How to Code a AI Trading bot (so you can make $$$)
Вставка
- Опубліковано 3 чер 2024
- 🐍 Get the free Python course
go.coursesfromnick.com/python
👨💻 Sign up for the Full Stack course and use UA-cam50 to get 50% off:
www.coursesfromnick.com/bundl...
Hopefully you enjoyed this video.
💼 Find AWESOME ML Jobs: www.jobsfromnick.com
🤖 Get the Code: github.com/nicknochnack/MLTra...
Disclaimer: This has been developed for academic purposes. Nothing herein is financial advice, and NOT a recommendation to trade real money. Please use common sense and always first consult a professional before trading or investing.
Oh, and don't forget to connect with me!
LinkedIn: bit.ly/324Epgo
Facebook: bit.ly/3mB1sZD
GitHub: bit.ly/3mDJllD
Patreon: bit.ly/2OCn3UW
Join the Discussion on Discord: bit.ly/3dQiZsV
Happy coding!
Nick - Наука та технологія
If you get a get_news() error when running the model make sure you're running alpaca-tade-api equal to 3.1.1 or greater. If not, run the following install and you should be good to go.
pip install alpaca-trade-api==3.1.1
I updated the README in GitHub so you should be sweet to smash it.
I get an api error that has something to do with the ML model IDK if you can help but if you do you may really help a junior dev get ahead.Starting backtest for MLTrader...
INFO:backtest_stats:Starting backtest...
Progress |--------------------------| 3.46% [Elapsed: 0:00:04 ETA: 0:02:07]2024-02-04 00:48:11,863: root: ERROR: Strategy.create_order() got an unexpected keyword argument 'type'
2024-02-04 00:48:11,865: root: ERROR: Traceback (most recent call last):
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 822, in run
self._run_trading_session()
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 773, in _run_trading_session
self._on_trading_iteration()
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 286, in func_output
result = func_input(self, *args, **kwargs)
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 310, in func_output
result = func_input(self, *args, **kwargs)
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 86, in func_output
return func_input(self, *args, **kwargs)
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 388, in _on_trading_iteration
on_trading_iteration()
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/tools/decorators.py", line 62, in func_output
frame, result = call_function_get_frame(func_input, *args, **kwargs)
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/tools/decorators.py", line 30, in call_function_get_frame
result = func(*args, **kwargs)
File "/Users/jasonwilliams/ML Project/MLTradingBot/tradingbot.py", line 56, in on_trading_iteration
order = self.create_order(
TypeError: Strategy.create_order() got an unexpected keyword argument 'type'
Exception in thread MLTrader:
Traceback (most recent call last):
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 822, in run
self._run_trading_session()
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 773, in _run_trading_session
self._on_trading_iteration()
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 286, in func_output
result = func_input(self, *args, **kwargs)
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 310, in func_output
result = func_input(self, *args, **kwargs)
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 86, in func_output
return func_input(self, *args, **kwargs)
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 388, in _on_trading_iteration
on_trading_iteration()
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/tools/decorators.py", line 62, in func_output
frame, result = call_function_get_frame(func_input, *args, **kwargs)
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/tools/decorators.py", line 30, in call_function_get_frame
result = func(*args, **kwargs)
File "/Users/jasonwilliams/ML Project/MLTradingBot/tradingbot.py", line 56, in on_trading_iteration
order = self.create_order(
TypeError: Strategy.create_order() got an unexpected keyword argument 'type'
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/threading.py", line 1016, in _bootstrap_inner
self.run()
File "/Users/jasonwilliams/anaconda3/envs/trader/lib/python3.10/site-packages/lumibot/strategies/strategy_executor.py", line 835, in run
raise RuntimeError("Exception encountered, stopping BackTest.") from e
RuntimeError: Exception encountered, stopping BackTest.
Creating trades plot...
Creating indicators plot...
Creating tearsheet...
2024-02-04 00:48:12,776: root: ERROR: Not enough data to create a tearsheet, at least 2 days of data are required. Skipping ((Sorry for the long explanation please help))
Thanks a million! I was going to ask 🙂
How can the bot be improved, what else can we do increase the accuracy and better the bots decision making on trades??
