LocalLLaMA !localllama@sh.itjust.works

Welcome to LocalLLaMA! Here we discuss running and developing machine learn

---

Read more

Members

3,372

Posts

354

Active Today

42

Created

2 yr. ago

Moderators

---

Distilling step-by-step: Outperforming larger language models with less training data and smaller model sizes

Efficient Fine-Tuning for Llama-v2-7b on a Single GPU

Exllama V2 released! Available in Ooba! Big speed upgrades!

Trying to run a local Story telling model with KoboldCpp

GitHub - nicholasyager/llama-cpp-guidance: A guidance compatibility layer for llama-cpp-python

How usable are AMD GPUs?

Pygmalion-2 has been released

Beating GPT-4 on HumanEval with a Fine-Tuned CodeLlama-34B

Supporting the Open Source AI Community | Andreessen Horowitz

What is your favorite offline LLM for technical utility, and have you noticed anything unexpected about certain models?

WizardLM introduce the newest WizardCoder 34B based on Code Llama

Is there a good reason why AMD APUs just aren't used with massive amounts of (V)RAM just like the Mac M2 is?

Code Llama: Open Foundation Models for Code | Meta AI Research

Making LLMs lighter with AutoGPTQ and transformers

Jon Durbin: Finished up a first stab at LMoE - LoRA mixture of experts

SeamlessM4T — Massively Multilingual and Multimodal Machine Translation

Hugging Face Releases IDEFICS: An Open-Access 80B Visual Language Model Replicating DeepMind's Flamingo

PEFT 0.5.0: Release GPTQ Quantization, Low-level API · huggingface/peft

GGUF PR has officially been merged into master!

Sam Bleckley's insighful explanation of why "Don't Fire Your Illustrator"