24/08/04

[shylee2021]

Achieving Faster Open-Source Llama3 Serving with SGLang Runtime (vs. TensorRT-LLM, vLLM) https://lmsys.org/blog/2024-07-25-sglang-llama3/

gemma 2 update https://huggingface.co/google/gemma-2-2b https://huggingface.co/google/shieldgemma-9b

https://huggingface.co/google/gemma-scope https://storage.googleapis.com/gemma-scope/gemma-scope-report.pdf https://www.neuronpedia.org/gemma-scope#main

AI models collapse when trained on recursively generated data https://www.nature.com/articles/s41586-024-07566-y

[junhwi]

https://pytorch.org/blog/torchchat-local-llm-inference/ https://github.com/pytorch/torchchat

https://github.blog/news-insights/product-news/introducing-github-models/

Seeking Neural Nuggets: Knowledge Transfer in Large Language Models from a Parametric Perspective

https://maszhongming.github.io/ParaKnowTransfer/

https://arxiv.org/abs/2310.11451

[seyong92]

Synthesizer Sound Matching Using Audio Spectrogram Transformers https://arxiv.org/abs/2407.16643

Wavespace: A Highly Explorable Wavetable Generator https://arxiv.org/abs/2407.19862

[asynclee]

e5 model in8 quantization https://medium.com/nixiesearch/how-to-compute-llm-embeddings-3x-faster-with-model-quantization-25523d9b4ce5