Permanently Deleted

CogVLM: Visual Expert for Pretrained Language Models

Presents CogVLM, a powerful open-source visual language foundation model that achieves SotA perf on 10 classic cross-modal benchmarks

repo: https://github.com/THUDM/CogVLM abs: https://arxiv.org/abs/2311.03079

A self-hosted BitTorrent indexer, DHT crawler, content classifier and torrent search engine with web UI, GraphQL API and Servarr stack integration.

A terminal workspace with batteries included

article: https://x.ai

trained a prototype LLM (Grok-0) with 33 billion parameters. This early model approaches LLaMA 2 (70B) capabilities on standard LM benchmarks but uses only half of its training resources. In the last two months, we have made significant improvements in reasoning and coding capabilities leading up to Grok-1, a state-of-the-art language model that is significantly more powerful, achieving 63.2% on the HumanEval coding task and 73% on MMLU.

Permanently Deleted

Permanently Deleted

The instance where I was using it changed it's rules to prevent bots from posting in it and I didn't care enough to search for another instance.

https://lemm.ee/c/issue_tracker?dataType=Post&page=1&sort=Active

`config_template.py py LEMMY_INSTANCE_URL = "" LEMMY_COMMUNITY_NAME = "" LEMMY_USERNAME = "" LEMMY_PASSWORD = "" GITHUB_API_BASE = "https://api.github.com" GITHUB_URL_BASE = "https://github.com" REPOSITORIES = ["LemmyNet/lemmy", "LemmyNet/lemmy-ui"] DB_FILE = "lemmy_github.db" DELAY = 1 MAX_BACKOFF_TIME = 300 PERSONAL_ACCESS_TOKEN = ""

github_lemmy_issue_reposter.py ```py import backoff import datetime import logging import requests import schedule import sqlite3 import time

from config import * from pythorhead import Lemmy from typing import Any, Dict, Generator, List, Optional, Tuple, Callable, TypeVar

T = TypeVar('T')

"[%(levelname)s]:%(asctime)s:%(name)s [%(filename)s:%(lineno)s - %(funcName)s()] %(message)s"

FORMAT = "%(message)s" logging.basicConfig( level=logging.INFO, format=FORMAT, handlers=[logging.FileHandler("debug.log", mode="w"), logging.StreamHandler()], )

def on_giveup(details: Dict[str, int]) -> None: logging.error(f"Failed to fetch issues after {details['tries']} attempts", exc_info=True)

def handle_errors(message: Optional[str] = None) -> Callable[[Callable[..., T]], Callable[..., T]]: def decorator(function: Callable[..., T]) -> Callable[..., T]: def wrapper(*args: Tuple[Any], **kwargs: Dict[str, Any]) -> T: try: return function(*args, **kwargs) except Exception as e: if message: logging.exception(f"{message} - Error in {function.name}:\n{e}") else: logging.exception(f"Error in {function.name}:\n{e}") raise

return wrapper

return decorator

class GitHubIssue: def init(self, issue_dict: dict[str, Any], github_repo: str) -> None: try: self.url = issue_dict["html_url"] logging.info(f"Creating issue {self.url}") self.state = issue_dict["state"] self.state_fmt = "[Closed]" if issue_dict["state"] == "closed" else "" self.repo_abbr = "[UI]" if "lemmy-ui" in github_repo else "[BE]" self.title = f"{self.state_fmt}{self.repo_abbr} {issue_dict['title']} #{issue_dict['number']}" self.title = self.title[:200] self.body = issue_dict["body"] if self.body is not None: self.body = self.body[:30000] self.user = issue_dict["user"]["login"] self.user_url = issue_dict["user"]["html_url"] self.updated_at = datetime.datetime.strptime(issue_dict["updated_at"], '%Y-%m-%dT%H:%M:%SZ') except Exception as e: log_message: str = ( f"Formatted issue:\n" f" - Repo: {github_repo}\n" f" - Issue State: {self.state}\n" f" - Repo Abbreviation: {self.repo_abbr}\n" f" - Title: {self.title}\n" f" - URL: {self.url}\n" f" - User: {self.user}\n" f" - User URL: {self.user_url}\n" f" - Updated At: {self.updated_at}\n" ) logging.exception(log_message) logging.exception(e)

@property def formatted_body(self) -> str: formatted_body: str = self.body try: if self.body is not None: formatted_body = self.body.replace("\n", "\n> ") formatted_body = f"> {formatted_body}\n> \n> Originally posted by {self.user} in #{self.number}" except Exception as e: logging.exception(f"Error formatting body for {self.url}\n{e}") return formatted_body

