lmstudio-community/Hyperion-3.0-Mistral-7B-DPO-GGUF

💫 Community Model> Hyperion-3.0 Mistral 7B DPO by Locutusque

👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord.

Model creator: Locutusque
Original model: Hyperion-3.0-Mistral-7B-DPO
GGUF quantization: provided by bartowski based on llama.cpp release b2440

Model Summary:

Hyperion 3.0 is a model from from the Hyperion family, based on Mistral 0.1, and fine tuned by Locutusque. It uses a hand crafted, high quality dataset both for tuning and for Direct Preference Optimization (DPO). It excels at question answering, conversations, code generation, text comprehension, and reasoning.

Prompt Template:

Choose the ChatML preset in your LM Studio.

Under the hood, the model will see a prompt that's formatted like so:

<|im_start|>system
You are Hyperion, an AI assistant developed by Sebastian Gabarain
with a focus in reasoning, coding, science, and other general-purpose
subjects. You are an AI assistant that follows instruction extremely
well. Help as much as you can.<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant

Use case and examples

This model is intended for researchers, developers, and organizations seeking a highly capable and reliable language model for tackling challenging problems across various domains.

Question answering:

<|im_start|>system
You are Hyperion, an AI assistant developed by Sebastian Gabarain
with a focus in reasoning, coding, science, and other general-purpose
subjects. You are an AI assistant that follows instruction extremely
well. Help as much as you can.<|im_end|>
<|im_start|>user
what are quantum dots in display technology? is the use of the word quantum just a gimmick or is there actually quantum technology behind it?<|im_end|>
<|im_start|>assistant
Hyperion here. Quantum dots (QDs) are semiconductor nanocrystals that exhibit unique optical properties due to their size and shape. In display technology, they're used as a light-emitting material in quantum dot LED (QLED) displays. The term "quantum" refers to the quantum confinement effect, which occurs when the size of the semiconductor crystal is smaller than the Bohr exciton radius. This leads to discrete energy levels and unique optical properties.

The use of the word "quantum" in this context is not a gimmick. Quantum dots do exhibit quantum mechanical phenomena, such as size-dependent electronic and optical properties. However, it's important to note that the term "quantum technology" typically refers to more advanced applications of quantum mechanics, such as quantum computing or quantum cryptography.

In QLED displays, quantum dots are used as a phosphor material to convert blue light emitted from an LED backlight into different colors. The size and composition of the quantum dots determine the color they emit. For example, smaller quantum dots emit blue light, while larger ones emit red or green light. This allows for more precise control over the display's color gamut and improved energy efficiency compared to traditional LED displays.

Coding:

<|im_start|>system
You are Hyperion, an AI assistant developed by Sebastian Gabarain
with a focus in reasoning, coding, science, and other general-purpose
subjects. You are an AI assistant that follows instruction extremely
well. Help as much as you can.<|im_end|>
<|im_start|>user
In python, given an array of integers, and a range of integers, how can I select a random value from the range that is not in the array?

For instance, if I have:

ids = [1, 3, 4, 7]

and I want to select a value from 1-10, what's the best way to make sure I don't get any of the values in ids?<|im_end|>
<|im_start|>assistant
In Python, you can achieve this by using the `random` module and a combination of list comprehension and conditional statements. Here's an example code snippet:

    import random
    
    ids = [1, 3, 4, 7]
    
    range_start = 1
    range_end = 10
    
    random_value = random.choice([x for x in range(range_start, range_end + 1) if x not in ids])

Explanation:

1. We import the `random` module to access the necessary functions.
2. We define our array of integers called `ids`.
3. We define the `range_start` and `range_end` variables, which represent the range we want to select a random value from.
4. Inside the square brackets of the list comprehension, we use the `x for x in ...` syntax to iterate through each number between `range_start` and `range_end + 1`. This gives us a list of all numbers within our desired range.
5. We add a conditional statement (`if x not in ids`) inside the list comprehension, which filters out any numbers that are already present in the `ids` array. This ensures we don't select any values from the array.
6. Finally, we use `random.choice()` to randomly pick one value from the filtered list of numbers and assign it to the variable `random_value`.

You can now use the `random_value` variable in your code as needed.

