David Hall
David Hall
### Description ```python import jax.numpy as jnp import jax.random as jr k = jr.PRNGKey(0) a = jr.randint(k, (10,), minval=10, maxval=20, dtype=jnp.int64) # no warning is printed! assert a.dtype == jnp.int32...
```java import dev.ludovic.netlib.LAPACK; import org.netlib.util.intW; class Main { public static void main(String[] args) { double[] arr = new double[400]; int[] piv = new int[20]; intW info = new intW(0); LAPACK.getInstance().dgetrf(20,...
Hi, a library I'm using is using pyyaml 5.4.1. Could we do this or change to >=5.4 ? Happy to do whichever you prefer.
involves using the parser to split up spans, then using the split spans to make a decision
(meaning the llama 2 tokenizer is quadratic in sequence length) len(str) vs time to tokenize: GPT-2's tokenizer is roughly linear as you would expect: ``` gpt2: 10000 0.012285999953746796 100000 0.13678820803761482...
### Description Mostly minimized? This shouldn't OOM on the 4 a6000's I'm using (and the unminimized version works fine on TPU) ```python import jax import jax.random as jrandom from jax.sharding...
**Describe the bug** The default parsing logic seems to want to call asdict on a dataclass field even if it has a registered decoding function: **To Reproduce** ```python def test_parse_helper_uses_custom_decoding_fn():...
Buffer donation would be nice. I don't see an issue for it so just opening it for tracking/asking if it's on the Apple JAX Metal Team's roadmap ``` >>> import...