llama-recipes
llama-recipes copied to clipboard
meta-llama
Scripts for fine-tuning Meta Llama with composable FSDP & PEFT methods to cover single/multi-node GPUs. Supports default & custom datasets for applications such as summarization and Q&A. Supporting a...
# What does this PR do? ## Feature/Issue validation/testing Please describe the tests that you ran to verify your changes and relevant result summary. Provide instructions so it can be...
# What does this PR do? This PR disables prefix tuning and limit llama_adapter to non-FSDP use case. Fixes # (issue) #359 ## Feature/Issue validation/testing Please describe the tests that...
# What does this PR do? Add recipes for LangChain agents. ## Feature/Issue validation/testing Notebooks tested in fresh conda env.
# What does this PR do? Without parentheses if-statement becomes a part of lambda which is incorrect. cc @lchu-ibm @KeitaW
# What does this PR do? FMBench readme updates for Llama3 on Inf2 and config.yml cleanup Fixes # (issue) ## Feature/Issue validation/testing Please describe the tests that you ran to...
# What does this PR do? Fixes #422 lm-evaluation-harness has introduced breaking changes causing `lm_eval.tasks' has no attribute 'initialize_tasks'` error in the current `recipes/evaluation/eval.py`. This PR is aiming to resolve...
### 🚀 The feature, motivation and pitch Is it possible to adapt the fine-tuning script for DPO finetuning? The current version seems to only work for next token prediction fine-tuning....
### 🚀 The feature, motivation and pitch Documentation and [examples](https://github.com/meta-llama/llama-recipes/tree/main/recipes/benchmarks/inference_throughput/on-prem) are for llama2 benchmarks. We would like to run llama3 on prem benchmarks. ### Alternatives _No response_ ### Additional context...
# What does this PR do? Introducing a new Llama3 example that is built on top of OctoAI model API provider and Weaviate vector DB to implement a helpful sales...
Metadata
Owner
Metadata
Scripts for fine-tuning Meta Llama with composable FSDP & PEFT methods to cover single/multi-node GPUs. Supports default & custom datasets for applications such as summarization and Q&A. Supporting a...