Re: [agi] Me training GPT-2 from scratch on various small datasets to compare it to my AI

I think you need a much larger training set. Are you really writing this in python? On Sun, Jul 11, 2021, 7:43 PM wrote: > https://pypi.org/project/aitextgen/ > To do it yourself: Go to the link on that page above called: Train a > GPT-2 model + tokenizer from scratch (GPU) >

[agi] Re: Me training GPT-2 from scratch on various small datasets to compare it to my AI

p.s..my post seems to have been widened and hidden, scroll sideways when needwords are covered on long lines -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Tb76967774bc4d8c2-M07fbd9450d696f76260388a6

[agi] Me training GPT-2 from scratch on various small datasets to compare it to my AI

https://pypi.org/project/aitextgen/ To do it yourself: Go to the link on that page above called: Train a GPT-2 model + tokenizer from scratch (GPU) 100KBs of enwik8.txt, here is some tests. Training took

Re: [agi] How does a machine "understand"? What is your definition of "understanding" for an AGI?

(Well, RL/ reflexes are in a sense learnt patterns really. And yeah, I consider RL sensory, motor is just motors, totally controlled by senses, no RL for them to work with.) -- Artificial General Intelligence List: AGI Permalink:

Re: [agi] How does a machine "understand"? What is your definition of "understanding" for an AGI?

Predicting truth is one thing, like GPT-2. Predicting hope is another thing, RL. -- Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Tf91e2eafa2515120-M2e85d130176ba31d16665c73 Delivery options:

Re: [agi] How does a machine "understand"? What is your definition of "understanding" for an AGI?

Text/image/sound prediction is guided by reward, so if you make it favor the word 'cheese', it will tend to predict cheese a lot more. It's just a simple adjustment permanently to the weights. It makes you say something for no reason, just instinct - to solve global warming, using fries, sleep,

Re: [agi] How does a machine "understand"? What is your definition of "understanding" for an AGI?

On 7/9/21, Daniel Jue wrote: > > In a 1977 paper by Louis Arnaud Reid in the Proceedings of the Aristotelian > Society #77 "Thinking, Feeling, Knowing" doi 10.1093/aristotelian/77.1.165 > , the author gives a (IMO) great argument about how "feeling" (not > necessarily those marked by noticeable