My summary of ModCon'23 Keynote about Mojo
My summary of ModCon'23 Keynote about Mojo.
# Mojo: Pythonic system programming language
- usability of Python + performance of C
- focus on AI stack
- can use existing Python modules in Mojo (uses CPython for interoperability)
- in terms of migrating your Python code it's still not 100% compatible (https://docs.modular.com/mojo/why-mojo.html#compatibility-with-python)
- just announced mojo 0.6
- added traits to express generic functions (https://www.modular.com/blog/mojo-traits-have-arrived)
- in 2024 Q1 will start opensource Mojo
- more about the language from the talk at LLVM 2023: https://www.modular.com/blog/mojo-llvm-2023
- llama2.🔥 (https://github.com/tairov/llama2.mojo) is faster than llama2.c on CPU, with much shorter code
# MAX: Modular Accelerated Xecution
- Announced MAX platform, https://www.modular.com/max
- Includes MAX Engine (model inference runtime and API library) + MAX Serving (serving library for the MAX Engine that provides full interoperability with existing serving systems, e.g. Triton) + Mojo
- With GPU support (NVIDIA only?)
- Free developer edition, paid enterprise edition
# Mojo: Pythonic system programming language
- usability of Python + performance of C
- focus on AI stack
- can use existing Python modules in Mojo (uses CPython for interoperability)
- in terms of migrating your Python code it's still not 100% compatible (https://docs.modular.com/mojo/why-mojo.html#compatibility-with-python)
- just announced mojo 0.6
- added traits to express generic functions (https://www.modular.com/blog/mojo-traits-have-arrived)
- in 2024 Q1 will start opensource Mojo
- more about the language from the talk at LLVM 2023: https://www.modular.com/blog/mojo-llvm-2023
- llama2.🔥 (https://github.com/tairov/llama2.mojo) is faster than llama2.c on CPU, with much shorter code
# MAX: Modular Accelerated Xecution
- Announced MAX platform, https://www.modular.com/max
- Includes MAX Engine (model inference runtime and API library) + MAX Serving (serving library for the MAX Engine that provides full interoperability with existing serving systems, e.g. Triton) + Mojo
- With GPU support (NVIDIA only?)
- Free developer edition, paid enterprise edition
Источник: gonzo-обзоры ML статей
2023-12-04 18:12:29