Anybody else gets this error?: TypeError: Strategy.create_order() got an unexpected keyword argument 'type'
but how to import the strategy code in the app ?
Duuuude. Love watching. The quickness at which you move step to step is fantastic. And the breakdown as you're doing it is fantastic. Fantastic.
he is so good at it. I learnt a lot from him.
Fantastic.
What programs do I need for this?
@@mdawg998learn python programming... m.ua-cam.com/video/tKTZoB2Vjuk/v-deo.html
Has this worked for you?
Instead of using your files, I love coding with you on each video. You are amazing sir! Thank you
Awesome, do you go back to the files as well occasionally? Sometimes I find it a little hard to get context of where the app is at.
Also OG trading logs, settings and data is in the GitHub repo in the log folder in case you want it y’all.
Crisp, smart way of explaining, necessary and sufficient information. That was helpful. Thank you.
Has this worked for you?
Hey Nicolas, really brilliant workshop and explanation to make em an idea on ML trading implementation. Thx a lot. You did great !
"Part I : Build the baseline bot":"00:00:43",
"Import libraries - lumibot etc":"00:00:56",
"Part II : Position sizing and limits":"00:10:11",
"Part III : Getting some news":"00:15:48",
"Part IV : Bringing in the ML Model (AI)":"00:23:43",
"Part V : Integrating our trading signal (sentiment)":"00:28:27"
The real MVP!
@@NicholasRenotte👌🏾it's complex and it works
🙋🏽♂️But me have questions😅
It's about the "lumibot" what is it and where can i install/get it😗
Can u drop ur WhatsApp number...so u can teach mii
Great video, very clean code and easy to code it along with you. 10/10!
Has this worked for you?
totally new to coding and as a professional trader this video is awesome to the point it seems like I already know how to code . splendid
I truly appreciate the effort you've put into sharing your knowledge and expertise. Looking forward to more great content from you in the future!
Thanks so much for checking it out! 🙏
Bro is back! Thank you :) Best regards from Germany
Ayyy thank you so much man!! Took a little break to study, pumped to be back!
Und bei hast du probiert
Love the thorough detail of building this bot from beginning to end!
Cheers Lewis!! You’re crushing it man, shorts are 🔥
What program is he using so I can begin?
What programs do I need for this?
@@mdawg998 Python and somewhere to code. I used VS Code for this.
Hey, it would be great if you continue the RL journey of yours. Coding more complex simpy environments, applying real-life cases etc. I would highly appreciate it!! Thanks for your content!
Oooh yeah RL is tricky but we can give it a go using lumibot now! Will study up while I’m travelling!
SIr i am currently learning algorithmic trading through several courses , however I stopped by coincidence at this video and observed how really good you are ! BUt unfortunately as I checked your channel there doesnt seem to be much more about the same topic as I thought , you have general coding stuff. In case you made a training courses I would be delighted to join in it , or if you can make more trading related videos about algo and quant trading if possible .... your information is so precious . Thank you
Welcome back! I really love to watch your codes.
Thanks so much man, good to be back!!
AMAZING VIDEO!! I've been looking for a video like this!
Thanks a mil 🙏
What programs do I need to do this?
Legend ! As a soon to be CS grad and trading enthusiast, this is a very informative video. Gives you a different perspective to the market, beyond the human side of it. Market makers aren’t sitting at their desk entering/exiting whale size positions by the second, computers are…code is ! I started to learn pine script (still might) to utilize scripting on tradingview, but thank you for displaying an exciting way to breaking into quant trading with python . Thank you 10000 times, earned a new subscriber!
What programs do I need for this?
@@mdawg998anaconda/ vs code (visual studio)
This Video is so helpful. I've watch it around 20 times
Your patience and composure are admirable! 😎
This is really good, Nic. Thank your very much!
Thanks a million Bryan!! 🙏
This is really cool. I learned a lot, wish to see more of these kind of things :)
Glad you enjoyed it, plenty more coming!
