Machine learning engineering rewards a particular kind of curiosity. Progress often requires digging into any surprising numerical discrepancy or slight wiggle in the performance graphs. Makes the job a lot more fun if you view those explorations as exciting rather than tedious.
CODE
-
AI-Generated Functional QR Codes Model Available on Hugging Face
By
–
AI-generated functional QR codes model on @huggingface
: https://
huggingface.co/DionTimmer/con
trolnet_qrcode
… -
Recording Spaces and Using LLMs for Transcript Analysis
By
–
I wish there were an easier way. If you host the space you can record it, and you can have AI turn the recording into a transcript (a variety of ways to do that) then you can send the transcript to GPT, or other LLMs, and then you can do a ton of things. Like ask it "please
-
Prototype Custom Chat UIs Multiple Models Minutes
By
–
This allows prototyping your own custom Chat UIs with multiple models options in few minutes.
-
Agent Toolkits Optimized for Newer Chat Models
By
–
A lot of agent toolkits were mostly designed to work with normal LLMs, not chat models But chat models are newer, faster, and generally better We're working on making all our agent toolkits usable with these chat models, s/o @fpingham for doing the first four!
-
QLoRA, Prefix Tuning, and LoRA: Fine-tuning Techniques Comparison
By
–
This was already a lot of coding. I am saving QLoRA for another day! @Yampeleg Regarding prefix tuning/LLaMA-Adapter vs LoRA. The performance is similar, but one advantage of the former is that it allows multimodal inputs (but that's also a post for another day :P)
-
Proprietary Data and LLM Training Strategy for Enterprises
By
–
A few topics: * Why proprietary data enables enterprises to build higher quality large language models * Should your organization fine-tune pre-trained models or train from scratch * What are the steps that need to be considered before training your first model
-
Stop Drawing Neural Networks Incorrectly
By
–
Please Stop Drawing Neural Networks Wrong https://
bit.ly/3IsqTZk #AI #MachineLearning #DeepLearning #LLMs #DataScience -
LLaMA-Adapter v2 Adds Trainable RMSNorm and Bias Parameters
By
–
Good question. Yes, because LLaMA-Adapter v2 is basically LLaMA-Adapter but it also has trainable RMSNorm parameters and trainable bias units (in FC layers)
-
MACARONS: 3D Reconstruction and Environment Exploration from RGB Images
By
–
MACARONS: Mapping And Coverage Anticipation with RGB Online Self-Supervision
— AK (@_akhaliq) 15 juin 2023
paper page: https://t.co/3aaipPlMcj
introduce a method that simultaneously learns to explore new large environments and to reconstruct them in 3D from color images only. This is closely related to the… pic.twitter.com/rj7GWrOy9fMACARONS: Mapping And Coverage Anticipation with RGB Online Self-Supervision paper page: https://
huggingface.co/papers/2303.03
315
… introduce a method that simultaneously learns to explore new large environments and to reconstruct them in 3D from color images only. This is closely related to the