@property def number(self) -> int: return int(self.url.split("/")[-1])

class GitHubComment: def init(self, comment_dict: dict[str, Any], issue_number: int) -> None: self.id = comment_dict["id"] self.body = comment_dict["body"] self.user = comment_dict["user"]["login"] self.user_url = comment_dict["user"]["html_url"] self.url = comment_dict["html_url"] self.issue_number = issue_number

@property def formatted_comment(self) -> str: formatted_body:str = self.body.replace("\n", "\n> ") formatted_body = f"> {formatted_body}\n> \n> Originally posted by {self.user} in #{self.issue_number}" return formatted_body

@handle_errors("Error initializing database") def initialize_database() -> sqlite3.Connection: logging.info("Initializing database") conn: sqlite3.Connection = sqlite3.connect(DB_FILE) cursor: sqlite3.Cursor = conn.cursor() cursor.execute( """ CREATE TABLE IF NOT EXISTS posts ( issue_number INTEGER PRIMARY KEY, lemmy_post_id INTEGER NOT NULL UNIQUE, issue_title TEXT, issue_body TEXT, updated_at TIMESTAMP DEFAULT NULL ) """ ) cursor.execute( """ CREATE TABLE IF NOT EXISTS comments ( github_comment_id INTEGER PRIMARY KEY, lemmy_comment_id INTEGER NOT NULL UNIQUE, comment_user TEXT, comment_body TEXT updated_at TIMESTAMP DEFAULT NULL ) """ ) cursor.execute( """ CREATE TABLE IF NOT EXISTS last_updated ( id INTEGER PRIMARY KEY, last_updated_time TIMESTAMP ); """ ) conn.commit() return conn

def get_last_updated_time() -> str: conn = sqlite3.connect(DB_FILE) cursor = conn.cursor() cursor.execute("SELECT last_updated_time FROM last_updated WHERE id = 1") last_updated_time: str = cursor.fetchone()[0] conn.close()

return last_updated_time

def update_last_updated_time() -> None: conn = sqlite3.connect(DB_FILE) cursor = conn.cursor() current_time = datetime.datetime.utcnow().isoformat()

cursor.execute("UPDATE last_updated SET last_updated_time = ? WHERE id = 1", (current_time,)) if cursor.rowcount == 0: cursor.execute("INSERT INTO last_updated (id, last_updated_time) VALUES (1, ?)", (current_time,))

conn.commit() conn.close() logging.info("Updated last updated time")

def update_post_time(post_id: int, updated_at: datetime.datetime) -> None: conn: sqlite3.Connection = sqlite3.connect(DB_FILE) cursor: sqlite3.Cursor = conn.cursor() time_formatted = updated_at.strftime('%Y-%m-%d %H:%M:%S') SQL = "UPDATE posts SET updated_at = ? WHERE lemmy_post_id = ?" cursor.execute(SQL, (time_formatted, post_id)) conn.commit() conn.close()

def check_updated_at(issue_number: int) -> Optional[Tuple[int, str, str, Optional[str]]]: logging.info(f"Checking last post update for {issue_number}") conn: sqlite3.Connection = sqlite3.connect(DB_FILE) cursor: sqlite3.Cursor = conn.cursor() SQL = "SELECT lemmy_post_id, issue_title, issue_body, updated_at FROM posts WHERE issue_number = ?" cursor.execute(SQL, (issue_number,)) result: Tuple[int, str, str, Optional[str]] = cursor.fetchone() conn.close()

if result is None: logging.info(f"No post found for {issue_number}") return None else: logging.info(f"Found post for {issue_number}") return result

@handle_errors("Error initializing Lemmy instance") def initialize_lemmy_instance() -> Lemmy: logging.info("Initializing Lemmy instance") lemmy = Lemmy(LEMMY_INSTANCE_URL) logging.info(f"Initialized Lemmy instance in {LEMMY_INSTANCE_URL}") lemmy.log_in(LEMMY_USERNAME, LEMMY_PASSWORD) logging.info(f"Logged in to Lemmy instance with user {LEMMY_USERNAME}") return lemmy

