bcarambio

3 ms·

1. 1.▲
   Acceleration is all you need (now)(octo.ai)

   1 points by bcarambio 3y ago|0 comments
2. 2.▲
   by bcarambio 3y ago

   Open Source LLM conversational memory using OctoAI and Pinecone, with code examples. If you saw yesterday's announcement from OpenAI about their limited access memory feature for ChatGPT and couldn't wait to try it because it open
3. 3.▲
   Memory for Open Source Models(octo.ai)

   2 points by bcarambio 3y ago|1 comments
4. 4.▲
   LLM Time Capsule part 1(medium.com)

   1 points by bcarambio 3y ago|1 comments
5. 5.▲
   by bcarambio 3y ago

   I started a project to snapshot the 10 most popular LLMs from the last six months. As expected, lots of churn, big hype cycles for Llama2 and Mixtral. Mixtral looks like it might stick around though...
6. 6.▲
   Performance insights for PyTorch deep learning applications(github.com)

   3 points by bcarambio 4y ago|1 comments
7. 7.▲
   by bcarambio 4y ago

   octoml-profile is a python library and cloud service designed to provide the simplest experience for assessing and optimizing the performance of PyTorch models on cloud hardware. With octoml-profile, you can easily run performance/cost
8. 8.▲
   by bcarambio 4y ago

   Hi from OctoML -- our team just put out some benchmarking results using GPT-J with a few model config/hw combinations that have relatively low latency: https:/octoml.ai/blog/evaluating-gpt-j-deployment-options/
9. 9.▲
   by bcarambio 4y ago

   Hi from OctoML -- our team just put out some benchmarking results using GPT-J with a few model config/hw combinations that have relatively low latency: /octoml.ai/blog/evaluating-gpt-j-deployment-options/