llama-recipes

llama-recipes copied to clipboard

guyuchao

Llama 3 Preferred RAG Prompting Format (xml tags vs. markdown vs. something else)

7

### 🚀 The feature, motivation and pitch Anthropic directly states that their models prefer context for longer prompts (like the usual RAG applications) to be inserted in XML tags. Some...

krumeto

Q&A RAG pipeline with MongoDB, Hugging Face and Llama3

3

# What does this PR do? **Description of Changes**: This update introduces an integrated pipeline using Retrieval-Augmented Generation (RAG) with MongoDB and Hugging Face's open-source Llama3 model for advanced question...

RichmondAlake

### System Info transformers==4.40.0, peft==0.10.0 ### Information - [X] The official example scripts - [ ] My own modified scripts ### 🐛 Describe the bug when i running this command:...

630bdd

FMBench: benchmarking Llama models on AWS

1

# What does this PR do? Adds a recipe (`FMBench`) for benchmarking Llama models (including Llama3) on AWS platforms (SageMaker, Bedrock). `FMBench` is an open-source Python package for benchmarking foundation...

aarora79

Update Getting_to_know_Llama.ipynb

2

# What does this PR do? update LLaMa family  Fixes # (issue) ## Feature/Issue validation/testing Please describe the tests that you ran to verify your changes and relevant result...

ghLcd9dG

# What does this PR do? This PR solve an old quesiton. prepare_model_for_int8_training has been deprecated for quite some time, with PEFT v0.10.0, it has been removed. Please use prepare_model_for_kbit_training...

ghLcd9dG

Dev

2

# What does this PR do? Fixes # (issue) ## Feature/Issue validation/testing Please describe the tests that you ran to verify your changes and relevant result summary. Provide instructions so...

pranjal-jaiswal

update VideoSummary, LiveData, llama-on-prem for Llama 3

2

# What does this PR do? VideoSummary: show how to ask Llama 3 to generate a summary of a long (almost 3 hours) youtube video (Yann LeCun vs Lex Fridman)...

jeffxtang

## Main Goal Building an e2e recipe for building chatbots where we need to fine-tune a model and wont be able to rely only on RAG. This is just a...

HamidShojanazeri