-
A couple of places in the `String` docs states that an operation is `O(n)`, usually when discussing a method that needs to shift elements over. However, this is not helpful to users who do not know O-…
-
Follow the bing_bert tutorial, my deepspeed_config is:
```json
{
"train_batch_size": 4096,
"train_micro_batch_size_per_gpu": 32,
"steps_per_print": 1000,
"prescale_gradients": false,
…
-
**A clear and concise description of what the problem is.**
Concisely: if you have data as a list of dicts or jsons then Pandas is quite slow. This is an extremely common problem when data is sent…
-
We don't actually *need* a lazy spine. An alternative representation for a queue, for example:
```haskell
data N10 n a = N10 !(Array n a) !()
data N11 n a = N11 !(Array n a) !() !(Array n a)
-- …
-
I want to compress the old chunk for the continuous aggregates view but when I try to enable the compression on its respective hyper table it give me following error:
ALTER TABLE _timescaledb_inter…
-
### Description
> Too bad we don't have a writer that uses tiny (like 8 bytes) block at first, but doubles size for each new block (16 bytes, 32 bytes next, etc.). Then we would naturally use log(siz…
-
A use case we have for https://github.com/timberio/vector is to rate limit certain noisy logs. The way we would like to define them is per callsite and to set a static window where only one event will…
-
For performance reasons, this should probably copy *ideas* and some (modified) code from the `fingertree` package, but not use its actual `FingerTree` type. Here's the general idea:
```haskell
dat…
-
We should implement streams in WebPPL. More generally, the functional reactive programming model would be pretty natural and useful to adopt. If our data were a stream we could do something like `data…
-
### Describe the bug
I am using ucx 1.14.1 to build my application. However, I am unable to immediately deregister memory using `UcpMemory.deregister()`. Instead, the UcpMemory is deregistered when t…