JAX

AI Community

Skill notes

Contents

Concepts & Theory Immutability The 4 Transformations Pytrees Code Examples

1. Developing a new Model

Define your parameters as a Pytree (dict/dataclass). Define your forward pass function (pure). Define your loss function. Use jax.value_and_grad to get gradients. Use jax.jit to speed up the update step.

2. Debugging Shapes/NaNs

Disable JIT: jax.config.update("jax_disable_jit", True) to debug with standard python tools. Use jax.debug.print inside JITted functions.

Fluxly profile Runtime and permissions are inferred by Fluxly static checks, not author declarations or runtime test results.
Domain
AI
Author-declared agents
No explicit declaration found; this is not inferred or tested compatibility
Static check
88 / 100 · heuristic scan, not runtime safety proof
Runtime evidence
Not run; no isolated runtime test evidence is currently recorded
Author / version / license
@diegosouzapw · no license declared
Token usage
Lean
Setup complexity
Plug-and-play
External API key
Not required
Operating systems
Unspecified (assume cross-platform)
Runtime requirements
Python
Detected file/system behavior
  • Read-only
  • Write / modify
Network behavior
Local-only
Install commands
None (reference only)

Heads up: Tool permissions are not declared; actual access depends on the runtime.

Author source file SKILL.md
# JAX Skill

JAX is Autograd and XLA, brought together for high-performance machine learning research.

## Contents

- [Concepts & Theory](reference.md)
  - Immutability
  - The 4 Transformations
  - Pytrees
- [Code Examples](examples.md)
  - `jit`, `grad`, `vmap`, `random` usage
  - Control Flow (`scan`, `cond`, `fori_loop`)
  - Parallelism (`sharding`)

## Common Workflows

### 1. Developing a new Model
1.  Define your parameters as a Pytree (dict/dataclass).
2.  Define your forward pass function (pure).
3.  Define your loss function.
4.  Use `jax.value_and_grad` to get gradients.
5.  Use `jax.jit` to speed up the update step.
6.  See [examples.md](examples.md) for snippets.

### 2. Debugging Shapes/NaNs
1.  Disable JIT: `jax.config.update("jax_disable_jit", True)` to debug with standard python tools.
2.  Use `jax.debug.print` inside JITted functions.

Discussion

Powered by GitHub Discussions. Sign in with GitHub to comment, react, or subscribe.