Kenvix ⭐
Kenvix ⭐
Maybe. Other platforms like Linux use UTF-8 as default and they should have no encoding issues.
The latest nvidia driver has fixed this issue. Simply download the latest **Windows** driver from https://www.nvidia.cn/drivers/lookup/
Hi, This issue is randomly triggered under **high concurrent** requests and cannot be reproduced by testing a single request message individually. The requests are initiated by a web translation plugin...
Hi, Just now, I did a simple observation of this issue. I found that it does **not** seem to be an "infinite loop" or "stuck," but rather the model continuously...
Hi, Based on my recent tests, as I previously guessed, the model falls into a loop, repeatedly outputting the same sentence until reached `max_tokens` . Below are the prompt, the...
For my instance, setting tp = 4 does not solve the problem
@jklj077 It seems like that VLLM does not load configuration of Qwen2 correctly. I noticed that vllm shows that `SamplingParams(n=1, best_of=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.0, temperature=0.6, top_p=1.0, top_k=-1, min_p=0.0 .... )`...
应该和 #513 是相同问题
Hi, @XuehaiPan The test code you provided does not seem relevant to this issue. In my case, using tmux or screen to keep nvitop running, you will find that the...
持续ping发包即可 ```shell docker exec -d ${CONTAINER_NAME} busybox ping -s 768 可到达的服务器IP地址 ```