May your gradient norms be smooth and monotonically decreasing.
@maximelabonne
-
Agentic AI Behaviors in Fitness Apps and Virtual Coaching
By
–
It's a good point, I can totally see people using a virtual coach with fitness apps. I'd bet on simple agentic behaviors, like filling out forms via tools/APIs.
-
Generative AI Enterprise Evolution: Top Use Cases 2023-2024
By
–
The State of Generative AI in the Enterprise Interesting report from Menlo Ventures that shows the evolution of Gen AI in companies from 2023 to 2024: • Uses cases: Code generation, chatbots, search, data, and meeting summarization are the top generative AI use cases in
-
RedPajama Paper: 46 Quality Filters for Pre-training Data
By
–
The new RedPajama paper is already a classic if you're into pre-training data It implements 46 quality filters with document-level and line-level granularity: • Natural language: fraction of all-caps words, terminal punctuation, etc.
• Repetitiveness: n-gram statistics
• -
Complete LLM Engineer Handbook: Updated Resources and Tools
By
–
Handy list of LLM-related resources with tools and articles by @panda_liyin It's pretty up-to-date and makes me want to work on the LLM course again GitHub: https://
github.com/SylphAI-Inc/LL
M-engineer-handbook
… -
Llama 3.1 70B Base Evaluation as Judge LLM
By
–
Really cool initiative, curious to see the results (is Llama 3.1 70B base the second-best judge LLM?)
-
Hugging Face SQL Console for Dataset Cleaning Guide
By
–
And @calebfahlgren wrote a short article on how to replicate it using Hugging Face's SQL console https://
huggingface.co/blog/cfahlgren
1/the-beginners-guide-to-cleaning-a-dataset
… I had no idea it was an option and it's a lot faster than my method -
Orca AgentInstruct Dataset: Cleaned Data Generation Techniques
By
–
Don't sleep on the Orca AgentInstruct dataset Here's a cleaned and reformatted version: https://
huggingface.co/datasets/mlabo
nne/orca-agentinstruct-1M-v1-cleaned
… I also wrote an article about this new data generation family of techniques on @huggingface a few months ago: https://
huggingface.co/blog/mlabonne/
agentic-datagen
… -
Unsloth AI Training Code Works Across Cloud Local Environments
By
–
It's a good point. The training code uses @UnslothAI and can run pretty much anywhere: cloud GPUs, Colab notebooks, local machines, etc. I'll add a blob about that in the readme.
-
LLM Engineers Handbook: Training and Deployment Guide Updated
By
–
I just updated the official repo of the LLM Engineer's Handbook, already sitting at 1.6k I recommend checking out the LLMOps pipelines, especially if you're a researcher. Train and deploy with AWS, monitor with Opik, and run with ZenML. GitHub: https://
github.com/PacktPublishin
g/LLM-Engineers-Handbook
…