Greta not respecting random seed?

I'm not getting ~~consistent~~ identical results from re-knitting [this little greta mcmc](https://github.com/cboettig/nimbios-transients/blob/a07abe7a5ca036ddfe1b9bd8a46f1e50642b53ef/code/greta.Rmd), running on GPU.

the differences are small, but non-identical, see [summary table in this diff](https://github.com/cboettig/nimbios-transients/commit/a07abe7a5ca036ddfe1b9bd8a46f1e50642b53ef)

Here's sessionInfo():

```r
sessionInfo()
R version 3.6.0 (2019-04-26)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.2 LTS

Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.7.1
LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.7.1

locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C LC_TIME=en_US.UTF-8
[4] LC_COLLATE=en_US.UTF-8 LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_US.UTF-8 LC_NAME=C LC_ADDRESS=C
[10] LC_TELEPHONE=C LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats graphics grDevices utils datasets methods base

other attached packages:
[1] greta_0.3.0.9002

loaded via a namespace (and not attached):
[1] Rcpp_1.0.1 rstudioapi_0.10 knitr_1.23 whisker_0.3-2 magrittr_1.5
[6] hms_0.4.2 progress_1.2.2 lattice_0.20-38 R6_2.4.0 rlang_0.3.4
[11] globals_0.12.4 tools_3.6.0 parallel_3.6.0 grid_3.6.0 packrat_0.5.0
[16] xfun_0.7 coda_0.19-2 htmltools_0.3.6 tfruns_1.4 yaml_2.2.0
[21] digest_0.6.19 assertthat_0.2.1 crayon_1.3.4 tensorflow_1.13.1 Matrix_1.2-17
[26] base64enc_0.1-3 codetools_0.2-16 rsconnect_0.8.13 evaluate_0.14 rmarkdown_1.13
[31] compiler_3.6.0 prettyunits_1.0.2 reticulate_1.12 jsonlite_1.6 future_1.13.0
[36] listenv_0.7.0 pkgconfig_2.0.2
```

Here's `tf_config()`:

```r
> tensorflow::tf_config()
TensorFlow v1.13.1 (/opt/venv/lib/python3.6/site-packages/tensorflow)
Python v3.6 (/opt/venv/bin/python)
> tensorflow::tf_gpu_configured()
TensorFlow built with CUDA: TRUE
GPU device name: /device:GPU:0[1] TRUE
```