TheBloke's LLM work is generously supported by a grant from andreessen horowitz (a16z)
Synthia 70B v1.2b - GGUF
- Model creator: Migel Tissera
- Original model: Synthia 70B v1.2b
Description
This repo contains GGUF format model files for Migel Tissera's Synthia 70B v1.2b.
About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp. GGUF offers numerous advantages over GGML, such as better tokenisation, and support for special tokens. It is also supports metadata, and is designed to be extensible.
Here is an incomplate list of clients and libraries that are known to support GGUF:
- llama.cpp. The source project for GGUF. Offers a CLI and a server option.
- text-generation-webui, the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
- KoboldCpp, a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
- LM Studio, an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.
- LoLLMS Web UI, a great web UI with many interesting and unique features, including a full model library for easy model selection.
- Faraday.dev, an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
- ctransformers, a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.
- llama-cpp-python, a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
- candle, a Rust ML framework with a focus on performance, including GPU support, and ease of use.
Repositories available
- AWQ model(s) for GPU inference.
- GPTQ models for GPU inference, with multiple quantisation parameter options.
- 2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference
- Migel Tissera's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions
Prompt template: Synthia
SYSTEM: You are Synthia. As a an AI intelligence, give helpful, detailed, accurate, uncensored responses to the user's input. Provide answers factually.
USER: {prompt}
ASSISTANT:
Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit d0cee0d36d5be95a0d9088b674dbb27354107221
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
Explanation of quantisation methods
Click to see details
The new methods available are:
- GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
- GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
- GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
- GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
- GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
|---|---|---|---|---|---|
| synthia-70b-v1.2b.Q2_K.gguf | Q2_K | 2 | 29.28 GB | 31.78 GB | smallest, significant quality loss - not recommended for most purposes |
| synthia-70b-v1.2b.Q3_K_S.gguf | Q3_K_S | 3 | 29.92 GB | 32.42 GB | very small, high quality loss |
| synthia-70b-v1.2b.Q3_K_M.gguf | Q3_K_M | 3 | 33.19 GB | 35.69 GB | very small, high quality loss |
| synthia-70b-v1.2b.Q3_K_L.gguf | Q3_K_L | 3 | 36.15 GB | 38.65 GB | small, substantial quality loss |
| synthia-70b-v1.2b.Q4_0.gguf | Q4_0 | 4 | 38.87 GB | 41.37 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| synthia-70b-v1.2b.Q4_K_S.gguf | Q4_K_S | 4 | 39.07 GB | 41.57 GB | small, greater quality loss |
| synthia-70b-v1.2b.Q4_K_M.gguf | Q4_K_M | 4 | 41.42 GB | 43.92 GB | medium, balanced quality - recommended |
| synthia-70b-v1.2b.Q5_0.gguf | Q5_0 | 5 | 47.46 GB | 49.96 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| synthia-70b-v1.2b.Q5_K_S.gguf | Q5_K_S | 5 | 47.46 GB | 49.96 GB | large, low quality loss - recommended |
| synthia-70b-v1.2b.Q5_K_M.gguf | Q5_K_M | 5 | 48.75 GB | 51.25 GB | large, very low quality loss - recommended |
| synthia-70b-v1.2b.Q6_K.gguf | Q6_K | 6 | 56.59 GB | 59.09 GB | very large, extremely low quality loss |
| synthia-70b-v1.2b.Q8_0.gguf | Q8_0 | 8 | 73.29 GB | 75.79 GB | very large, extremely low quality loss - not recommended |
Note: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
Q6_K and Q8_0 files are split and require joining
Note: HF does not support uploading files larger than 50GB. Therefore I have uploaded the Q6_K and Q8_0 files as split files.