Hey there! This is an amazing video. The best Bot video, and also one of the best Python videos in general! Question -
Would this be difficult to setup to buy crypto instead? I'm going to work on that this week lol
necessary and sufficient information. That was helpful. Thank you.
This video is an automatic subscription trigger lol thank you!
Bro is back with some insane projects !
😉 just needed a little time to cook!
I think you should have used floor() vs. round() to set the quantity. round() may round up to the next integer if the decimal > .5. That should cause an error since it will cost more than what you wanted to risk. floor() rounds down.
Yep I think you’re right! Good spot!
Yes or use // operator i beleive that would do the same thing .
Great video, What version of NVIDIA driver do you use to speed up the machine learning processing? For me, I can only use old version 11040 driver to make the script work, but it processes slow. Newer NVIDIA driver makes "Torch" out of memory after 20% Progress then stop.
love the library summary!
Yeah? Got it, will keep it in for the next vid!
Very interesting code and explanation! I have several question: Is there any check on sufficient funds in the investment account? If the funds in the account is low, how would the bot react? Is there protection in the form of a residual reserve of funds for any taxes for the positions for positions that are already active?
Nicholas BEEN putting out that consistent high quality content.
DIEGO!! Thanks a mil!!
Oriental blue tonic has become my daily reminder that self-care can be both indulgent and beneficial. It's a treat for both body and soul 🙏
when trading you should be considering the Volume of the stock trading aswel as price. If a share drops or rises on small volume you might want to wait for beter volume to give a better signal.
you are amazing sir please make more programming videos on trading👌👌
You got it!
You're a legend mate, thank you!
Cheers man, thanks so much for checking it out! 🙏
Great video! Have you considered building a version of it in which it still implements the stop loss, but picks stocks at random instead of news and sentiment? In fact, it would be interesting to combine the two into one app and then see how one performs against the other. Similar to the experts against the darts
Now I understand there are more to learn in this trading each day, I hope it works for other of different stocks not just SPY
Great video and awesome idea! I think you can even publish some papers, Thank you❤
Cheers from Greece! Thanks for this.
Ayo, thanks happening Greece!! I’m not too far from you right now, in Barcelona!
καλησπερα αδερφε! σου απαντησα γτ ειδα εισαι ελληνας.. εχεις δοκιμασει και ειναι legit ενα τετοιο bot? ειναι κομπλε η φαση?
για σου ρε φιλε. καλος ειναι. δε το εχω παιξει ακομα. τι να σου πω; καλο ειναι. @@user-kr1us1mz3i
@@user-kr1us1mz3i και εγω Ελληνας ειμαι .εχετε φτιαξει κωδικα για trading ?
Great video. I am a trader who has been licensed since 2015. I’ve only recently (since 2020) been playing with python.
I decided to have a little fun and change the ticker to stocks, funds, and other indexes. Swings from outperformance of 34% CAGR from S&P400 and Nasdaq was negative $1B. With a b. My plan is to add in screening functionality to build a small basket of stocks and then let this loose.
Thanks for doing the video.
Hey brady when u changed the ticker did the sheets titles change as well? I was trying to do QQQ or TSLA but it kept on giving me SPY as the title
@@ethan9936 it turned from what was downloaded from GitHub, then new files were created to support the functionality of what I envision. I added a screener that built basket portfolios prior to the backtest. I didn’t care if the title was still SPY since that is often the benchmark.
He's back with a legendary project
Thanks so much!!
Could you make a follow up video on how to deploy/utilize this bot?
Would try to create an interactive streamlit dashboard for this looks like a very cool POC
Love your videos man❤
Oh yeah, a little UI to manage it on top?
@@NicholasRenotte Yes please continue this series. This is very underrated and having something like streamlit and a UI will be very helpful. Also, thanks for this video!!!!
Welcome back!
All the way from South Africa!
Ayyyy thanks so much!!
Hi Nicholas.
Great Video.
Is possible you build a python code for use XTB ?
