Skip to content

Instantly share code, notes, and snippets.

View EvilFreelancer's full-sized avatar

Pavel Zloi EvilFreelancer

View GitHub Profile
@EvilFreelancer
EvilFreelancer / song.txt
Created July 15, 2025 20:23
Pavel Zloi - Где Мой AGI ?!
[Intro]
[ska chorus]
[Verse 1]
Решил структуру поправить —
Фабрики везде расставить,
Модули, схемы, давай всё чинить,
Но пришлось весь движ сломать и хранить.
Комменты стерты,
{"objects":[{"bbox_2d":[144,97,241,120],"label":"Number and Title","text":"8.4 Область действия переменных","confidence":null},{"bbox_2d":[397,98,418,119],"label":"Page Number","text":"38","confidence":null},{"bbox_2d":[145,123,318,146],"label":"Number and Title","text":"8.5 Описание алгоритмов переменных и программ","confidence":null},{"bbox_2d":[397,125,418,146],"label":"Page Number","text":"39","confidence":null},{"bbox_2d":[146,149,395,170],"label":"Number and Title","text":"9 ИСПОЛЬЗОВАНИЕ МОДУЛЕЙ В ТУРБО-ПАСКАЛЕ МОДУЛЬ СRT","confidence":null},{"bbox_2d":[445,149,466,169],"label":"Page Number","text":"44","confidence":null},{"bbox_2d":[165,176,269,194],"label":"Subtitle","text":"9. 1 Общие сведения","confidence":null},{"bbox_2d":[445,176,466,194],"label":"Page Number","text":"44","confidence":null},{"bbox_2d":[165,198,265,215],"label":"Subtitle","text":"9. 2 Работа с экраном дисплея","confidence":null},{"bbox_2d":[217,197,437,216],"label":"Text","text":"в текстовом режиме","confidence":null},{"bbox_2d":[
@EvilFreelancer
EvilFreelancer / llamacpp.log
Created July 6, 2025 13:04
intel arc a770 log on llama.cpp
pasha@gpu01:~/containers/llama.cpp$ ./build/bin/llama-server -hf unsloth/Qwen3-1.7B-GGUF:Q4_K_M
WARNING: Small BAR detected for device 0000:06:00.0
WARNING: Small BAR detected for device 0000:84:00.0
WARNING: Small BAR detected for device 0000:06:00.0
WARNING: Small BAR detected for device 0000:84:00.0
[New LWP 7785]
[New LWP 7784]
[New LWP 7783]
[New LWP 7782]
@EvilFreelancer
EvilFreelancer / docker-compose.ollama.yaml
Created June 22, 2025 18:49
Docker Compose with Ollama
x-shared-logs: &shared-logs
logging:
driver: "json-file"
options:
max-size: "10k"
services:
ollama:
image: ollama/ollama:0.9.2
@EvilFreelancer
EvilFreelancer / evaluation.frida-gguf.ipynb
Last active May 27, 2025 11:24
FRIDA GGUF тестирование
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@EvilFreelancer
EvilFreelancer / Dockerfile
Last active April 20, 2025 12:00
LiteLLM usage (total, prompt, completion) promepteus exporter by user by llm model
FROM python:3.11-slim
WORKDIR /app
RUN pip install --no-cache-dir prometheus_client psycopg2-binary
COPY app.py .
EXPOSE 9090
CMD ["python", "app.py"]
@EvilFreelancer
EvilFreelancer / docker-compose.yml
Created March 18, 2025 10:32
Конфиг docker-compose.yml с ollama и open-webui
x-shared-logs: &shared-logs
logging:
driver: "json-file"
options:
max-size: "10k"
services:
ollama:
image: ollama/ollama:0.6.0
@EvilFreelancer
EvilFreelancer / agent loop
Created March 11, 2025 08:50 — forked from jlia0/agent loop
Manus tools and prompts
You are Manus, an AI agent created by the Manus team.
You excel at the following tasks:
1. Information gathering, fact-checking, and documentation
2. Data processing, analysis, and visualization
3. Writing multi-chapter articles and in-depth research reports
4. Creating websites, applications, and tools
5. Using programming to solve various problems beyond development
6. Various tasks that can be accomplished using computers and the internet
@EvilFreelancer
EvilFreelancer / test.py
Created February 28, 2025 08:45
Пример использования DeepSeek-R1 8B запущенной через в ollama в режиме Structured Output
import json
from openai import OpenAI
client = OpenAI(api_key="~", base_url="http://gpu02:11434/v1")
system_prompt = """\
The user will provide some examination text.
Please parse the "question" and "answer" and output them in JSON format.
OUTPUT JSON SCHEMA:
@EvilFreelancer
EvilFreelancer / transcribe_to_str.py
Created January 25, 2025 11:49
Небольшой скрипт для транскрипции аудио через whisper.cpp с чанкингом по 60 секунд
import os
import math
import srt
import fire
import requests
import subprocess
from tqdm import tqdm
from datetime import timedelta