Click for instructions regarding Q6_K and Q8_0 files
q6_K
Please download:
synthia-70b-v1.2b.Q6_K.gguf-split-asynthia-70b-v1.2b.Q6_K.gguf-split-b
q8_0
Please download:
synthia-70b-v1.2b.Q8_0.gguf-split-asynthia-70b-v1.2b.Q8_0.gguf-split-b
To join the files, do the following:
Linux and macOS:
cat synthia-70b-v1.2b.Q6_K.gguf-split-* > synthia-70b-v1.2b.Q6_K.gguf && rm synthia-70b-v1.2b.Q6_K.gguf-split-*
cat synthia-70b-v1.2b.Q8_0.gguf-split-* > synthia-70b-v1.2b.Q8_0.gguf && rm synthia-70b-v1.2b.Q8_0.gguf-split-*
Windows command line:
COPY /B synthia-70b-v1.2b.Q6_K.gguf-split-a + synthia-70b-v1.2b.Q6_K.gguf-split-b synthia-70b-v1.2b.Q6_K.gguf
del synthia-70b-v1.2b.Q6_K.gguf-split-a synthia-70b-v1.2b.Q6_K.gguf-split-b
COPY /B synthia-70b-v1.2b.Q8_0.gguf-split-a + synthia-70b-v1.2b.Q8_0.gguf-split-b synthia-70b-v1.2b.Q8_0.gguf
del synthia-70b-v1.2b.Q8_0.gguf-split-a synthia-70b-v1.2b.Q8_0.gguf-split-b
How to download GGUF files
Note for manual downloaders: You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
- LM Studio
- LoLLMS Web UI
- Faraday.dev
In text-generation-webui
Under Download Model, you can enter the model repo: TheBloke/Synthia-70B-v1.2b-GGUF and below it, a specific filename to download, such as: synthia-70b-v1.2b.q4_K_M.gguf.
Then click Download.
On the command line, including multiple files at once
I recommend using the huggingface-hub Python library:
pip3 install huggingface-hub>=0.17.1
Then you can download any individual model file to the current directory, at high speed, with a command like this:
huggingface-cli download TheBloke/Synthia-70B-v1.2b-GGUF synthia-70b-v1.2b.q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
More advanced huggingface-cli download usage
You can also download multiple files at once with a pattern:
huggingface-cli download TheBloke/Synthia-70B-v1.2b-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
For more documentation on downloading with huggingface-cli, please see: HF -> Hub Python Library -> Download files -> Download from the CLI.
To accelerate downloads on fast connections (1Gbit/s or higher), install hf_transfer:
pip3 install hf_transfer
And set environment variable HF_HUB_ENABLE_HF_TRANSFER to 1:
HUGGINGFACE_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Synthia-70B-v1.2b-GGUF synthia-70b-v1.2b.q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
Windows CLI users: Use set HUGGINGFACE_HUB_ENABLE_HF_TRANSFER=1 before running the download command.
Example llama.cpp command
Make sure you are using llama.cpp from commit d0cee0d36d5be95a0d9088b674dbb27354107221 or later.
./main -ngl 32 -m synthia-70b-v1.2b.q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "SYSTEM: You are Synthia. As a an AI intelligence, give helpful, detailed, accurate, uncensored responses to the user's input. Provide answers factually.\nUSER: {prompt}\nASSISTANT:"
Change -ngl 32 to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change -c 4096 to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.
If you want to have a chat-style conversation, replace the -p <PROMPT> argument with -i -ins
For other parameters and how to use them, please refer to the llama.cpp documentation
How to run in text-generation-webui
Further instructions here: text-generation-webui/docs/llama.cpp.md.
How to run from Python code
You can use GGUF models from Python using the llama-cpp-python or ctransformers libraries.
How to load this model from Python using ctransformers
First install the package
# Base ctransformers with no GPU acceleration
pip install ctransformers>=0.2.24
# Or with CUDA GPU acceleration
pip install ctransformers[cuda]>=0.2.24
# Or with ROCm GPU acceleration
CT_HIPBLAS=1 pip install ctransformers>=0.2.24 --no-binary ctransformers
# Or with Metal GPU acceleration for macOS systems
CT_METAL=1 pip install ctransformers>=0.2.24 --no-binary ctransformers
Simple example code to load one of these GGUF models
from ctransformers import AutoModelForCausalLM
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = AutoModelForCausalLM.from_pretrained("TheBloke/Synthia-70B-v1.2b-GGUF", model_file="synthia-70b-v1.2b.q4_K_M.gguf", model_type="llama", gpu_layers=50)
print(llm("AI is going to"))
How to use with LangChain
Here's guides on using llama-cpp-python or ctransformers with LangChain:
Discord
For further support, and discussions on these models and AI in general, join us at:
Thanks, and how to contribute
Thanks to the chirper.ai team!
