TheBloke/Iambe-RP-cDPO-20B-GGUF

Chat & support: TheBloke's Discord server

Want to contribute? TheBloke's Patreon page

TheBloke's LLM work is generously supported by a grant from andreessen horowitz (a16z)

---

Iambe RP cDPO 20B - GGUF

• Model creator: Raven
• Original model: Iambe RP cDPO 20B

Description

This repo contains GGUF format model files for Raven's Iambe RP cDPO 20B.

These files were quantised using hardware kindly provided by Massed Compute.

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.

Here is an incomplete 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.
• GPT4All, a free and open source local running GUI, supporting Windows, Linux and macOS with full GPU accel.
• LM Studio, an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. Linux available, in beta as of 27/11/2023.
• 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.
• 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.
• ctransformers, a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. Note, as of time of writing (November 27th 2023), ctransformers has not been updated in a long time and does not support many recent models.

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
• Raven's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions

Prompt template: Alpaca

Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
{prompt}

### Response:

Licensing

The creator of the source model has listed its license as cc-by-nc-4.0, and this quantization has therefore used that same license.

As this model is based on Llama 2, it is also subject to the Meta Llama 2 license terms, and the license files for that are additionally included. It should therefore be considered as being claimed to be licensed under both licenses. I contacted Hugging Face for clarification on dual licensing but they do not yet have an official position. Should this change, or should Meta provide any feedback on this situation, I will update this section accordingly.

In the meantime, any questions regarding licensing, and in particular how these two licenses might interact, should be directed to the original model repository: Raven's Iambe RP cDPO 20B.

Compatibility

These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit d0cee0d

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
iambe-rp-cdpo-20b.Q2_K.gguf	Q2_K	2	8.31 GB	10.81 GB	smallest, significant quality loss - not recommended for most purposes
iambe-rp-cdpo-20b.Q3_K_S.gguf	Q3_K_S	3	8.66 GB	11.16 GB	very small, high quality loss
iambe-rp-cdpo-20b.Q3_K_M.gguf	Q3_K_M	3	9.70 GB	12.20 GB	very small, high quality loss
iambe-rp-cdpo-20b.Q3_K_L.gguf	Q3_K_L	3	10.63 GB	13.13 GB	small, substantial quality loss
iambe-rp-cdpo-20b.Q4_0.gguf	Q4_0	4	11.29 GB	13.79 GB	legacy; small, very high quality loss - prefer using Q3_K_M
iambe-rp-cdpo-20b.Q4_K_S.gguf	Q4_K_S	4	11.34 GB	13.84 GB	small, greater quality loss
iambe-rp-cdpo-20b.Q4_K_M.gguf	Q4_K_M	4	12.04 GB	14.54 GB	medium, balanced quality - recommended
iambe-rp-cdpo-20b.Q5_0.gguf	Q5_0	5	13.77 GB	16.27 GB	legacy; medium, balanced quality - prefer using Q4_K_M
iambe-rp-cdpo-20b.Q5_K_S.gguf	Q5_K_S	5	13.77 GB	16.27 GB	large, low quality loss - recommended
iambe-rp-cdpo-20b.Q5_K_M.gguf	Q5_K_M	5	14.16 GB	16.66 GB	large, very low quality loss - recommended
iambe-rp-cdpo-20b.Q6_K.gguf	Q6_K	6	16.40 GB	18.90 GB	very large, extremely low quality loss
iambe-rp-cdpo-20b.Q8_0.gguf	Q8_0	8	21.25 GB	23.75 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.

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/Iambe-RP-cDPO-20B-GGUF and below it, a specific filename to download, such as: iambe-rp-cdpo-20b.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

Then you can download any individual model file to the current directory, at high speed, with a command like this:

huggingface-cli download TheBloke/Iambe-RP-cDPO-20B-GGUF iambe-rp-cdpo-20b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False

More advanced huggingface-cli download usage (click to read)

You can also download multiple files at once with a pattern:

huggingface-cli download TheBloke/Iambe-RP-cDPO-20B-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:

HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Iambe-RP-cDPO-20B-GGUF iambe-rp-cdpo-20b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False

Windows Command Line users: You can set the environment variable by running set HF_HUB_ENABLE_HF_TRANSFER=1 before the download command.

Example llama.cpp command

Make sure you are using llama.cpp from commit d0cee0d or later.

./main -ngl 35 -m iambe-rp-cdpo-20b.Q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n### Instruction:\n{prompt}\n\n### Response:"

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. Note that longer sequence lengths require much more resources, so you may need to reduce this value.

