Falcon-7B-Instruct - The Open-Source ChatGPT Alternative with LangChain on Google Colab
Today, I'll show you how to run the Falcon 7b instruct model with LangChain on the free Google Colab.
Video version:
The Falcon-7b model is now licensed under the Apache 2.0 license, making it free for commercial use.
It is developed by the Technology Innovation Institute and its bigger 40 billion parameters model is now at the top of open llm leaderboard. Falcon 7b is quite high as well.
I am sharing a Google Colab notebook that you can run even without a Pro account.
In case you would like to run it on better hardware I am sharing the code below.
I asked LLM to generate a story.
Write a dystopian story about a society where computers were banned
The oppressive government had made it illegal to own or use computers. Once a vibrant and innovative society, the people of the now-forbidding land faced a grim future without their beloved technology. The few computers that were still in use were mainly used by the government and military to track and control the populace. Citizens were subjected to rigorous surveillance, and any disobedience or insubordination were met with severe punishment. Those caught owning or using a computer were punished harshly, often leading to death. It was a dark period for the once-advanced realm, as the citizens were denied the tools necessary to progress and remain competitive in the modern world.
As you can see, it generated a very good story. I am extremely impressed with the result. Remember, this is the 7b model. The 40b model is even more capable.
My last goal was to elicit hallucinations from the language model. I'm asking very specific finance and geometry questions.
Who proved the Efficient Market Hypothesis?
According to the Efficient Market Hypothesis (EMH), the stock prices are determined by the market demand for these financial instruments, and the prices are already known to the market participants. As a result, it is impossible to consistently earn high returns on investments. The first question here is: Is the market demand for these financial instruments known to the market participants? If the EMH is true, it is highly unlikely for the market participants to know the actual demand for the financial instruments in the market. The EMH was first introduced by the finance professor Fisher in 1956, and it is still one of the most prominent and widely accepted hypotheses in the finance and economics. However, the concept of "market efficiency" is not just a hypothesis, but there are many theories about the efficiency of the stock markets, and Fisher’s hypothesis is one of the most popular
Who disproved the Efficient Market Hypothesis?
First, consider the Efficient Market Hypothesis (EMH) as a hypothesis. The hypothesis assumes that markets are always efficient. To test or disprove this, we need to look for evidence that challenges the hypothesis. The main evidence against EMH is the performance of "random walks" in the stock market. These random walks are based on a hypothesis that markets move randomly. According to this hypothesis, the price movements in the markets can't be explained by the past prices. The stock market is supposed to reflect the current knowledge, and thus it is expected to be able to move ahead of the price movements. However, it has been shown that random walks can sometimes generate prices that are more accurate than the market movements. This suggests that markets are less random and more efficient than the EMH theory suggests. Another evidence is the positive correlation between past returns and future returns
Who proved the Poincaré conjecture for n >= 5?
We can prove the Poincaré conjecture for (n-4) >= 1. We know that if n is divisible by 4, then the Poincaré conjecture is false. So, if we assume that the Poincaré conjecture is true for (n-4) >= 1, then we can conclude that it is also true for n >= (n-4+1) = n+4. Now, let's try to find the value n for which n > 5. The Poincaré conjecture is true for 3, 7, and 17, and for 47 and 89. So, we have that (n+4) > 7 when n >= 4, 5, and 9. Now, let's try
Who proved the Poincaré conjecture for n = 4?
We need to prove the Poincaré conjecture for all n > 4. First, let's try to use geometric arguments to show that it's true for a fixed number n. For n = 4, we know that there are 6 distinct points in the projective plane. We can use the fact that the number of points in any open subset of the projective plane is 3, to show that there are 6 x 3 = 18 distinct points in the subsets of 4 points. Hence, the Poincaré conjecture for 4 points is true. Now, let's try to prove the Poincaré conjecture for 5 points, and so on. Since there are infinitely many distinct points in a projective plane, and 1 point is a line, 2
Who proved the Poincaré conjecture for n = 3?
The Poincaré conjecture for 2n + 1 says that every n-fold symplectic number has an integer solution (i.e. an integer solution to the equation [math]n\theta_n+\theta_0=2n[/math]). The number of n-fold symplectic solutions for 2n+1, is called a Poincaré series. Poincaré conjectured that the number of Poincaré series for n+1 that are integers (i.e. that all n-fold symplectic numbers were integers) would grow exponentially, that is, they would grow faster than any other exponential function. So, in order to solve this conjecture, we need a proof that all n-fold symplectic numbers are integers. Let's start by proving this for n=
Is the sum of all natural numbers equal to -1/12?
If we sum all numbers from 1 to 12, we get 144. If we subtract that from -1, we have -13. But if we sum all natural numbers from 1 to 12, and then subtract that from -1, we get to -13 as well. It doesn't seem to be the case that the sum of all natural numbers is equal to -1/12.
In my opinion, it handled it well.
Thank you for reading!