Trade is better for stocks, Crypto or Forex , or you can use for everything that you want ?
Hi, Nicholas! Thank you very much for this tutorial. Got a question. How much time does it take to understand the particular topic, so that you can explain it so well? 30 minutes of coding, knowing exactly what to import, how to use and why, with brief and elegant explanation. I'm really impressed.
Building the original code takes a few weeks depending on how big the project is. Understanding it and memorising how to put it together for a video took 3 days for this one.
Hey bro love the vid new sub....
I just wanted to know where you learnt this all I'm currently in the position where I've finished high school planning to go uni for computer science, just wanted some advice or maybe some knowledge where you came on your journey to build this skillset of coding and having social media presence etc..
Wow your code is great, I love it❣❤
Love your videos! What about Video classification models?
Hmmmm just realised I never did a deep video classification tutorial. On the list.
Hey nick nice video as always... i havent used TF in a long time so if i start i would feel like learning a new framework so planning to shift to pytorch and based on all huggingface videos i guess pytorch is better(also cant find that many TF videos regarding hugging face)??
I'm a developer myself, but I'll get my 14 year old to start with a project like this since he dreams of being BIG in active trading/investment. New sub here!
great video, it explains alot. i will like if you could do a video analysis about Reef Chain
Thank you so much Nic! This is beautiful 👏🏾🙌🏾
Yooooo! How you doing man?! Thank you so much!!
Yeah I am doing good@@NicholasRenotte
Great explainer video Nick🎉Does the GitHub files also include the stats server you show on the video to compare the Bot’s performance to SPY?
Thanks for your wonderful video . Is it maybe possible to use some forecasting model inside the strategy and use it and the sentiment to open or close an order based on them ? and if so the forecasting model should be a single step model or a multi step ?
Could defs bring forecasting into it, sometimes the simplest solutions are best but I could def try that!
Thank you for the great video:)
Anytime, cheers for checking it out!!
❤❤Welcome back we miss you and your videos so much❤❤❤
Ohh thank you so much!! I’m back, needed a little break to study up some stuff but I’m here!!
HI Nicholas, great video, thanks so much for sharing! I have followed step by step and got the point where my bot is now up and running paper trading in alpaca - only thing is it's not really doing anything yet? is this normal or have i missed something?
question: why did you use pytorch and tokenizer to use the transformer when from tranformers you could just use pipeline?
Excellent
I have an idea, the next video demonstrating real-time violence detection using YOLOv8 or another advanced model. Explore its capabilities to enhance security and prevention in various scenarios.
and thank you
Thank you for the excellent code, Nicholas. May I ask you to add the Gemini platform API for a comparison please as I'm having trouble getting that to go using your code and I need some guidance?
Thank you so much for your video and for sharing your insights on Lumibot! Would it be possible for you to create a video explaining how to use an AI agent like Mistral to achieve similar results?
Sure, I actually built it with Mixtral to begin with but it was a little slow and had to do a bunch of prompt engineering. Can share the code though.
This is so exciting. Thank you.
Noob's question : How is the tearsheet generated ? I didn't notie it in the code.
This is so much to try to understand for a noob
Haven't seen the video yet but looks like a lot of fun!!
🙏 thank you Juan!!
You are the Messi of the programmers!!! Thanks for building it and sharing it!!
Ayyy thanks so much, if only my football was close as well 😅
You're incredible. Well done
Thanks so much 🙏
Very good video bro, can i use this code in MetaTrader 5?
Good stuff! Thank you very much Nick! Do you mind sharing you laptop specs as well as any recommendation on a decent laptop to run local llms? I’m looking to buy one to run local llms for coding and don’t really know much. Thanks a bunch again man!
This video was run on my deep learning machine 2080 super and an AMD cpu with 32GB of RAM. But I originally developed it on my MacBook Pro M1 Max with 32 gb of RAM.
I've made my own trading algo as well, but i'm working on a different way.. by using an image classifier to look for specific graph patterns from a bunch of indicators while scalping at the min, plus news
Ohhh man, I was like I wonder if I can get an image captioning model to classify a candlestick pattern as another strategy! How you going with your strat?