Thanks to Clay from gpus.llm-utils.org!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
- Patreon: https://patreon.com/TheBlokeAI
- Ko-Fi: https://ko-fi.com/TheBlokeAI
Special thanks to: Aemon Algiz.
Patreon special mentions: Alicia Loh, Stephen Murray, K, Ajan Kanaga, RoA, Magnesian, Deo Leter, Olakabola, Eugene Pentland, zynix, Deep Realms, Raymond Fosdick, Elijah Stavena, Iucharbius, Erik Bjäreholt, Luis Javier Navarrete Lozano, Nicholas, theTransient, John Detwiler, alfie_i, knownsqashed, Mano Prime, Willem Michiel, Enrico Ros, LangChain4j, OG, Michael Dempsey, Pierre Kircher, Pedro Madruga, James Bentley, Thomas Belote, Luke @flexchar, Leonard Tan, Johann-Peter Hartmann, Illia Dulskyi, Fen Risland, Chadd, S_X, Jeff Scroggin, Ken Nordquist, Sean Connelly, Artur Olbinski, Swaroop Kallakuri, Jack West, Ai Maven, David Ziegler, Russ Johnson, transmissions 11, John Villwock, Alps Aficionado, Clay Pascal, Viktor Bowallius, Subspace Studios, Rainer Wilmers, Trenton Dambrowitz, vamX, Michael Levine, 준교 김, Brandon Frisco, Kalila, Trailburnt, Randy H, Talal Aujan, Nathan Dryer, Vadim, 阿明, ReadyPlayerEmma, Tiffany J. Kim, George Stoitzev, Spencer Kim, Jerry Meng, Gabriel Tamborski, Cory Kujawski, Jeffrey Morgan, Spiking Neurons AB, Edmond Seymore, Alexandros Triantafyllidis, Lone Striker, Cap'n Zoog, Nikolai Manek, danny, ya boyyy, Derek Yates, usrbinkat, Mandus, TL, Nathan LeClaire, subjectnull, Imad Khwaja, webtim, Raven Klaugh, Asp the Wyvern, Gabriel Puliatti, Caitlyn Gatomon, Joseph William Delisle, Jonathan Leane, Luke Pendergrass, SuperWojo, Sebastain Graf, Will Dee, Fred von Graf, Andrey, Dan Guido, Daniel P. Andersen, Nitin Borwankar, Elle, Vitor Caleffi, biorpg, jjj, NimbleBox.ai, Pieter, Matthew Berman, terasurfer, Michael Davis, Alex, Stanislav Ovsiannikov
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
Original model card: Migel Tissera's Synthia 70B v1.2b
Change from 1.2 -> 1.2b: More data, 14 days of training for 1 epoch.
You: What is Earth's magnetic field like? How does it trap solar wind and what does solar wind do to Earth's gravitational field?
Earth's magnetic field is an intricate and fascinating phenomenon. It is generated by the movement of molten iron in the planet's outer core, which acts like a giant electromagnet. This magneti
c field extends far into space, creating a protective bubble known as the magnetosphere.
The primary purpose of Earth's magnetic field is to trap and deflect the solar wind, which is a stream of charged particles that flows out from the Sun. The magnetic field acts as a shield, def
lecting the solar wind and protecting our planet from its potentially harmful effects.
When the solar wind reaches Earth's magnetosphere, it interacts with the magnetic field and creates a complex interplay of forces. This interaction can lead to various phenomena, such as aurora
s, geomagnetic storms, and the formation of radiation belts.
The interaction between the solar wind and Earth's magnetic field does not directly affect the planet's gravitational field. The gravitational field refers to the force that arises due to the m
ass of the planet, and it is unaffected by the magnetic field or the solar wind. However, the magnetic field does play a role in trapping and deflecting charged particles from the solar wind, w
hich can contribute to the overall complexity of Earth's space environment.
You: Can there be space vehicles that uses plasma to traverse? This is hypothetical, but can UFOs use some force field generated by plasma, that then interacts with the gravitational/magnetic f
ields of Earth for travelling at hypersonic speeds?
