Baseten Blog | Page 9

ML models

Pinning ML model revisions for compatibility and security

Pin versions of open source packages like PyPi's transformers to avoid breaking changes or security issues; similarly, pin model revisions for stability.

ML models

Deployment and inference for open source text embedding models

Text embedding models convert text into semantic vectors. Numerous open source models cater to search, recommendation, classification & LLM-augmented retrieval.

Product

New in October 2023

All-new model management, a text embedding model that matches OpenAI, and misgif, the most fun you’ll have with AI all week.

ML models

Jina AI’s jina-embeddings-v2: an open source text embedding model that matches OpenAI’s ada-002

Jina AI released jina-embeddings-v2-base-en, a text embedding model that matches OpenAI’s ada-002 model in both benchmark performance and context window length.

Product

New in September 2023

Mistral 7B LLM, GPU comparisons, model observability features, and an open source AI event series

GPU guides

NVIDIA A10 vs A100 GPUs for LLM and Stable Diffusion inference

This article compares two popular GPUs—the NVIDIA A10 and A100—for model inference and discusses the option of using multi-GPU instances for larger models.

Product

New in August 2023

Truss' latest update addresses key ML model serving issues. Discover how to speed up SDXL inference to 3s and build ChatGPT-like apps with Llama 2 & Chainlit.

Model performance

SDXL inference in under 2 seconds: the ultimate guide to Stable Diffusion optimization

SDXL 1.0 initially takes 8-10 seconds for a 1024x1024px image on A100 GPU. Learn how to reduce this to just 1.92 seconds on the same hardware.

Hacks & projects

Build your own open-source ChatGPT with Llama 2 and Chainlit

Llama 2 rivals GPT-3.5 in quality and powers ChatGPT. Chainlit helps build ChatGPT-like interfaces. This guide shows creating such interfaces with Llama 2.

1891014