-
I have problem reading messages from IBM WebSphere MQ. It seems that some field(s) in MQMD header has type which is not handled properly. Some messages are read without exception.
When I turn off inc…
-
Thank you for taking the time to submit an issue!
## Background information
### What version of Open MPI are you using? (e.g., v3.0.5, v4.0.2, git branch name and hash, etc.)
v4.1.1 release
…
-
```
I have a new Haswell Core i7-4770k processor. Here's what i7z says:
i7z DEBUG: Found Intel Processor
i7z DEBUG: Stepping 3
i7z DEBUG: Model c
i7z DEBUG: Family 6
i7z DEBUG: Processor T…
-
Sometimes, when I save my edits, a warning message saying "Error reading chunk:" appears. You might be thinking that there is something after the colon, but there is none. When this happens, it is alw…
-
@artpol84 FYI As discussed with @rhc54 in [an issue reported in the PMIx repo](https://github.com/pmix/pmix/issues/1427), it seems something in the OpenMPI runtime 'breaks' the PMIx infrastructure so …
-
### 🐛 Describe the bug
```
def config_model(self):
flags = self.FLAGS.MODEL
model = self.get_model(flags)
model.cuda(device=self.device)
if self.world_size > 1:
if flags…
-
### Background
For various reasons, SquashFS is a viable alternative to the current recommendation of unpacking a tarball (with `ch-tar2dir`) into a tmpfs. SquashFS mounts are read-only. Issue #96 …
-
I've been getting reports lately of problems launching OMPI and PRRTE jobs under Slurm. Apparently, Slurm is now attempting to modify the ORTE/PRRTE internal `srun` command line used to launch the `mp…
rhc54 updated
4 months ago
-
### 🐛 Describe the bug
This error only happens when tensordict is imported (not used at all). Basically any backward pass after using flexattention fails.
```python
import torch
from einops im…
-
**LocalAI version:**
[localai/localai:v2.20.1-cublas-cuda12](https://hub.docker.com/layers/localai/localai/v2.20.1-cublas-cuda12/images/sha256-76e5637be78c1904046ef99f708dbefaa068b90a27cb335587…