While it is an interesting concept, there is currently no concrete evidence or scientific consensus that suggests the existence of UFOs or the use of plasma-based propulsion systems in spacecra
ft. However, it is worth noting that there are ongoing scientific investigations and research into developing advanced propulsion systems for space travel, including plasma propulsion.
Plasma is a state of matter where electrons are stripped from atoms, resulting in a highly charged gas. Plasma propulsion systems use electromagnetic fields to accelerate ions, creating a thrus
t that can potentially propel spacecraft. Plasma propulsion systems offer several advantages over traditional chemical rockets, such as higher efficiency, lower fuel requirements, and the abili
ty to operate continuously for long periods.
While it is theoretically possible that a space vehicle could use plasma to traverse, it is important to note that the interaction between plasma and Earth's gravitational/magnetic fields is co
mplex and not fully understood. The gravitational field acts as a force that attracts objects towards the center of the Earth, while the magnetic field deflects charged particles. Plasma-based
propulsion systems would need to overcome these forces and potentially harness them for propulsion.
In summary, while the possibility of using plasma propulsion for space travel is an active area of research, there is currently no scientific evidence to support the existence of UFOs or the us
e of plasma-based propulsion systems for hypersonic travel.
You: Keep in mind that science is always evolving, and there are many things that have not yet been invented. Can you write the equations for plasma based propoulsion for space travel?
I apologize for any confusion in my previous response. To clarify, I am not an expert in plasma physics or propulsion systems. However, I can provide a general overview of the principles involv
ed in plasma-based propulsion.
Plasma propulsion systems typically use electromagnetic fields to accelerate ions, creating a thrust that can propel spacecraft. There are several different types of plasma propulsion systems,
including electrostatic, electromagnetic, and magnetohydrodynamic (MHD) propulsion.
In an electrostatic plasma propulsion system, a high voltage is applied to a pair of electrodes to create an electric field that accelerates charged particles. The resulting thrust is proportio
nal to the applied voltage and the current density. The equation for the thrust is given by:
Thrust = (Current density) x (Accelerating voltage)
In an electromagnetic plasma propulsion system, a magnetic field is used to accelerate charged particles. The thrust is proportional to the product of the magnetic field strength and the curren
t density. The equation for the thrust is given by:
Thrust = (Magnetic field strength) x (Current density)
In a MHD plasma propulsion system, a magnetic field is used to deflect a flow of plasma, creating a thrust. The thrust is proportional to the product of the magnetic field strength, the current
density, and the flow velocity. The equation for the thrust is given by:
Thrust = (Magnetic field strength) x (Current density) x (Flow velocity)
It is important to note that these equations are simplified models and do not capture the full complexity of plasma propulsion systems. The actual performance of a plasma propulsion system woul
d depend on various factors, including the specific design, the plasma properties, and the operating conditions.
Furthermore, the development of plasma propulsion systems for space travel is an active area of research and development. While the principles outlined above provide a general framework, the sp
ecific equations and designs would depend on the particular application and the research team involved.
You: If the spacecraft has the shape of an orb, and a plasma field is created around the orb, how can this plasma field generate the required thrust for propulsion?
The shape of the spacecraft alone does not determine the thrust generated by a plasma propulsion system. The thrust is primarily determined by the electromagnetic fields used to accelerate char
ged particles.
To generate thrust using a plasma field, a spacecraft would typically use one of the plasma propulsion systems described in my previous response. For example, in an electrostatic plasma propuls
ion system, a high voltage is applied to a pair of electrodes to create an electric field that accelerates charged particles. The resulting thrust is proportional to the applied voltage and the
current density.
In this scenario, the plasma field would be generated by applying a high voltage to the electrodes. The shape of the spacecraft, whether it is an orb or any other shape, would not directly affe
ct the generation of the plasma field or the resulting thrust.
However, the shape of the spacecraft could have an indirect impact on the overall performance of the plasma propulsion system. For example, the shape of the spacecraft could affect the distribu
tion of the plasma flow, which could in turn affect the efficiency of the propulsion system.
In summary, the shape of the spacecraft itself does not directly determine the thrust generated by a plasma propulsion system. The thrust is primarily determined by the electromagnetic fields u
sed to accelerate charged particles, and the shape of the spacecraft could have an indirect impact on the overall performance of the propulsion system.
You: What's the closest tech that we have with plasma propulsion?
