just software optimizations. still FP16 weights.
CODE
-
Compiler Optimizations and Memory Management for FP16 Models
By
–
compiler optimizations, better memory management etc. same FP16 weights.
-
Reverse Engineering Apple’s ScreenContinuityUI Framework and Swift Challenges
By
–
The actual app is just a thin shell around ScreenContinuityUI.framework. Swift really doesn't make reverge engineering easy, I miss the good old ObjC times. I see some Combine use too. There's also some code for Telemetry… guess that's for Apple to track how reliable the
-
ToolComp: New Tool Use Benchmark with Process Supervision
By
–
This is a new benchmark (ToolComp) that encompasses a broader range of Tool Use scenarios than prior benchmarks. Uniquely, we are further evaluating the models utilizing process supervision labels. We also split the benchmark into Enterprise and Chat use cases to differentiate
-
GRADIO-CODER: Instruction Fine-Tuned StarCoder for Gradio Development
By
–
GRADIO-CODER model: https://
huggingface.co/matrixglitch/G
RADIO-CODER
… This language model is the version 0.0 of a Gradio Coding Assistant. It is an instruction fine-tuned version of StarCoder that is designed to provide assistance to developers who use gradio. -
Unified Memory Management Across AI Models Implementation
By
–
I know, I need that so I can deal with unified memory across models
-
Sonnet 3.5 vs O1: Multi-file Editing Performance Comparison
By
–
i agree sonnet 3.5 for coding. except o1 seems to be better at multi-file edits.
-
Making AI Templates Successful: Ideas and Strategy
By
–
https://
x.com/hwchase17/stat
us/1836790660063510701
… gave a shout out in blog! I'm bullish on templates and want to make them successful. what ideas/advice would you have for this? -
Three Cognitive Architecture Templates Launch: ReAct, RAG, Data Enrichment
By
–
we're launching a small number (3) of templates for common cognitive architectures
— Harrison Chase (@hwchase17) 19 septembre 2024
these are configurable (choose LLM/vector store of your choice)
Starting with:
– ReAct agent
– RAG chatbot
– data enrichment (research) agent
What should we add next? https://t.co/bqOclMrhUNwe're launching a small number (3) of templates for common cognitive architectures these are configurable (choose LLM/vector store of your choice) Starting with:
– ReAct agent
– RAG chatbot
– data enrichment (research) agent What should we add next? -
Standard Deviations of Shoot AI Model Weights and Convolution Matrix
By
–
Wow, I just finished working on something really cool. I calculated the standard deviations (and many other things) of the layer weights of the 500 different models that people trained on Shoot AI. From that, I deduced, in ULTRA BRIEF, a kind of convolution matrix and