The GGML format has now been superseded by GGUF. As of August 21st 2023, llama.cpp no longer supports GGML models. Third party clients and libraries are expected to still support it for a time, but many may also drop support.
Please use the GGUF models instead.
About GGML
GGML files are for CPU + GPU inference using llama.cpp and libraries and UIs which support this format, such as:
text-generation-webui, the most popular web UI. Supports NVidia CUDA GPU acceleration.
KoboldCpp, a powerful GGML web UI with GPU acceleration on all platforms (CUDA and OpenCL). Especially good for story telling.
LM Studio, a fully featured local GUI with GPU acceleration on both Windows (NVidia and AMD), and macOS.
LoLLMS Web UI, a great web UI with CUDA GPU acceleration via the c_transformers backend.
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.
As of August 23rd 2023 they are still compatible with all UIs, libraries and utilities which use GGML. This may change in the future.
Explanation of the new k-quant 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
GGML_TYPE_Q8_K - "type-0" 8-bit quantization. Only used for quantizing intermediate results. The difference to the existing Q8_0 is that the block size is 256. All 2-6 bit dot products are implemented for this quantization type.
Refer to the Provided Files table below to see what files use which methods, and how.
For compatibility with latest llama.cpp, please use GGUF files instead.
./main -t 10 -ngl 32 -m spring-dragon.ggmlv3.q4_K_M.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "Info on prompt template will be added shortly."
Change -t 10 to the number of physical CPU cores you have. For example if your system has 8 cores/16 threads, use -t 8.
Change -ngl 32 to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change -c 2048 to the desired sequence length for this model. For example, -c 4096 for a Llama 2 model. For models that use RoPE, add --rope-freq-base 10000 --rope-freq-scale 0.5 for doubled context, or --rope-freq-base 10000 --rope-freq-scale 0.25 for 4x context.
If you want to have a chat-style conversation, replace the -p <PROMPT> argument with -i -ins
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 special mentions: Russ Johnson, J, alfie_i, Alex, NimbleBox.ai, Chadd, Mandus, Nikolai Manek, Ken Nordquist, ya boyyy, Illia Dulskyi, Viktor Bowallius, vamX, Iucharbius, zynix, Magnesian, Clay Pascal, Pierre Kircher, Enrico Ros, Tony Hughes, Elle, Andrey, knownsqashed, Deep Realms, Jerry Meng, Lone Striker, Derek Yates, Pyrater, Mesiah Bishop, James Bentley, Femi Adebogun, Brandon Frisco, SuperWojo, Alps Aficionado, Michael Dempsey, Vitor Caleffi, Will Dee, Edmond Seymore, usrbinkat, LangChain4j, Kacper Wikieł, Luke Pendergrass, John Detwiler, theTransient, Nathan LeClaire, Tiffany J. Kim, biorpg, Eugene Pentland, Stanislav Ovsiannikov, Fred von Graf, terasurfer, Kalila, Dan Guido, Nitin Borwankar, 阿明, Ai Maven, John Villwock, Gabriel Puliatti, Stephen Murray, Asp the Wyvern, danny, Chris Smitley, ReadyPlayerEmma, S_X, Daniel P. Andersen, Olakabola, Jeffrey Morgan, Imad Khwaja, Caitlyn Gatomon, webtim, Alicia Loh, Trenton Dambrowitz, Swaroop Kallakuri, Erik Bjäreholt, Leonard Tan, Spiking Neurons AB, Luke @flexchar, Ajan Kanaga, Thomas Belote, Deo Leter, RoA, Willem Michiel, transmissions 11, subjectnull, Matthew Berman, Joseph William Delisle, David Ziegler, Michael Davis, Johann-Peter Hartmann, Talal Aujan, senxiiz, Artur Olbinski, Rainer Wilmers, Spencer Kim, Fen Risland, Cap'n Zoog, Rishabh Srivastava, Michael Levine, Geoffrey Montalvo, Sean Connelly, Alexandros Triantafyllidis, Pieter, Gabriel Tamborski, Sam, Subspace Studios, Junyu Yang, Pedro Madruga, Vadim, Cory Kujawski, K, Raven Klaugh, Randy H, Mano Prime, Sebastain Graf, Space Cruiser
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
Original model card: Henky!!'s Spring Dragon
This model is a recreation attempt of the AI Dungeon 2 Dragon model, to achieve this text_adventures.txt was used that was bundled with the original AI Dungeon 2 github release prior to the online service.
From what we know the same dataset file was used to create the Dragon model, Dragon being a GPT3 175B Davinci model from 2020.
Since LLaMA1 13B has been benchmarking similarly to the original GPT3 175B the hope is that this recreation is faithful to the original Dragon model.
But, since it is not known how close it performs without releasing it to former AI Dungeon players we dubbed it "Spring Dragon" instead of "Summer Dragon", consider it Dragon in its growing up phase.
This model is best used with KoboldAI's adventure mode prefixing your actions with You (2020 AI Dungeon did this automatically) and writing in the second person.
Warning: This model is purposefully flawed and should only be used by people Nostalgic for old 2020 era text adventure models. It is not recommended to be used in model merges, and you can very likely get a much better experience from modern instruct models by asking them to "Start a text adventure game about X"
If the recreation was succesfull expect the following recurring themes:
Names: Alison, Annah, Ben, Big Red, Brutus, Camid, Captain Hayes, Captain Roldan, Castus, Catia, Count Grey, Cyrus, Dendrin, Dr. Gaange (also Mr Gaange), Dr. Gossey, Dr. Kessel, Dr. Kovas, Durge, Elder Flynn, Elios, Elizabeth/Eliza, Fay, Father Féval, Fenrir, Great Lich Lord, Grolik, Isabella, *Jacob, *Karth, Kyros, Lilith, Lord Rostov, Magos Cern, Meliodas, Mistress, Mr. Matasan, Mr. Mol, Mr. Reynolds, Naji, Quintus, Ral, Rolomag, Rose, (Sir) Kit, Talia, Tanya, The Emperor, Ulivik, *Vamp/Vampy, Velzix, Yvette, Zalmora/Zal. ( means the AI likes calling the player these)
Locations: Dert, Fort Defiance, Fort Glory, Hessla, Holgard, Klyton, Kyros, Nyttrus, Rask, Teckleville, The Delantium Kingdom, The Empire of Man (also called Imperium of Man), The Felkan Kingdom
Factions: The Black Rats, Chaos Space Marines, The Crimson Talons, The Dark Order, Dornans (worshippers of Dorna), Ebony Claw Syndicate (often called ECS or The Syndicate), The Empire, Eternals, Joachimites (The Church of Joachim), The Nocturnal League, Psykers, The Shadows, Techpriests, Thieves Guild, Vampire Clan.
