-
When enabling `-promscrape.cluster` in vmagent, it's impossible to know which shard we're using through the general service.
It would be helpful to have a seperate svc created for each shard vmagent …
-
**Describe the bug**
This option is disabled by default and produces imbalanced ingesters, which is never desirable.
**Expected behavior**
The option should be enabled always and it should not be…
-
In stable diffusion group-norm produces block sharded, row-major tensors that we feed into a matmul, so they need to be tilized. Current tilize op does not support block-sharded inputs.
Can we add…
-
### Parent Issue
#16438
### Detail of Subtask
Support sharding migration:
1. use workState to migrate
2. report sharding count to hakeeper
if shardingCount == 0, means sharding migration compl…
-
Painfully discovered and debugged during Grok-1 bringup. The `mul` operator (and inline `*` operator) do not support sharded inputs. They silently produce bad results for _some_ of the cores. Oh, and …
-
### Overview of the Issue
There is a bug in the way the shard ranges are generated for 103 & 107 shards.
For example.
```
vtctldclient --server=$host:15999 GenerateShardRanges 102 | tail -n 2
…
-
**Question**
Any plan for sharded cluster deployments on community operator?
-
I was wondering if there was a straightforward way to convert from sharded to monolithic checkpoint for a subsequent conversion to hf format (not a direct conversion sharded -> hf).
I've read you ca…
-
### Issue type
Performance
### Have you reproduced the bug with TensorFlow Nightly?
No
### Source
source
### TensorFlow version
v2.15.0-11-g63f5a65c7cd 2.15.1
### Custom code
No
### OS platf…
-
- [ ] - main hs m.org ?
- [ ] - send on pain.agency to avoid ratelimit
- [ ] - appservice api for deeper hs probing?
- [ ] - redaction of soft-failed events? (redaction on not main hs for ratelimit…