-
### System Info
I am experimenting with TRT LLM and `flan-t5` models. My simple goal is to build engines with different configurations and tensor parallelism, then review performance. Have a DGX syst…
-
1 minion host running 1 real-time to offline task only allows 1 subtask to run at a time. Since there are no parallelism, minion cannot scale if the realtime table becomes too huge.
Note: Mapper p…
-
Running this code:
```
parallel_task = dsl.ParallelFor(repo_tasks.output, parallelism=50)
with parallel_task as item:
# this is a func_to_container_op with boto3
unpack_task =…
-
### First check
- [X] I added a descriptive title to this issue.
- [X] I used the GitHub search to find a similar request and didn't find it.
- [X] I searched the Prefect documentation for this featu…
-
**Description**
We are playing with the rolling upgrade version of the services.
In this case, we have a service deployed as global in three nodes, and we are testing the rollback feature when t…
-
## Requested Functionality
When running a large parallel job on a k8s cluster, I might have many more tasks that need to be run than I have pods to run them. In that case, it would be very useful to …
-
### Issue Type
Others
### Source
binary
### Secretflow Version
1.6.1b0
### OS Platform and Distribution
WSL Ubuntu 22.04
### Python version
3.10.13
### Bazel version
_No response_
### GCC/…
-
Julia comes with OpenBLAS and that's our default libblastrampoline backend. https://github.com/OpenMathLib/OpenBLAS/pull/4577 (which was a continuation of https://github.com/OpenMathLib/OpenBLAS/pull…
-
It takes a lot of time to run it now. Thanks!
-
### What steps did you take
from kfp.v2 import dsl
from kfp import dsl as v1dsl
Using v2 compatible pipeline, using a combination of
with v1dsl.SubGraph(parallelism=1):
with dsl.Par…