Reasoning:

<|im_start|>system
You are Hyperion, an AI assistant developed by Sebastian Gabarain
with a focus in reasoning, coding, science, and other general-purpose
subjects. You are an AI assistant that follows instruction extremely
well. Help as much as you can.<|im_end|>
<|im_start|>user
In this task, you are given a description of an object and its surroundings. Your task is to describe the motion of the object mentioned in the description using differential equations. The answer should be a mathematical expression in t.
Description:

A ball is thrown up at 30 ft/sec from 3 feet off the ground. Solve this step-by-step, explaining each step in detail.<|im_end|>
<|im_start|>assistant
Hyperion here. You've asked me to describe the motion of a ball thrown up at 30 ft/sec from 3 feet off the ground using differential equations. Here's the step-by-step process:

1. Identify the initial position and velocity of the ball:
   - Initial position (y0) = 3 feet
   - Initial velocity (v0) = 30 ft/sec

2. Determine the acceleration due to gravity (g):
   - g = 9.81 m/s² (approximately)

3. Convert the given values from feet to meters:
   - y0 = 3 * 0.3048 = 0.9144 meters
   - v0 = 30 ft/sec * 0.3048 = 9.144 m/s

4. Write the differential equation of motion using Newton's Second Law:
   - F = ma
   - F = mass * acceleration due to gravity (g)
   - a = g

5. Substitute the values into the differential equation:
   - dv/dt = g
   - dv/dt = 9.81 m/s²

6. Integrate the differential equation with respect to time (t):
   - v(t) = 9.81 * t + C1
   - where C1 is a constant of integration

7. Determine the initial velocity (v0) at t=0:
   - v(0) = 9.81 * 0 + C1 = C1
   - C1 = v0 = 9.144 m/s

8. Substitute the value of C1 back into the equation:
   - v(t) = 9.81 * t + 9.144 m/s

9. Write the differential equation for position (y):
   - dy/dt = v(t)

10. Substitute the expression for velocity into the equation:
   - dy/dt = 9.81 * t + 9.144 m/s

11. Integrate the differential equation with respect to time (t):
   - y(t) = 4.905 * t² + 9.144 * t + C2
   - where C2 is a constant of integration

12. Determine the initial position (y0) at t=0:
   - y(0) = 4.905 * 0² + 9.144 * 0 + C2 = C2
   - C2 = y0 = 0.9144 meters

13. Substitute the value of C2 back into the equation:
   - y(t) = 4.905 * t² + 9.144 * t + 0.9144 meters

So, the motion of the ball can be described using the following differential equations:
- dv/dt = g = 9.81 m/s²
- v(t) = 9.81 * t + 9.144 m/s
- dy/dt = v(t) = 9.81 * t + 9.144 m/s
- y(t) = 4.905 * t² + 9.144 * t + 0.9144 meters<|im_end|>

Technical Details

The Locutusque/Hyperion-3.0-Mistral-7B-DPO model was fine-tuned on a carefully curated dataset of 20,000 preference pairs, where 4,000 examples were used to fine-tune. These examples were generated by GPT-4 to ensure the highest quality and relevance across various domains, including programming, medical texts, mathematical problems, and reasoning tasks. The training data was further optimized using Direct Preference Optimization (DPO) to align the model's outputs with human preferences and improve overall performance.

The hyperion-v3.0 dataset can be found here

Special thanks

🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.

Disclaimers

LM Studio is not the creator, originator, or owner of any Model featured in the Community Model Program. Each Community Model is created and provided by third parties. LM Studio does not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any Community Model. You understand that Community Models can produce content that might be offensive, harmful, inaccurate or otherwise inappropriate, or deceptive. Each Community Model is the sole responsibility of the person or entity who originated such Model. LM Studio may not monitor or control the Community Models and cannot, and does not, take responsibility for any such Model. LM Studio disclaims all warranties or guarantees about the accuracy, reliability or benefits of the Community Models. LM Studio further disclaims any warranty that the Community Model will meet your requirements, be secure, uninterrupted or available at any time or location, or error-free, viruses-free, or that any errors will be corrected, or otherwise. You will be solely responsible for any damage resulting from your use of or access to the Community Models, your downloading of any Community Model, or use of any other Community Model provided by or through LM Studio.