Our friends at Unbody launched today on Product Hunt! Let's support them! Unbody is an AI-native content API that automates your full pipeline.

Connect your data and start building with a single line of code.

Build with Weaviate • vector database

producthunt.com/posts/unbody

arxiv.org/abs/2404.08634…
Knut Jägersberg Chidambara .ML. Kevin Corella N. 🔬🔭 🇪🇺
Eveline Ruehlin Estela Mandela Dr. de l'immatériel
Mack 4.0 #Tech4Good Deus ex Machina BusinessIntelligence
Andrew Neff
HeidiLeeStockenstrom☮️ Fati Sule Dr. Monika Sonu I Founder Healthinnovationtoolbox
Dr. Khulood Almani | د.خلود المانع Prof. Sally Eaves Nikos Kafritsas
Alex Dimakis
Great paper!

I find this puzzling: Why does Anthropic seem to be crawling the web more extensively than Google?

Qwen1.5-110B released 🚀

BusinessIntelligence Dr. Khulood Almani | د.خلود المانع Dr. Theophano Mitsa ☦️🇬🇷🇺🇸 Tony Moroney Françoise Morvan Fati Sule Eveline Ruehlin Prof. Sally Eaves Dr. Monika Sonu I Founder Healthinnovationtoolbox nat://TheAIObserverX

qwenlm.github.io/blog/qwen1.5-1…

marktechpost.com/2024/04/25/sno…
480B Parameters!!!
Knut Jägersberg Chidambara .ML. Kevin Corella N. 🔬🔭 🇪🇺
Eveline Ruehlin Estela Mandela Dr. de l'immatériel
Mack 4.0 #Tech4Good Deus ex Machina BusinessIntelligence
Andrew Neff
HeidiLeeStockenstrom☮️ Fati Sule Dr. Monika Sonu I Founder Healthinnovationtoolbox
Dr. Khulood Almani | د.خلود المانع Prof. Sally Eaves Nikos Kafritsas

I guess you might have tried the demo (huggingface.co/spaces/Qwen/Qw…). Now the weights of Qwen1.5-110B are out! Temporarily only the base and chat models, AWQ and GGUF quantized models are about to be released very soon!

Blog: qwenlm.github.io/blog/qwen1.5-1…

Hugging Face:…

🤠 Qwen1.5-110B model weights released. Qwen2 is on the way, let's take it step by step!
👇🏻 Enjoy it!
hf.co/spaces/Qwen/Qw…

Llama 3 70B is a top tier performance LLM

BusinessIntelligence Dr. Khulood Almani | د.خلود المانع Dr. Theophano Mitsa ☦️🇬🇷🇺🇸 Tony Moroney Françoise Morvan Fati Sule Eveline Ruehlin Prof. Sally Eaves Dr. Monika Sonu I Founder Healthinnovationtoolbox nat://TheAIObserverX

'Quantizing Llama 3 8B seems more harmful compared to other models'

The 8B model is packed so full of information it's tensores can no longer be as robustly mathematically/ structurally encoded compared to the older 7Bs.

Similar thoughts were explored in the paper - 'How Good…

Fine tune LLAMA3 on million scale dataset in consumer GPU using QLora, Deepspeed

medium.com/@sumandas0/fin…

Time flies, so I've updated my performance LLM collections

huggingface.co/KnutJaegersberg

dont you just love how the industry standard for the major labs is to release a model thats very good, then over the following weeks, make it almost entirely useless and tell no one at all.

i think theres a word for that particular type of thing.

Anthropic 's claude is now…

Could replace 'serendipity' by 'impact'.

Aka 'The amount of impact that will occur in your life is directly proportional to the degree to which you do something you're passionate about combined with the total number of people to whom this is effectively communicated.'…

🥳Multi-modal Phi-3-mini is here! #LLaVA -Phi-3-mini outperforms LLaVA-v1.5-7B and matches the performance of LLaVA-Llama-3-8B in multiple benchmarks.
😊For easy applications, #GGUF weights are provided.
👉github.com/InternLM/xtuner
AK #Phi3

Meta presents Layer Skip

Enabling Early Exit Inference and Self-Speculative Decoding

We present LayerSkip, an end-to-end solution to speed-up inference of large language models (LLMs). First, during training we apply layer dropout, with low dropout rates for