elvircrn

Overclocking the GPU

sudo /home/nvidia/jetson_clocks.sh --store
sudo /home/nvidia/jetson_clocks.sh
sudo /home/nvidia/jetson_clocks.sh --restore

elvircrn

import re
import urllib
import urllib.request
import os
import concurrent.futures
import zipfile
import shutil


def download_stuff(comic_url, comic_name):

elvircrn

# Enable mouse mode (tmux 2.1 and above)
set -g mouse on

set -g default-terminal "screen-256color" # colors!
setw -g xterm-keys on
set -s escape-time 10                     # faster command sequences
set -sg repeat-time 600                   # increase repeat timeout
set -s focus-events on

elvircrn

#include <ATen/cuda/CUDAContext.h>
#include <c10/util/Float8_e4m3fn.h>

#include "../per_token_group_quant_8bit.h"

#include <cmath>

#include <cuda_fp16.h>
#include <cuda_bf16.h>

elvircrn

apiVersion: leaderworkerset.x-k8s.io/v1
kind: LeaderWorkerSet
metadata:
  name: wide-ep-llm-d-decode
  labels:
    llm-d.ai/inferenceServing: "true"
    llm-d.ai/model: Qwen3-30B-A3B
    llm-d.ai/role: decode
spec:
  replicas: 1

elvircrn

apiVersion: leaderworkerset.x-k8s.io/v1
kind: LeaderWorkerSet
metadata:
  name: wide-ep-llm-d-prefill
  labels:
    llm-d.ai/inferenceServing: "true"
    llm-d.ai/model: Qwen3-30B-A3B
    llm-d.ai/role: prefill
spec:
  replicas: 1