abs/2308.07317 Key points are
1) a curated dataset: removing similar & duplicate questions
2) finetuning and merging Low Rank Approximation (LoRA) modules: focusing on the non-attention modules -
ETH Zürich DLSC Course Introduction and Deep Learning
By
–
ETH Zürich DLSC: Course Introduction https://
bit.ly/47oy2F1 #AI #MachineLearning #DeepLearning #LLMs #DataScience -
Efficient Fine-Tuning Llama 7B on Single GPU Workshop
By
–
Upcoming #workshop w/ @DeepLearningAI_
: Efficiently Fine-Tune #Llama7b on a Single GPU! We'll discuss the challenges of #finetuning LLMs and show you how to tackle them. Topics include: parameter efficient tuning, deployment strategies, RAG & more. https://
eventbrite.com/e/efficient-fi
ne-tuning-for-llama-7b-on-a-single-gpu-tickets-696331224437?aff=Psm
… -
LLMs for Structured Data Extraction and Web Automation
By
–
An underrated aspect of LLMs is using them for structured data extraction Extracting knowledge triplets is a great use case! I gave it our most recent blog post about @MultiON_AI (
https://
blog.langchain.dev/multion-x-lang
chain-powering-next-gen-web-automation-navigation-with-ai/
…) and it came up with the below – how did it do @DivGarg9 ??