moluzhui

Results 12 comments of moluzhui

Thanks for your answer. @HJianBo . If I want to `emqx/emqx-edge:v3.2.4` set the username and password at startup, what should I do? and in which file should I view the...

So, if I set `enable_acl_cache` to off, does the modify table implementation take effect in time, or do I need to change some other configuration?

@HJianBo sorry, I didn't catch your second one See emqx.conf results below ``` cat etc/emqx.conf | grep acl_deny_action acl_deny_action = ignore zone.external.acl_deny_action = ignore zone.internal.acl_deny_action = ignore ``` What I...

Hi, has this function been merged? I'd like to use this feature. Thank you

> ``` > Feb 28 13:55:16 iZ0jl5att67k7fqmbzp4j2Z ollama[189492]: llama.cpp:11942: The current context does not support K-shift > ``` > > [#5975](https://github.com/ollama/ollama/issues/5975) Does num_ctx increase mean that a large number of...

> maybe it is crashed. Indeed, I asked again to wait forever, and the GPU was not in use, so I had to restart the ollama service. That's too bad

I see num_ctx and other settings in [#5975](https://github.com/ollama/ollama/issues/5975#issuecomment-2306851184) are shown below, may I ask whether these two parameters have the best recommendation for different GPU memory? H200 and 760GiB video...

> Set [`OLLAMA_LOAD_TIMEOUT=30m`](https://github.com/ollama/ollama/blob/98d44fa39d22c9c6f86fb964dd3bb13a38356371/envconfig/config.go#L248) in the server environment. Is this an occasional phenomenon? There is no such problem on another server of the same type

> ``` > Feb 26 19:46:29 iZbp1d0j83sijy1mh3b3s9Z ollama[21816]: time=2025-02-26T19:46:29.575+08:00 level=INFO source=gpu.go:612 msg="Unable to load cudart library /usr/lib64/libcuda.so.418.226.00: symbol lookup for cuCtxCreate_v3 failed: /usr/lib64/libcuda.so.418.226.00: undefined symbol: cuCtxCreate_v3" > ``` > >...

The problem remains unresolved. I upgraded CUDA to version **11.4** and before that upgraded the driver to a compatible version **470.256.02**. However, after `ollama run deepseek-r1:1.5b "same question"` is executed,...