-
The attachments contain details of the warning that I encountered while working on some dataset. Kindly review it. In case of bug, fix it.
WARNING:tensorflow:Entity could not be transformed and wi…
-
It'd be convenient to provide support for use of Keras or PyTorch models in model selection. There are two issues:
1. Keras/PyTorch models don't conform to the Scikit-learn API.
2. Keras models ar…
-
### Checklist
- [ ] The issue exists after disabling all extensions
- [X] The issue exists on a clean installation of webui
- [ ] The issue is caused by an extension, but I believe it is caused by a …
-
We have been trying to migrate from Elasticsearch version 1.7.6 to the latest version (8.15) in our company and discovered that the latest version has become much slower. To find the reason for this d…
-
Now that we have high level graphs in our collections, we can do some more complex optimizations. We've done a little bit of this with blockwise fusion, read_parquet and column selection, and root fu…
-
I don't *think* this is an existing issue or anywhere in the docs, but my apologies if it is already mentioned somewhere. I occasionally run jobs on 1500+ worker clusters (using dask-gateway with the …
-
I have a code writing with trillinos, and i want to call the amgcl routine from my trillinos code. Does amgcl supoor thtis?
-
Hi, what are the requirement for NVLINK to function. I have 2 machine - one is regular pcie 3090 - 2 x cards in nvlink - works good and nvlink shows activity via :
nvidia-smi nvlink -gt r
and DGX-1…
-
Currently the only time the Chapel compiler will inline code is if the function is prefixed with `inline`. Can/should we support automatic inlining on top of whatever inlining the backend does?
Pro…
-
Observations:
- Using fused optim (e.g. Adam) + `ipex.optimize(..., level='01', dtype=torch.bfloat16)` **no error**
- non fused optimizer (e.g. AdamW) + `ipex optimize(..., level='01', dtype=torch.f…