Oh hai Max, I remember you from the TC comment section :)
Here's a Hacker News card:
Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
21–30 of 47 posts
Re: Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
#22This is remarkably more coherent than previous attempts at M:tG card generation with a deep RNN [1] but on the other hand, the ability text on the few cards I've generated so far seems oddly famliar. In fact, I got one that is identical to an actual card. Generated by GPT-2 [2]: Tephraderm {1}{R} (common) Sorcery Each player sacrifices a land. Actual M:tG card (copied by hand by me and keeping the same notation): Tre…
I go into it more in the README but yes, GPT-2 overfits Magic cards very quickly despite my best efforts (usually it shifts the mana cost/rarity a bit more though than your example).
Re: Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
#23Would be awesome if you added the ability to choose the card's image in the render (like source from imgur or something).
Re: Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
#24This is remarkably more coherent than previous attempts at M:tG card generation with a deep RNN [1] but on the other hand, the ability text on the few cards I've generated so far seems oddly famliar. In fact, I got one that is identical to an actual card. Generated by GPT-2 [2]: Tephraderm {1}{R} (common) Sorcery Each player sacrifices a land. Actual M:tG card (copied by hand by me and keeping the same notation): Tre…
I go into it more in the README but yes, GPT-2 overfits Magic cards very quickly despite my best efforts (usually it shifts the mana cost/rarity a bit more though than your example).
Hm. Sorry to hear that. I'm looking into the pregenerated dump you posted on github earlier and I'm seeing a lot of cards that are basically copies of existing cards.
These are from mtg-gpt-2-cloud-run/generated_card_dumps/temp_0_7/cards_0.txt and just the top few rows (also, only the ones I immediately recognised):
Shrieking Specter {6}{B} (common)
Creature ~ Specter (4/4)
Flying
Whenever Shrieking Specter attacks, defending player discards a card.
Luminous Bonds {2}{W} (common)
Enchantment ~ Aura
Enchant creature
Enchanted creature can't attack or block.
Mwonvuli Acid-Moss {3}{G} (uncommon)
Sorcery
Destroy target land. Search your library for
a forest card and put that card onto the battlefield
tapped. Then shuffle your library.
Druid'S Deliverance {1}{G} (common)
Instant
Prevent all combat damage that would be dealt this turn. Populate.
Don't ask me why I immediately recognised Shrieking Specter.
It would take a bit of work to figure out which of the generated cards are mashups of existing cards, or the characteristics of existing cards with new names, and which ones are original. Do you have a way to figure this out?
Hey, this is an interesting project regardless. At least with M:tG cards it's very easy to see when GPT-2's output is identical to its input. Imagine trying to figure that out with arbitrary text...
Edit: Oh, the README. Reading it now.
Re: Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
#25Amazing. I wonder if there will ever be a CCG comprised entirely of AI-generated cards.
Re: Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
#26Amazing. I wonder if there will ever be a CCG comprised entirely of AI-generated cards.
I don't know how they're created, but I know that Keyforge is randomized somehow.
Re: Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
#27This is remarkably more coherent than previous attempts at M:tG card generation with a deep RNN [1] but on the other hand, the ability text on the few cards I've generated so far seems oddly famliar. In fact, I got one that is identical to an actual card. Generated by GPT-2 [2]: Tephraderm {1}{R} (common) Sorcery Each player sacrifices a land. Actual M:tG card (copied by hand by me and keeping the same notation): Tre…
I go into it more in the README but yes, GPT-2 overfits Magic cards very quickly despite my best efforts (usually it shifts the mana cost/rarity a bit more though than your example).
To avoid copying the training set you can increase the temperature at sampling time. This should trade off between accuracy (maximum likelihood) and quirkiness (more stochastic background).
Re: Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
#28This is really great. Is there any other AI similar to this one? Maybe one that generates AI illustrations for the cards...?
generating images for generated cards:
https://andymakesgames.tumblr.com/post/167733819029/urzas-dr...
Re: Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
#29Earlier quoted context omitted.
The first one I generated was a 1/1 creature costing 1 black mana, that had the ability "if you control 8 or more lands during your upkeep you win the game". It's been a while since I've played Magic, but that seems a little overpowered to me :)
Not far off these days. Might not see play in most formats. Would be cool in standard or limited though.
Too weak in modern, basically like any other bomb in limited but less interesting, and too non-interactive in standard.
Re: Show HN: Make your own AI-generated Magic: The Gathering cards with GPT-2
#30Earlier quoted context omitted.
There is this project that used an RNN to generate cards' text: https://www.mtgsalvation.com/forums/magic-fundamentals/custo... I think towards the end of the thread the initiator of the thread and some of the people in the impromptu community that formed around him started discussing how to generate illustrations and even did it to some extent.
That project turned into RoboRosewater ( https://twitter.com/RoboRosewater ), which has unfortunately been defunct for several years.
As a LoadingReadyRun (and Desert Bus) fan, I was waiting to see if RoboRosewater came up.
I'm not sure it's quite fair, as RoboRosewater is funnier when it's garbled. Some of its cards have been… memorable :)