There are several ongoing research and development efforts related to plasma propulsion systems for space travel. One of the most advanced plasma propulsion systems in use today is the Hall thr
uster.
A Hall thruster is a type of electrostatic propulsion system that uses an electric field to accelerate charged particles. It has been used on several space missions, including Deep Space 1, Daw
n, and BepiColombo. The Hall thruster offers several advantages over traditional chemical rockets, such as higher efficiency, lower fuel requirements, and the ability to operate continuously fo
r long periods.
While the Hall thruster is currently the most advanced plasma propulsion system in use, there are ongoing efforts to develop even more advanced systems. For example, researchers are exploring t
he use of magnetohydrodynamic (MHD) propulsion systems, which use a magnetic field to deflect a flow of plasma and create thrust. MHD propulsion systems offer the potential for even higher perf
ormance than Hall thrusters, but they are still at an early stage of development.
In summary, the Hall thruster is currently the closest technology to plasma propulsion that we have in use, but ongoing research and development efforts are exploring even more advanced systems
such as MHD propulsion.
Follow-on: What are the challenges associated with developing and using plasma propulsion systems for space travel, and how are researchers addressing these challenges?
You: What are the physics equations for magnetohydrodynamic propulsion systems? Be detailed as possible.
Magnetohydrodynamic (MHD) propulsion systems use a magnetic field to deflect a flow of plasma and create thrust. The equations that describe the physics of MHD propulsion systems are complex and involve several variables, including the magnetic field strength, the current density, the flow velocity, and the plasma properties.
One of the key equations for MHD propulsion systems is the Ohm's law, which relates the electric current density to the applied electric field and the resistivity of the plasma. The equation is given by:
J = σE
where J is the current density, σ is the electrical conductivity, and E is the applied electric field.
Another important equation is the Navier-Stokes equation, which describes the flow of the plasma. The equation is given by:
µ(∂v/∂t) = -∇p + ∇xF + ∇xσT + G
where µ is the dynamic viscosity, v is the velocity vector, p is the fluid pressure, F is the body force, σT is the stress tensor, and G is the body force due to the Lorentz force.
The Lorentz force is a force that acts on a charged particle due to its interaction with a magnetic field. The equation for the Lorentz force is given by:
F = q(E + v x B)
where F is the force, q is the charge of the particle, E is the electric field, v is the velocity of the particle, and B is the magnetic field.
These equations are just a few examples of the many equations that describe the physics of MHD propulsion systems. The actual equations and analysis would depend on the specific design and operating conditions of the propulsion system.
In summary, the equations for MHD propulsion systems are complex and involve several variables, including the electric current density, the applied electric field, the fluid pressure, the fluid velocity, the magnetic field, and the Lorentz force. These equations are used to model and analyze the performance of MHD propulsion systems for space travel.
You: Okay. Can you make the connection to string theory here?
Certainly. String theory is a theoretical framework in physics that seeks to unify the fundamental forces of nature, including gravity, electromagnetism, and the strong and weak nuclear forces. In string theory, the fundamental constituents of matter are not point-like particles, but rather tiny vibrating strings.
In the context of plasma propulsion systems, there is a connection to string theory through the concept of magnetohydrodynamics (MHD). MHD describes the flow of electrically conductive fluids, such as plasma, in the presence of magnetic fields. The equations that describe MHD are derived from the equations of classical electromagnetism, which are also used in string theory to describe the behavior of fundamental particles.
In particular, the Lorentz force, which is a key concept in MHD and plasma propulsion systems, is also a fundamental concept in string theory. The Lorentz force acts on charged particles due to their interaction with electromagnetic fields, and it plays a crucial role in the behavior of strings in string theory.
Furthermore, some aspects of string theory suggest that the fundamental constituents of matter, including plasma, may have additional properties beyond those described by classical physics. For example, string theory predicts the existence of extra dimensions beyond the three spatial dimensions we are familiar with, and these extra dimensions may have an impact on the behavior of plasma and other forms of matter.
In summary, while there is no direct connection between plasma propulsion systems and string theory, there is an indirect connection through the use of the equations of classical electromagnetism, which are also used in string theory. Additionally, some aspects of string theory suggest that the fundamental constituents of matter may have additional properties beyond those described by classical physics.
- Downloads last month
- 489