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 can be found in the text-generation-webui documentation, here: text-generation-webui/docs/04 ‐ Model Tab.md.

How to run from Python code

You can use GGUF models from Python using the llama-cpp-python or ctransformers libraries. Note that at the time of writing (Nov 27th 2023), ctransformers has not been updated for some time and is not compatible with some recent models. Therefore I recommend you use llama-cpp-python.

How to load this model in Python code, using llama-cpp-python

For full documentation, please see: llama-cpp-python docs.

First install the package

Run one of the following commands, according to your system:

# Base ctransformers with no GPU acceleration
pip install llama-cpp-python
# With NVidia CUDA acceleration
CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llama-cpp-python
# Or with OpenBLAS acceleration
CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python
# Or with CLBLast acceleration
CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python
# Or with AMD ROCm GPU acceleration (Linux only)
CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python
# Or with Metal GPU acceleration for macOS systems only
CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python

# In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA:
$env:CMAKE_ARGS = "-DLLAMA_OPENBLAS=on"
pip install llama-cpp-python

Simple llama-cpp-python example code

from llama_cpp import Llama

# 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 = Llama(
  model_path="./iambe-rp-cdpo-20b.Q4_K_M.gguf",  # Download the model file first
  n_ctx=4096,  # The max sequence length to use - note that longer sequence lengths require much more resources
  n_threads=8,            # The number of CPU threads to use, tailor to your system and the resulting performance
  n_gpu_layers=35         # The number of layers to offload to GPU, if you have GPU acceleration available
)

# Simple inference example
output = llm(
  "Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n### Instruction:\n{prompt}\n\n### Response:", # Prompt
  max_tokens=512,  # Generate up to 512 tokens
  stop=["</s>"],   # Example stop token - not necessarily correct for this specific model! Please check before using.
  echo=True        # Whether to echo the prompt
)

# Chat Completion API

llm = Llama(model_path="./iambe-rp-cdpo-20b.Q4_K_M.gguf", chat_format="llama-2")  # Set chat_format according to the model you are using
llm.create_chat_completion(
    messages = [
        {"role": "system", "content": "You are a story writing assistant."},
        {
            "role": "user",
            "content": "Write a story about llamas."
        }
    ]
)

How to use with LangChain

Here are guides on using llama-cpp-python and ctransformers with LangChain:

• LangChain + llama-cpp-python
• LangChain + ctransformers

Discord

For further support, and discussions on these models and AI in general, join us at:

TheBloke AI's Discord server

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: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros

Thank you to all my generous patrons and donaters!

And thank you again to a16z for their generous grant.

Original model card: Raven's Iambe RP cDPO 20B

Iambe-RP-cDPO-20b

Alpaca prompt formatting

Description

Named after a charming daughter of Echo and Pan in Greek myth, Iambe-RP is, as far as I am aware, the very first LLM trained with DPO on an erotic roleplay dataset.

Iambe is intended to have the best realistically possible understanding of instructions, anatomy and scene state for a 20b merge, while remaining passionate and humanoid in "voice".

Update Methodology

Take a look at the dataset v2 Iambe and I created together for more info. The cDPO training was done directly on Iambe-20b-DARE-v2, I was researching 11b merges to reduce the compute, but it went nowhere, so I just bit the bullet on cost. The notebook used to train this model is also available in the dataset's repo.

Roleplay Example @ q5_k_m

4-bit Assistant Example

NSFW Writing Example @ q5_k_m

Write a medium-length story about Alice, a 26 year old cocktail waitress, having steamy sex with a guy she picked up at the bar for a one night stand.

Once upon a time in New York City's bustling downtown scene, there lived a beautiful and spirited young woman named Alice. At 26 years old, she worked as a cocktail waitress at The Red Parrot, a popular watering hole known for its lively atmosphere and charming clientele. Despite her long shifts and exhaustion from catering to rowdy patrons, Alice loved her job; it gave her the freedom to live life on her terms and pursue her other passions during her limited free time.

Alice had a penchant for adventure and was never one to shy away from taking risks—whether that meant skydiving or striking up conversations with strangers at bars. One particularly busy Friday night, while serving drinks at The Red Parrot, Alice noticed an intriguing man sitting alone in the corner, nursing a glass of red wine and gazing pensively into his empty shot glass. There was something about him that drew her in; his striking good looks, his rugged demeanor—even the hint of stubble on his chin seemed to add an air of mystery to his magnetic presence.