@backoff.on_exception( backoff.expo, (requests.exceptions.RequestException, TypeError), max_time=MAX_BACKOFF_TIME, on_giveup=on_giveup, ) def fetch_github_data(url: str) -> List[Dict[str, Any]]: global LAST_REQUEST_TIME try: headers = { "Accept": "application/vnd.github+json", "Authorization": f"Bearer {PERSONAL_ACCESS_TOKEN}", "X-GitHub-Api-Version": "2022-11-28", } time_elapsed = time.time() - LAST_REQUEST_TIME required_delay = max(0, DELAY - time_elapsed) time.sleep(required_delay) response = requests.get(url, headers=headers) LAST_REQUEST_TIME = time.time() logging.info(f"Fetched data from {url}") res: List[Dict[str, Any]] = response.json() return res except requests.exceptions.RequestException as e: logging.exception(f"Error fetching data from {url}\n{e}") raise

def check_existing_post(issue_number: str) -> Optional[int]: conn: sqlite3.Connection = sqlite3.connect(DB_FILE) cursor: sqlite3.Cursor = conn.cursor() SQL = "SELECT lemmy_post_id FROM posts WHERE issue_number=?" cursor.execute(SQL, (issue_number,)) post_id: Optional[tuple[int]] = cursor.fetchone() if post_id: return post_id[0] return None

def insert_post_to_db(issue: GitHubIssue, lemmy_post_id: Optional[int]) -> None: try: conn: sqlite3.Connection = sqlite3.connect(DB_FILE) cursor: sqlite3.Cursor = conn.cursor() SQL = "INSERT INTO posts (issue_number, lemmy_post_id, issue_title, issue_body, updated_at) VALUES (?, ?, ?, ?, ?)" cursor.execute(SQL, (issue.number, lemmy_post_id, issue.title, issue.formatted_body, issue.updated_at)) conn.commit() logging.info(f"Inserted new Lemmy post {lemmy_post_id} into the database") except sqlite3.Error as e: logging.exception(f"Error inserting post into the database for issue {issue.title} with url {issue.url}\n{e}") raise

def insert_comment_to_database(cursor: sqlite3.Cursor, github_comment_id: int, lemmy_comment_id: int, comment: GitHubComment) -> None: try: SQL = "INSERT INTO comments (github_comment_id, lemmy_comment_id, comment_user, comment_body) VALUES (?, ?, ?, ?)" cursor.execute(SQL, (github_comment_id, lemmy_comment_id, comment.user, comment.formatted_comment,)) logging.info(f"Inserted comment {github_comment_id} into the database") except Exception as e: logging.exception(f"Error encountered while inserting comment {github_comment_id} to database\n{e}")

@backoff.on_exception( backoff.expo, (requests.exceptions.RequestException, TypeError), max_time=MAX_BACKOFF_TIME, on_giveup=on_giveup, ) def create_lemmy_post(lemmy: Any, community_id: int, issue: GitHubIssue) -> Optional[int]: lemmy_post_id: Optional[int] = None lemmy_post_id = lemmy.post.create(community_id, issue.title, url=issue.url, body=issue.body)["post_view"]["post"]["id"] lemmy_url = f"{LEMMY_INSTANCE_URL}/post/{lemmy_post_id}" logging.info(f"Posted issue {lemmy_url}")

return lemmy_post_id

@backoff.on_exception( backoff.expo, (requests.exceptions.RequestException, TypeError), max_time=MAX_BACKOFF_TIME, on_giveup=on_giveup, ) def create_lemmy_comment(lemmy: Any, post_id: Optional[int], comment: GitHubComment) -> Optional[int]: logging.info(f"Creating new Lemmy comment in {LEMMY_INSTANCE_URL}/post/{post_id}")

if not post_id: logging.warning("Post ID is empty. Skipping comment creation") return None

response = lemmy.comment.create(post_id, comment.formatted_comment) lemmy_comment_id:int = response["comment_view"]["comment"]["id"] logging.info(f"Successfully created Lemmy comment {LEMMY_INSTANCE_URL}/comment/{lemmy_comment_id}")

return lemmy_comment_id

def get_total_issues(github_repo: str) -> int: url: str = f"https://api.github.com/repos/{github_repo}" data: List[Dict[str, Any]] = fetch_github_data(url) total_issues: int = data["open_issues_count"] return total_issues

def fetch_issues(github_repo: str, last_updated_time: str) -> Generator[Dict[str, Any], None, None]: page = 1 per_page = 100 issues_url = (f"{GITHUB_API_BASE}/repos/{github_repo}/issues?state=all&since={last_updated_time}&per_page={per_page}")

while True: page_url = f"{issues_url}&page={page}" issues: List[Dict[str, Any]] = fetch_github_data(page_url)

