An problem that continuously shows irrelevant example (Quicksort) answer

[lililolo0927]

I am currently using ktransformers:0.2.1 with docker image and testing DeepSeek V3-Q4_kM with ktransformers.local_chat .
(the used command : python -m ktransformers.local_chat --gguf_path /models/gguf/DeepSeek-V3-Q4_K_M --model_path deepseek-ai/DeepSeek-V3)

But the first prompt goes through successfully, but if by chance I press Enter in the middle or if there is an Enter after the first prompt, the response continues with a completely unrelated Quicksort algorithm C++ implementation, which seems wrong. How can I fix this?
Currently, my execution environment is:
CPU: INTEL(R) XEON(R) SILVER 4509Y, 16 cores per socket, 125GB DRAM / GPU: NVIDIA [GeForce RTX 4090] 24GB.

It’s already quite slow, and having this irrelevant response follow up is highly inefficient.

============ Showing as the following:
Chat: Certainly! Below is a C++ implementation of the Quicksort algorithm:

#include <iostream>
#include <vector>
#include <algorithm>

using namespace std;
// Function to partition the array and return the pivot index
int partition(vector<int>& arr, int low, int high) {
    int pivot = arr[high]; // Choosing the last element as the pivot
    int i = low - 1; // Index of the smaller element

    for (int j = low; j < high; j++) {
        // If the current element is smaller than or equal to the pivot
        if (arr[j] <= pivot) {
            i++; // Increment the index of the smaller element
            swap(arr[i], arr[j]); // Swap the elements
        }

===============

Additionally, is there any way to optimize performance in the above environment? It seems that neither the GPU nor the CPU is being fully utilized, and it’s slow. If you have any suggestions, please let me know