Our great sponsors
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
I have seen the OCaml versions to be faster than Python(I did try to compare it with python here - https://github.com/ArulselvanMadhavan/mini_dalle). I haven't seen the OCaml version to be faster than Rust. It feels like it's close. With few optimizations, it might close the gap with Rust. The main benefit I see over with Rust is development time. The compiler being fast and the FP benefits that come with OCaml makes it easier to build a model fairly fast. While working on it, I would occasionally try to run the Rust library to debug issues, the rust compiler is just several times slower to build and start running the project, while the OCaml compiler would build and have completed few iterations by the time rust compiler is ready to run. This wasn't surprising either. It was just easier and faster to build a working a model
Related posts
- An OCaml port of min-DALLE
- Show HN: Toolkit for LLM Fine-Tuning, Ablating and Testing
- Compressing Images with Neural Networks
- SUPIR Full Tutorial + 1 Click 12GB VRAM Windows & RunPod / Linux Installer + Batch Upscale + Comparison With Magnific
- Visualizer for neural network, deep learning and machine learning models