-
Thank you for providing this excellent benchmark and sharing the evaluation results across various models!
When we tested the MSMARCO task using the evaluation code provided in this GitHub reposito…
-
When running a training loop using MPI, the `EvalCallback` doesn't seem to make use of the parallelisation:
for example, in this `train` function:
https://github.com/hardmaru/slimevolleygym/blo…
-
### Describe the feature request
In `frps` we can specify ports for tcp, kcp and quic, however in `frpc.toml` we only get to specify a single `transport.protocol` and `serverPort`.
It would be nic…
-
Hi,
Are you working on implementing the fetch api? Not even the complex consumer behavior, just the kafka fetch protocol.
-
Hi
Your opnsense-speedtest is perfect for what I need as we are having irregular speed fluctuations from our ISP and this will help me to provide conclusive evidence (in conjunction with a script t…
-
with 4at/uc, depending on version, will segfault randomly after doing a few temperatures, or right after setting hamiltonian.
![image](https://user-images.githubusercontent.com/37928020/216989822-1…
-
### Describe the bug
When I used copyFile to copy a lot of audio files (around 5GB), I had to wait a long time and the CPU usage was up to 99%. But when I implemented it in Rust on the backend, it wa…
-
### Describe the bug
When the response time of the downstream service of the reverse-proxy service fluctuates, the CPU fluctuation is very obvious
### To Reproduce
[Downstream.zip](https://…
-
The commit with the notability system came with notability tiers which affix titles to people based on how famous they are. The tricky part is that between the wild fluctuations inherent in the way fa…
-
If I fixed the power, them the clock and mh/s moving a little bit and sometimes lock bu LHR.
But if I fixed the core clock, the power is changing around 130~150W with stable Mh/s.
What´s better for …