@NicholasRenotte The strategy works on its own for forex ( 4 multi timeframe macd + ema ), but the pattern is easier to detect visually by looking at all of the indicators at the same time. I created an image classifier following your video and it works but I still have some false positives during the signal detection. I'm getting the indicator data from mql4/5 through zmq and python is plotting before sending it to the classifier. It takes about a day to have enough data to train the model. Your videos helped me a lot. Kudos on all the content you get out there.
I’m doing something similar. I found a website that you can pay and get minute to minute quotes, right now I have about ten years of data from ten different stocks. Each has a year of basically minute to minute candles, I am experimenting with reinforcement learning
How are you going with this? Does it work?
@@miguelzuniga I’m trying basically the same thing but just looking at quote data instead of images, how much data did you use to train it on and for how long? After about 8 hours of training of the same year of AAPL stock mine wasn’t working.
I think that the ML part can be extrapolated and run on all the stocks in last 5 days to find which are the stocks with positive sentiment and focus on them instead of the opposite. which is what traders do going to the news in the morning before starting to look at the markets....
Oh so find stocks with positive sentiment and invest in those?
@@NicholasRenotte could be an idea....
Nice video and explanations. I am not able to install Trensflow on my computer; I receive an error. Do you have a video for Windows users?
also...take into account spreads and commissions per broker...as well as slippage per symbol....its easy to backtest...much harder to replicate true trading conditions...lovely stuff bro..
Oh 10000% agree.
Excellent. Thank you.
Except...
For trading, as a somebody that has just started learning that, I tend to believe that it is
"Buy the rumours sell the news".
With that the sentiment should be analyzed in both. Not just news.
is it possible to write this to use multiple symbols within the same strategy? you would basically be just creating more strategies with different symbols/%cash @ risk then adding all your strategies to the traders and with the run_all function it would be operating with all the strategies you added?
Hi interesting video. A quick question for anyone who has the time. This is the first Ive heard about alpaca, For some reason I thought that it is a broker in the states.
My question is, if someone builds a bot with this api, my take is, with not much effort the bot can be used to trade through mt4, mt5, ctrader, ninja trader and others, then it can be used with any broker that offers access using any of the mentioned trading platforms.
If so that sounds more practical than learning a whole bunch of api's
Did I get that rightly cheers, and thanks for sharing you code.
So buy at X & sell at X*1.2 or X*0.95. Got it!
Idea of a trading algorith should be first drawn on table / paper .. analysed prior to codding.. see how much sense it has.
Only 1 symbol & just with that rule.. meh.
If maket goes up.. you'll win some.. if market goes down, you'll keep loosing that 0.05X..
You need to analyse volume/ prices / various indexes to know if price will fall/ rise ..
Great content thanks, much appreciated! How would you revise the code to trade simultaneously more than one etfs/stocks instead of just one like the SPY example? What am asking is how to adjust this code for a portfolio of several instruments instead of one ETF
Thats what I am tryinbg to figure out
Thanks for the effort ❤
Thanks for checking it out!! 🙏
Man wish I dedicated my life in coding back when I was young seeing how much fun you have building this and that 😂 gonna admit mad jealous
It's never late mate!
What they said ^ but also….i am having a helluva time with it 😅. Stoked you’re enjoying it!!
Start now. Python is easy-ish. JavaScript also.
Hi, great tutorial, easy to follow, very educational.... q 4 u ... it seems like the alpaca-trade-api has been deprecated and it does not install on python 3.12... do you have any suggestions as i am stuck and cant proceed to complete the educational trading bot ... any help from the community and or you would be great... thanks
as everyone gets on the ai trading bot id expect the future of what causes volatility will change. I just dont see how the system functions with ai analysis. I'm curious if ai likes the way we have structured markets globally. That may have more long term implications for investment infrastructure then scripting some trades
For Mac users, what do you recommend to run Python for projects like these? Cloud client or desktop? Which one?
