We love watching the community build with The AI Scientist since its release! We have a new community section in our repo and support more recent frontier models: https://
github.com/SakanaAI/AI-Sc
ientist
… Time to highlight awesome topics that people have explored with our system! Thread
CODE
-
AI Scientist Community Explores New Frontier Models
By
–
-
Computer Vision for Legacy System Automation: Specialization and Risk Management
By
–
When envisioned in the context of routine tasks using legacy systems, which nobody wants to use or maintain, I can imagine a computer vision-based approach working out well. It'd need solid specialization and risk management though!
-
GGUF Equivalent Needed for Diffusers Models
By
–
I kinda wish there was an equivalent to GGUF for Diffusers models DDUF?
-
Outlines Algorithms Ported to Rust for Faster Structured Generation
By
–
I've said it before: Structured generation is the future The @dottxtai team has ported its core Outlines algos to rust, which means from now on, structured LLM generation is: – faster
– more robust
– embeddable into any ML software (no need for Python anymore) ^#3 is my -
Install and Configure AutoTrain Advanced for Llama3 8B ORPO Fine-tuning
By
–
$ pip install autotrain-advanced
$ autotrain –config https://
raw.githubusercontent.com/huggingface/au
totrain-advanced/refs/heads/main/configs/llm_finetuning/llama3-8b-orpo.yml
… -
Keras Celebrates 10 Years with 500K Daily Downloads
By
–
Keras will turn 10 this March. As of late 2024, it's getting downloaded half a million times every day on average. I recently drafted a doc that had a section titled "Keras 2035".
-
Keras vision: Long-term thinking for future generations of deep learning
By
–
I remember doing a little speech to a number of TF team members at my desk back in 2019 — about why they should think long-term, about how I wanted my future children (not even in the plans at the time) to eventually learn deep learning with Keras, 20 or so years in the future.
-
Claude 3.5 Sonnet Shows Robustness in Complex Tasks
By
–
One huge change I'm seeing in the new Claude 3.5 Sonnet model…it's more robust. I'm chucking complex decisions and code at it, and for each prompt, it's not only giving me "standard" output, it's giving flags, considerations, next steps, and alt decisions unprompted. LOVE.
-
DuoAttention: Efficient Long-Context LLM Inference with Retrieval
By
–
DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Heads https://
arxiv.org/abs/2410.10819 https://
github.com/mit-han-lab/du
o-attention
… #MIT @songhan_mit