RWKV-LM
RWKV-CUDA
RWKV-LM | RWKV-CUDA | |
---|---|---|
84 | 3 | |
13,086 | 217 | |
1.8% | 0.9% | |
9.1 | 2.9 | |
9 days ago | about 2 months ago | |
Python | Cuda | |
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.
RWKV-LM
-
Do LLMs need a context window?
https://github.com/BlinkDL/RWKV-LM#rwkv-discord-httpsdiscord... lists a number of implementations of various versions of RWKV.
https://github.com/BlinkDL/RWKV-LM#rwkv-parallelizable-rnn-w... :
> RWKV: Parallelizable RNN with Transformer-level LLM Performance (pronounced as "RwaKuv", from 4 major params: R W K V)
> RWKV is an RNN with Transformer-level LLM performance, which can also be directly trained like a GPT transformer (parallelizable). And it's 100% attention-free. You only need the hidden state at position t to compute the state at position t+1. You can use the "GPT" mode to quickly compute the hidden state for the "RNN" mode.
> So it's combining the best of RNN and transformer - great performance, fast inference, saves VRAM, fast training, "infinite" ctx_len, and free sentence embedding (using the final hidden state).
> "Our latest version is RWKV-6,*
- People who've used RWKV, whats your wishlist for it?
- Paving the way to efficient architectures: StripedHyena-7B
-
Understanding Deep Learning
That is not true. There are RNNs with transformer/LLM-like performance. See https://github.com/BlinkDL/RWKV-LM.
-
Q-Transformer: Scalable Reinforcement Learning via Autoregressive Q-Functions
This is what RWKV (https://github.com/BlinkDL/RWKV-LM) was made for, and what it will be good at.
Wow. Pretty darn cool! <3 :'))))
-
Personal GPT: A tiny AI Chatbot that runs fully offline on your iPhone
Thanks for the support! Two weeks ago, I'd have said longer contexts on small on-device LLMs are at least a year away, but developments from last week seem to indicate that it's well within reach. Once the low hanging product features are done, I think it's a worthy problem to spend a couple of weeks or perhaps even months on. Speaking of context lengths, recurrent models like RWKV technically have infinite context lengths, but in practice the context slowly fades away after a few thousands of tokens.
-
"If you see a startup claiming to possess top-secret results leading to human level AI, they're lying or delusional. Don't believe them!" - Yann LeCun, on the conspiracy theories of "X company has reached AGI in secret"
This is the reason there are only a few AI labs, and they show little of the theoretical and scientific understanding you believe is required. Go check their code, there's nothing there. Even the transformer with it's heads and other architectural elements turns out to not do anything and it is less efficient than RNNs. (see https://github.com/BlinkDL/RWKV-LM)
-
The Secret Sauce behind 100K context window in LLMs: all tricks in one place
I've been pondering the same thing, as simply extending the context window in a straightforward manner would lead to a significant increase in computational resources. I've had the opportunity to experiment with Anthropics' 100k model, and it's evident that they're employing some clever techniques to make it work, albeit with some imperfections. One interesting observation is that their prompt guide recommends placing instructions after the reference text when inputting lengthy text bodies. I noticed that the model often disregarded the instructions if placed beforehand. It's clear that the model doesn't allocate the same level of "attention" to all parts of the input across the entire context window.
Moreover, the inability to cache transformers makes the use of large context windows quite costly, as all previous messages must be sent with each call. In this context, the RWKV-LM project on GitHub (https://github.com/BlinkDL/RWKV-LM) might offer a solution. They claim to achieve performance comparable to transformers using an RNN, which could potentially handle a 100-page document and cache it, thereby eliminating the need to process the entire document with each subsequent query. However, I suspect RWKV might fall short in handling complex tasks that require maintaining multiple variables in memory, such as mathematical computations, but it should suffice for many scenarios.
On a related note, I believe Anthropics' Claude is somewhat underappreciated. In some instances, it outperforms GPT4, and I'd rank it somewhere between GPT4 and Bard overall.
-
Meta's plan to offer free commercial AI models puts pressure on Google, OpenAI
> The only reason open-source LLMs have a heartbeat is they’re standing on Meta’s weights.
Not necessarily.
RWKV, for example, is a different architecture that wasn't based on Facebook's weights whatsoever. I don't know where BlinkDL (the author) got the training data, but they seem to have done everything mostly independently otherwise.
https://github.com/BlinkDL/RWKV-LM
disclaimer: I've been doing a lot of work lately on an implementation of CPU inference for this model, so I'm obviously somewhat biased since this is the model I have the most experience in.
-
Eliezer Yudkowsky - open letter on AI
I think the main concern is that, due to the resources put into LLM research for finding new ways to refine and improve them, that work can then be used by projects that do go the extra mile and create things that are more than just LLMs. For example, RWKV is similar to an LLM but will actually change its own model after every processed token, thus letting it remember things longer-term without the use of 'context tokens'.
RWKV-CUDA
- People who've used RWKV, whats your wishlist for it?
-
Accelerate PyTorch with Taichi: Data Preprocessing & High-performance ML Operator Customization
This repo introduces an interesting example of customizing an ML operator in CUDA. The author developed an RWKV language model using sort of a one-dimensional depthwise convolution custom operator. The model in itself does not involve large amounts of computation, but still runs slow because PyTorch does not have native support for it. So, the author customized the operator in CUDA and used a set of optimization techniques, such as loop fusion and Shared Memory, achieving a performance 20x better than he did with PyTorch.
-
[R] RWKV-v2-RNN : A parallelizable RNN with transformer-level LM performance, and without using attention
It's using my custom CUDA kernel ( https://github.com/BlinkDL/RWKV-CUDA ) to speedup training, so only GPU for now. On the other hand, you don't need CUDA for inference, and it is very fast even on CPUs.
What are some alternatives?
koboldcpp - Run GGUF models easily with a KoboldAI UI. One File. Zero Install.
token-shift-gpt - Implementation of Token Shift GPT - An autoregressive model that solely relies on shifting the sequence space for mixing
flash-attention - Fast and memory-efficient exact attention
web-rwkv - Implementation of the RWKV language model in pure WebGPU/Rust.
text-generation-webui - A Gradio web UI for Large Language Models with support for multiple inference backends.
RWKV-v2-RNN-Pile - RWKV-v2-RNN trained on the Pile. See https://github.com/BlinkDL/RWKV-LM for details.
nanoGPT - The simplest, fastest repository for training/finetuning medium-sized GPTs.
SmallInitEmb - LayerNorm(SmallInit(Embedding)) in a Transformer to improve convergence
gpt4all - GPT4All: Run Local LLMs on Any Device. Open-source and available for commercial use.
AI-Writer - AI 写小说,生成玄幻和言情网文等等。中文预训练生成模型。采用我的 RWKV 模型,类似 GPT-2 。AI写作。RWKV for Chinese novel generation.
llama - Inference code for Llama models
RWKV-infctx-trainer - RWKV infctx trainer, for training arbitary context sizes, to 10k and beyond!