promptbench
TornadoVM
promptbench | TornadoVM | |
---|---|---|
4 | 22 | |
2,103 | 1,123 | |
9.0% | 2.8% | |
9.2 | 9.9 | |
13 days ago | 3 days ago | |
Python | Java | |
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.
promptbench
-
Show HN: Times faster LLM evaluation with Bayesian optimization
Fair question.
Evaluate refers to the phase after training to check if the training is good.
Usually the flow goes training -> evaluation -> deployment (what you called inference). This project is aimed for evaluation. Evaluation can be slow (might even be slower than training if you're finetuning on a small domain specific subset)!
So there are [quite](https://github.com/microsoft/promptbench) [a](https://github.com/confident-ai/deepeval) [few](https://github.com/openai/evals) [frameworks](https://github.com/EleutherAI/lm-evaluation-harness) working on evaluation, however, all of them are quite slow, because LLM are slow if you don't have infinite money. [This](https://github.com/open-compass/opencompass) one tries to speed up by parallelizing on multiple computers, but none of them takes advantage of the fact that many evaluation queries might be similar and all try to evaluate on all given queries. And that's where this project might come in handy.
- FLaNK Weekly 31 December 2023
- FLaNK 25 December 2023
- Promptbench: A Unified Library for Evaluating and Understanding LLMs
TornadoVM
-
Intel Gaudi 3 AI Accelerator
You don't need to use C++ to interface with CUDA or even write it.
A while ago NVIDIA and the GraalVM team demoed grCUDA which makes it easy to share memory with CUDA kernels and invoke them from any managed language that runs on GraalVM (which includes JIT compiled Python). Because it's integrated with the compiler the invocation overhead is low:
https://developer.nvidia.com/blog/grcuda-a-polyglot-language...
And TornadoVM lets you write kernels in JVM langs that are compiled through to CUDA:
https://www.tornadovm.org
There are similar technologies for other languages/runtimes too. So I don't think that will cause NVIDIA to lose ground.
- Java VectorAPI compatiblity with TornadoVM GPU programming framework
- Java GPU pre/post processing with ONNX RT and TornadoVM
- FLaNK Stack 05 Feb 2024
- FLaNK 25 December 2023
- GPU Acceleration for Python, JavaScript, Ruby from Java with Truffle
- TornadoVM v1.0 Released
- TornadoVM 1.0
-
From CPU to GPU and FPGAs: Supercharging Java Applications with TornadoVM [video]
Presented by Juan Fumero, PhD & Research Fellow (The University of Manchester, UK) during the JVM Language Summit 2023 (Santa Clara CA).
More information on TornadoVM can be found at https://www.tornadovm.org/
Tags: #Java #JVMLS #GPU #FPGA #OpenJDK #GraalVM #AI
What are some alternatives?
awesome-gpt-prompt-engineering - A curated list of awesome resources, tools, and other shiny things for GPT prompt engineering.
Aparapi - The New Official Aparapi: a framework for executing native Java and Scala code on the GPU.
osgameclones - Open Source Clones of Popular Games
openapi4j - OpenAPI 3 parser, JSON schema and request validator.
opencompass - OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets.
GraalVMREPL - REPL (read–eval–print loop) shell built on top of JavaFX and GraalVM stack, incorporating GraalJS, GraalPython, TruffleRuby and FastR
JavaOnRaspberryPi - Sources and scripts for the book "Getting started with Java on the Raspberry Pi"
kattlo-cli - Kattlo CLI Project
Zolver - Automatic jigsaw puzzle solver
junodb - JunoDB is PayPal's home-grown secure, consistent and highly available key-value store providing low, single digit millisecond, latency at any scale.
FLiPStackWeekly - FLaNK AI Weekly covering Apache NiFi, Apache Flink, Apache Kafka, Apache Spark, Apache Iceberg, Apache Ozone, Apache Pulsar, and more...
jr - JR: streaming quality random data from the command line