if not issues: break

for issue_dict in issues: yield issue_dict

page += 1

@backoff.on_exception( backoff.expo, (requests.exceptions.RequestException, TypeError), max_time=MAX_BACKOFF_TIME, on_giveup=on_giveup, ) def edit_lemmy_post(lemmy: Any, lemmy_post_id: int, issue: GitHubIssue) -> None: lemmy.post.edit(lemmy_post_id, name=issue.title, url=issue.url, body=issue.body)

def process_issues(lemmy: Any, community_id: int, github_repo: str) -> None: last_updated_time = get_last_updated_time() update_last_updated_time() for issue_dict in fetch_issues(github_repo, last_updated_time): process_issue(lemmy, community_id, github_repo, issue_dict)

def process_issue(lemmy: Any, community_id: int, github_repo: str, issue_dict: dict[str, Any]) -> None: issue: GitHubIssue = GitHubIssue(issue_dict, github_repo) res: Optional[Tuple[int, str, str, Optional[str]]] = check_updated_at(issue.number)

if res is None: create_new_lemmy_post(lemmy, community_id, github_repo, issue) else: lemmy_post_id, existing_title, existing_body, updated_at = res if updated_at is None or has_enough_time_passed(updated_at, issue.updated_at): update_issue_if_needed(lemmy, lemmy_post_id, existing_title, existing_body, issue) process_comments(lemmy, lemmy_post_id, github_repo, issue) update_post_time(lemmy_post_id, issue.updated_at)

def has_enough_time_passed(old_updated_at_str: str, new_updated_at: datetime.datetime) -> bool: old_updated_at = datetime.datetime.strptime(old_updated_at_str, '%Y-%m-%d %H:%M:%S') time_difference: datetime.timedelta = new_updated_at - old_updated_at return time_difference >= datetime.timedelta(hours=2)

def update_issue_if_needed(lemmy: Any, lemmy_post_id: int, existing_title: str, existing_body: str, issue: GitHubIssue) -> None: if existing_title != issue.title or existing_body != issue.formatted_body: edit_lemmy_post(lemmy, lemmy_post_id, issue)

def create_new_lemmy_post(lemmy: Any, community_id: int, github_repo: str, issue: GitHubIssue) -> None: lemmy_post_id: Optional[int] = post_issue_to_lemmy(lemmy, community_id, issue) insert_post_to_db(issue, lemmy_post_id) process_comments(lemmy, lemmy_post_id, github_repo, issue)

def post_issue_to_lemmy(lemmy: Any, community_id: int, issue: GitHubIssue) -> Optional[int]: try: logging.info(f"Start posting issue {issue.title} to community {community_id}") lemmy_post_id: Optional[int] = create_lemmy_post(lemmy, community_id, issue) return lemmy_post_id except Exception as e: logging.exception(f"Error posting issue {issue.title} to community {community_id}\n{e}") return None

def process_comments(lemmy: Any, post_id: Optional[int], github_repo: str, issue: GitHubIssue) -> None: try: logging.info(f"Posting comments from issue #{issue.number} to Lemmy post {LEMMY_INSTANCE_URL}/post/{post_id}") comments_url: str = f"{GITHUB_API_BASE}/repos/{github_repo}/issues/{issue.number}/comments" comments: Dict[str, Any] = fetch_github_data(comments_url) for comment_data in comments: if isinstance(comment_data, str): logging.warning(f"Skipping comment {comment_data}") continue process_comment(lemmy, github_repo, comment_data, post_id, issue.number) except Exception as e: logging.exception(f"Error posting comments to lemmy post {post_id}\n{e}")

def process_comment(lemmy: Any, github_repo: str, comment_data: Dict[str, Any], post_id: Optional[int], issue_number: int) -> None: conn: sqlite3.Connection = sqlite3.connect(DB_FILE) cursor: sqlite3.Cursor = conn.cursor() comment = GitHubComment(comment_data, issue_number)

existing_comment_id: Optional[int] = get_existing_comment_id(cursor, comment.id) if existing_comment_id: logging.info(f"Skipping existing comment with GitHub comment ID: {comment.id}") return

post_comment_to_lemmy(cursor, lemmy, github_repo, comment, post_id, issue_number) conn.commit()

