langroid
trl
langroid | trl | |
---|---|---|
15 | 13 | |
1,698 | 8,291 | |
21.4% | 6.2% | |
9.8 | 9.7 | |
1 day ago | 4 days ago | |
Python | Python | |
MIT License | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
langroid
-
OpenAI: Streaming is now available in the Assistants API
This was indeed true in the beginning, and I don’t know if this has changed. Inserting messages with Assistant role is crucial for many reasons, such as if you want to implement caching, or otherwise edit/compress a previous assistant response for cost or other reason.
At the time I implemented a work-around in Langroid[1]: since you can only insert a “user” role message, prepend the content with ASSISTANT: whenever you want it to be treated as an assistant role. This actually works as expected and I was able to do caching. I explained it in this forum:
https://community.openai.com/t/add-custom-roles-to-messages-...
[1] the Langroid code that adds a message with a given role, using this above “assistant spoofing trick”:
https://github.com/langroid/langroid/blob/main/langroid/agen...
- FLaNK Stack 29 Jan 2024
-
Ollama Python and JavaScript Libraries
Same question here. Ollama is fantastic as it makes it very easy to run models locally, But if you already have a lot of code that processes OpenAI API responses (with retry, streaming, async, caching etc), it would be nice to be able to simply switch the API client to Ollama, without having to have a whole other branch of code that handles Alama API responses. One way to do an easy switch is using the litellm library as a go-between but it’s not ideal (and I also recently found issues with their chat formatting for mistral models).
For an OpenAI compatible API my current favorite method is to spin up models using oobabooga TGW. Your OpenAI API code then works seamlessly by simply switching out the api_base to the ooba endpoint. Regarding chat formatting, even ooba’s Mistral formatting has issues[1] so I am doing my own in Langroid using HuggingFace tokenizer.apply_chat_template [2]
[1] https://github.com/oobabooga/text-generation-webui/issues/53...
[2] https://github.com/langroid/langroid/blob/main/langroid/lang...
Related question - I assume ollama auto detects and applies the right chat formatting template for a model?
-
Pushing ChatGPT's Structured Data Support to Its Limits
we (like simpleaichat from OP) leverage Pydantic to specify the desired structured output, and under the hood Langroid translates it to either the OpenAI function-calling params or (for LLMs that don’t natively support fn-calling), auto-insert appropriate instructions into tje system-prompt. We call this mechanism a ToolMessage:
https://github.com/langroid/langroid/blob/main/langroid/agen...
We take this idea much further — you can define a method in a ChatAgent to “handle” the tool and attach the tool to the agent. For stateless tools you can define a “handle” method in the tool itself and it gets patched into the ChatAgent as the handler for the tool.
-
Ask HN: How do I train a custom LLM/ChatGPT on my own documents in Dec 2023?
Many services/platforms are careless/disingenuous when they claim they “train” on your documents, where they actually mean they do RAG.
An under-appreciate benefit of RAG is the ability to have the LLM cite sources for its answers (which are in principle automatically/manually verifiable). You lose this citation ability when you finetune on your documents.
In Langroid (the Multi-Agent framework from ex-CMU/UW-Madison researchers) https://github.com/langroid/langroid
-
Build a search engine, not a vector DB
This resonates with the approach we’ve taken in Langroid (the Multi-Agent framework from ex-CMU/UW-Madison researchers): our DocChatAgent uses a combination of lexical and semantic retrieval, reranking and relevance extraction to improve precision and recall:
https://github.com/langroid/langroid/blob/main/langroid/agen...
-
HuggingChat – ChatGPT alternative with open source models
In the Langroid library (a multi-agent framework from ex-CMU/UW-Madison researchers) we have these and more. For example here’s a script that combines web search and RAG:
https://github.com/langroid/langroid/blob/main/examples/docq...
-
SuperDuperDB - how to use it to talk to your documents locally using llama 7B or Mistral 7B?
Thanks, also found Langdroid: https://github.com/langroid/langroid/blob/main/README.md
- memory in ConversationalRetrievalChain removed
- [D] github repositories for ai web search agents
trl
- FLaNK Stack 29 Jan 2024
-
OOM Error while using TRL for RLHF Fine-tuning
I am using TRL for RLHF fine-tuning the Llama-2-7B model and getting an OOM error (even with batch_size=1). If anyone used TRL for RLHF can please tell me what I am doing wrong? Code details can be found in the GitHub issue.
-
[D] Tokenizers Truncation during Fine-tuning with Large Texts
SFTtrainer from huggingface
-
New Open-source LLMs! 🤯 The Falcon has landed! 7B and 40B
For lora - PEFT seems to work. I don't have patience to wait 5 hours, but modifying this example seems to work. You don't even need to modify that much, as their model just as neo-x uses query_key_value name for self-attention.
-
[D] Using RLHF beyond preference tuning
They have examples of making GPT output more positive (code) by using a sentiment model as reward. There are other examples about reducing toxicity, summarization here: https://github.com/lvwerra/trl/tree/main/examples . Should be fairly simple to modify the sentiment example and try the calculator reward you mentioned above.
-
[R] 🤖🌟 Unlock the Power of Personal AI: Introducing ChatLLaMA, Your Custom Personal Assistant! 🚀💬
You can use this -> https://github.com/lvwerra/trl/blob/main/examples/sentiment/scripts/gpt-neox-20b_peft/merge_peft_adapter.py
-
[R] Stanford-Alpaca 7B model (an instruction tuned version of LLaMA) performs as well as text-davinci-003
Just the hh directly. From the results it seems like it might possibly be enough but I might also try instruction tuning then running the whole process from that base. I will also be running the reinforcement learning by using a Lora using this as an example https://github.com/lvwerra/trl/tree/main/examples/sentiment/scripts/gpt-neox-20b_peft
-
[R] A simple explanation of Reinforcement Learning from Human Feedback (RLHF)
This package is pretty simple to use! https://github.com/lvwerra/trl
- Transformer Reinforcement Learning
- trl: Train transformer language models with reinforcement learning
What are some alternatives?
simpleaichat - Python package for easily interfacing with chat apps, with robust features and minimal code complexity.
lm-human-preferences - Code for the paper Fine-Tuning Language Models from Human Preferences
modelfusion - The TypeScript library for building AI applications.
alpaca-lora - Instruct-tune LLaMA on consumer hardware
autogen - A programming framework for agentic AI. Discord: https://aka.ms/autogen-dc. Roadmap: https://aka.ms/autogen-roadmap
trlx - A repo for distributed training of language models with Reinforcement Learning via Human Feedback (RLHF)
vectordb - A minimal Python package for storing and retrieving text using chunking, embeddings, and vector search.
LLaMA-8bit-LoRA - Repository for Chat LLaMA - training a LoRA for the LLaMA (1 or 2) models on HuggingFace with 8-bit or 4-bit quantization. Research only.
Adala - Adala: Autonomous DAta (Labeling) Agent framework
sparsegpt-for-LLaMA - Code for the paper "SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot" with LLaMA implementation.
chidori - A reactive runtime for building durable AI agents
llama-recipes - Scripts for fine-tuning Meta Llama3 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 number of candid inference solutions such as HF TGI, VLLM for local or cloud deployment. Demo apps to showcase Meta Llama3 for WhatsApp & Messenger.