GitHub – simonw/language-models-on-the-command-line: Handout for a talk I gave about LLM and CLI tools https://
bit.ly/3Wgpg8p
#AI #MachineLearning #DeepLearning #LLMs #DataScience
CODE
-
Language Models on Command Line: CLI Integration Guide
By
–
-
Turbo LoRA: Faster Fine-Tuning with Improved Inference Efficiency
By
–
Get ready for Turbo #LoRA, our new approach to parameter-efficient #finetuning. You get: 2-3x faster text generation throughput vs. base models Task-specific response quality on par with LoRA Improved inference cost efficiency
Read more: -
Fine-tuning Embedding Models for RAG Systems
By
–
Fine-tune Embedding models for Retrieval Augmented Generation (RAG) https://
bit.ly/4cuPhqm
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
FastHTML Should Have Been Option T5 Choice
By
–
They should have gone with option T5: just use FastHTML. https://
youtube.com/watch?v=aWfYxg
-Ypm4
… -
Scaling AI Systems Across Fleet with Microservices Architecture
By
–
how you gonna scale this across a fleet using microservices huh?
-
FastHTML Streamlit Pattern: Real-Time Form Results Demo
By
–
If you're looking for a Streamlit-style pattern in FastHTML, then this is the full code for a basic example with a form on the left and some results updated in real time on the right (just the form vals in this case): https://
github.com/AnswerDotAI/fa
sthtml-example/blob/main/htmx/formdemo.py
… -
Llama 3.1 70B Recipe Adapted with LoRA Optimization
By
–
I adapted this recipe to Llama 3.1 70B using @failspy
's Meta-Llama-3-70B-Instruct-abliterated-v3.5 and optimized the LoRA rank. GGUF quants are ready and there shouldn't be any issue with the tokenizer. Special thanks to him and to grimjim for the model and this technique. -
Lorablation: Two-Step LoRA Extraction for Model Abliteration
By
–
Lorablation is a simple two-step process: 1/ Extract a LoRA by comparing two models: a censored and an abliterated Llama 3.0 2/ Merge the LoRA with a censored Llama 3.1 to abliterate it It means that we directly applied a Llama 3 LoRA to a 3.1 model and it just works!
-
htmx2ft adds flexible syntax output support
By
–
Ah what the hell here's another quick release to add a last minute PR. `htmx2ft` now supports either children-first or attr-first syntax output. https://
github.com/AnswerDotAI/fa
sthtml/pull/193
… -
htmx2ft releases children-first and attr-first syntax output
By
–
Ah what the hell here's another quick release to add a last minute PR. `htmx2ft` now supports either children-first or attr-first syntax output.