def post_comment_to_lemmy(cursor: sqlite3.Cursor, lemmy: Any, github_repo: str, comment: GitHubComment, post_id: Optional[int], issue_number: int) -> None: lemmy_post_url = f"{LEMMY_INSTANCE_URL}/post/{post_id}" comment_url = f"{GITHUB_URL_BASE}/{github_repo}/issues/{issue_number}#issuecomment-{comment.id}" logging.info(f"Posting comment {comment.url} to Lemmy post {lemmy_post_url}") lemmy_comment_id: Optional[int] = create_lemmy_comment(lemmy, post_id, comment)

if not lemmy_comment_id: logging.exception(f"Error creating Lemmy comment {lemmy_comment_id} to {lemmy_post_url} from Github comment {comment.url}") return

logging.info(f"Posted comment {comment_url} to Lemmy post {lemmy_post_url}") insert_comment_to_database(cursor, comment.id, lemmy_comment_id, comment)

def get_existing_comment_id(cursor: sqlite3.Cursor, github_comment_id: int) -> Optional[int]: logging.info(f"Checking if comment with GitHub comment ID: {github_comment_id} exists") cursor.execute("SELECT lemmy_comment_id FROM comments WHERE github_comment_id=?", (github_comment_id,)) existing_comment = cursor.fetchone() if existing_comment is not None: logging.info(f"Found existing comment with GitHub comment ID: {github_comment_id}") existing_comment_id: int = existing_comment[0] return existing_comment_id else: logging.info(f"No existing comment found with GitHub comment ID: {github_comment_id}") return None

def fetch_issue_data(github_repo: str) -> List[Tuple[str, Optional[int]]]: logging.info("Fetching the GitHub issue number and Lemmy post ID for all issues") conn: sqlite3.Connection = sqlite3.connect(DB_FILE) cursor: sqlite3.Cursor = conn.cursor() SQL = "SELECT issue_url, lemmy_post_id FROM posts WHERE issue_url LIKE ?" issues_url = f"https://github.com/{github_repo}/issues/%" issue_data = cursor.execute(SQL, (issues_url,)).fetchall() logging.info(f"Fetched {len(issue_data)} issues") return issue_data

def process_repo(lemmy: Any, community_id: int, github_repo: str) -> None: try: logging.info(f"Processing repository {github_repo}") process_issues(lemmy, community_id, github_repo) except Exception as e: logging.exception(f"Error occurred while processing repository {github_repo}\n{e}")

def main() -> None: logging.info("Running main function") initialize_database() lemmy = initialize_lemmy_instance() community_id = lemmy.discover_community(LEMMY_COMMUNITY_NAME)

for github_repo in REPOSITORIES: process_repo(lemmy, community_id, github_repo)

def run_periodically() -> None: logging.info("Starting periodic run") schedule.every(1).hours.do(main)

while True: try: schedule.run_pending() except Exception as e: logging.exception(f"Error occurred during scheduling\n{e}") time.sleep(60)

if name == "main": try: logging.info("Starting script") main() run_periodically() except Exception as e: logging.exception(f"Error occurred during script execution\n{e}") ```

requirements.txt pythorhead==0.12.3 schedule==1.2.0 backoff==2.2.1 feedparser==6.0.10

Hey everyone, I was wondering what you think about having a bot for nightly builds to test the latest changes and discover regressions. This would allow some instances to test the latest changes and discover regressions so that we don't get stuck with those until the next big release. I think this would be a great way to improve the user experience. What do you think?

I'm looking for more cost-effective alternatives to Perplexity.ai that offer Claude 2 or similar integration along with search capabilities for factual assistance, ideally around $10/month instead of the $20/month subscription fee for Perplexity.ai. I've found Phind.com, a developer-focused search engine that uses GPT-4 to answer technical questions with code examples and detailed explanations. While it may not be as good as Perplexity.ai, it offers more free uses. Are there any other options that combine large context and search features at a lower price point?

Fellow Lemmy users,

The Lemmy development team is considering adding a new tag system that would allow us to tag posts with keywords. This could make it easier to search for and find content on Lemmy.

Before implementing this, the team would like our feedback as users. Specifically:

• Do you think having post tags would be helpful on Lemmy? Why or why not?

• How should tags be displayed and integrated into Lemmy?

Please share your thoughts on whether you'd find a tag system useful, and if so, how you'd want it implemented. The dev team reads the feedback and will use it to decide how to proceed.

To give your input, you can comment or vote here or on the GitHub issue[^1].

Thanks for helping shape Lemmy! This is our community, so please speak up.

Looking forward to hearing your thoughts.

• RFC document

[^1]: GitHub — Post tags