Respect to you son, i have wanted to do this for a long time but i dont have the skill in coding,.
I will follow and copy you. If i become wealthy... I WILL send you a percentage...
Thank you for your time and mind.
Let me know how you go!! Looking forward to it 🙏
How are you doing right now?
bad news... I killed the system i built it on and had to format the partition lol and also sadly my mum has been ill, in hospital and all sorts so i have not had the time to re build. Now you have reminded me I will do it this evening if i can. I rally liked it before i killed it...@@Soonicx1
Great video Nicholas! I just subscribed to learn much more. Does this work for meme coins as well?
Possibly, I’d need to test. I gave it a crack on BTCUSD and it looked promising but again that was only during back testing
@@NicholasRenotte Thanks so much for this I'm struggling to get a backtest working for BTCUSD I understand the ticker for Yfinance is BTC-USD but still doesn't spend anything. I feel like I'm missing something really obvious
How would you go about the same set sup but for a synthetic index market that does not require news events
All you need is volume profile. 😊😊😊
do you have a video that explains coding for someone just starting
Do you have an alternative to Alpaca? It's currently not offered in Canada.
Please make more of these types of videos, so cool!
Trading vids or AI vids or both?!
@@NicholasRenotte BOTH Please !
Trading vids please and btw your skills are unmatched, keep up the awesome work!@@NicholasRenotte
What video of yours should I watch first as an absolute beginner? Do you have an order you recommend?
I will watch the video later, thanks a lot. Is this AI trading bot actually useful for making money or it´s just for learning purposes?.
Interesting project! I appreciate you explaining it with diagrams too! ❤ Yet again, live coding doesn’t reflect the true problem solving process which should take much longer and with a lot more struggles😅 it’s a nice experience anyway
I don’t know how I can package four weeks of crying over my keyboard into one video 😅 haha thanks a mil man, glad you enjoyed it!!
@@NicholasRenottea picture is worth a thousand words. An imprint of QWERTY on one’s forehead explains any coding task
@@kennydubroff7772 LOL mechanical keyboards are rough to sleep on
@@NicholasRenotte also to repeatedly bang your head against
@@NicholasRenotte Add a bonus 1000x timelapse clip!
You should revisit this and figure out what you could do to stop that big fall, like surely there's a decent strategy to do that. If you can generate that much it's just a matter of figuring out how you can hold onto that. My first thought is surely there's an indicator that this part in the chart/market signaled on a long time frame that it could have taken account of
Agreed I didn’t want to overly fit the strategy to handle that but I was thinking a trailing stop loss could’ve saved us there.
Thanks for the awesome video. I am having some problems with my authentication, but I have submitted a request into Alpaca to figure this out. I can not wait to run this on my paper account! Very cool. Thanks for putting this together! I would think that having the program allow an individual stock to be named at the start of running the program would be cool. Thoughts?
Definitely! It’s easy enough to spin it up for other symbols. I think I might do another video with some other symbols/asset classes soon.
Awesome tutorial!
Thanks so much @Machiuka!!
Hi, I fell in love with your video, I thought it was really good, could you create an analysis script for IQ Optio?
Would you mind outlining the reasearch you did/resources you looked at to build this for those who want to start looking more into this type of stuff?
Can tou set it to run trades multiple times. For example the bot trading with £10.00 a day. But hundreds of times. Even if 60% of the bots loose your hiping that 10% of the bots win really big. The bot has the patience to start really slowly but also the disaplin to stick it out until it makes a million
Awesome tutorial! I build/run complex Deep learning and sentiment analysis to trade in cryptos. We're currently managing about 500K a funds.
Awesome!! DL can be kind of slow in terms of latency, curious what timeframe are you trading on?
@@NicholasRenotte absolutely. We mainly trade on 6h intervals. The secret is months of optimization + an army of GPU's crunching numbers 24/7.
@@mpbarry46 on average we outperform the market with 30-42%. This could be more but we're not taking risky trades.
do you have a video explaining q,k and v in transformers and how they are derived?,