jax.random.PRNGKey — JAX documentation
The resulting key carries the default PRNG implementation, as determined by the optional impl argument or, otherwise, by the jax_default_prng_impl config flag. seed (int | ArrayLike) – a 64- or 32-bit integer used as the value of the key. impl (PRNGSpecDesc | None | None) – optional string specifying the PRNG implementation (e.g. 'threefry2x32') A PRNG key, consumable by random functions as well as split and fold_in. KeyArray Previous jax.random module Next jax.random.key By The JAX authors © Copyright 2024, The JAX Authors. NumPy and SciPy documentation are copyright the respective authors..
jax.random.PRNGKey # jax.random. PRNGKey ( seed , * , impl = None ) [source] # Create a legacy PRNG key given an integer seed. This function produces old-style legacy PRNG keys, which are arrays of dtype uint32 . For more, see the note in the PRNG keys section. When possible, jax.random.key() is recommended for use instead. The resulting key does not carry a PRNG implementation. The returned key matches the implementation given by the optional impl argument or, otherwise, determined by the jax_default_prng_impl config flag. Callers must ensure that same implementation is set as the default whe
Explore this link on the map →related reading
- jax.random module — JAX documentationjax.readthedocs.io
- Illustrating Reinforcement Learning from Human Feedback (RLHF)huggingface.co
- Pseudorandom Number Generators | Computer Securitytextbook.cs161.org
- Myths about /dev/urandom2uo.de
- /dev/random - Wikipediaen.wikipedia.org
- jason (@jxnlco) / Xtwitter.com
- Library Functions | CS 161 Spring 2024sp24.cs161.org
- Permuted congruential generator - Wikipediaen.wikipedia.org
- torch.randn — PyTorch 2.12 documentationdocs.pytorch.org
- Pseudorandom generator - Wikipediaen.wikipedia.org
- random_graphs.pdfinst.eecs.berkeley.edu
- Types of generators - The Rust Rand Bookrust-random.github.io