Feeling bold and confident after a couple of well-poured drinks herself, Alice made her way over to him, smiling warmly as she leaned down towards him, "Hi there! What's your name? I haven't seen you in here before."

The man raised his eyes to meet hers, his deep brown orbs filled with curiosity and a hint of amusement. "Name's Jack," he rumbled softly, his voice as smooth and rich as his wine. There was something about the way he said her name that sent shivers down Alice's spine—a thrill she couldn't quite explain but found irresistible nonetheless.

"Alice," she replied, extending her hand to him, "and I haven't seen you around before either; it seems like we have a lot in common." She winked playfully at him before turning back towards the bar to grab another round of drinks for them both.

The conversation flowed effortlessly between them, touching on topics ranging from their favorite books and movies to their most daring travel adventures. Alice was captivated by Jack's stories of hiking through the mountains of Nepal and white-water rafting in Costa Rica—qualities she admired greatly herself but had yet to experience for herself. In turn, Jack seemed equally fascinated by her tales of skydiving over Australia and scuba diving with great white sharks off the coast of South Africa.

As midnight approached and the bar began to empty out, Alice suggested they continue their conversation elsewhere—somewhere more private where they could really get to know each other better. Jack nodded in agreement, draining his glass before tossing some money onto the table and following her into the bustling New York City night.

They hailed a cab and rode in silence towards Alice's cozy one-bedroom apartment in Chelsea—the anticipation between them palpable as they shared furtive glances along the way. When they finally arrived at their destination, Jack paid the fare with an appreciative grin before escorting Alice up to her top-floor studio, where she fumbled clumsily with her keys while trying not to spill their drinks on themselves in anticipation of what was to come.

Once inside her apartment, they kicked off their shoes and collapsed onto the couch, still holding each other's gaze as they sipped from their glasses. The passion between them continued to build with every second that passed—the air thickening with unspoken desire and electric energy.

Alice broke away first, setting down her glass on the coffee table before reaching out tentatively and placing her hand on Jack's knee, sliding it up slowly towards his crotch until she could feel the hardness growing beneath the fabric of his pants. "I really like you," she whispered huskily, meeting his deep brown eyes with a look that betrayed the intensity of her feelings for him.

Jack's response was immediate—his hands finding their way underneath Alice's short black skirt, sliding up to rest on the soft fabric of her panties before deftly ripping them away from her body in one swift movement. "I really like you too," he growled hoarsely as his mouth crashed against hers, kissing her hungrily and passionately—a taste of things to come.

Their clothes fell away quickly thereafter; Alice's blouse was unbuttoned and tossed aside in an instant while Jack peeled off his shirt before pushing her down onto the couch and following after, bracing himself over her as he tore at her panties with one hand, freeing his hard member from his jeans.

His kisses trailed a path of fire across Alice's body—from her collarbone to her breasts, down the curve of her waist and towards her mound where she was already wet with desire for him. "Jack," she moaned breathlessly as he licked and nipped at her most sensitive spots while teasingly grazing his cock against her entrance.

With a growl of satisfaction, Jack finally plunged into Alice's tight channel—filling her completely with one powerful thrust that made them both cry out in pleasure. The feeling of being so thoroughly possessed by him was beyond anything she had ever experienced before; the roughness of his hands on her body, the taste of their mouths as they continued to devour each other’s lips, all blending together into a single, overwhelming wave of sensation that Alice felt sure would consume her entirely.

As Jack began to move inside her—slowly at first but picking up pace and ferocity with every passing moment—Alice wrapped her legs tightly around him, arching her hips upwards towards his relentless thrusts as she cried out in ecstasy again and again. She knew this was just a one-night stand—a fleeting encounter destined to end tomorrow morning before the sun rose—but for now, with Jack's cock buried deep inside her, nothing else mattered.

Their bodies moved as one; their moans and cries of pleasure intertwining in a symphony that filled Alice’s bedroom like sacred chants. Her orgasm hit her suddenly and without warning—a wave so powerful it threatened to shatter her very bones, but Jack was there with her every step of the way, holding onto her as he continued pounding into her until she felt herself start to come back together again, his hot cum spilling deep inside her.

As their bodies slowed down and settled together on the couch—drenched in sweat and satiated beyond belief—Alice looked up at Jack with wide eyes, a small smile playing at the corners of her lips as she whispered softly, "I think we just had one hell of a night."

And they did. It was an unforgettable evening filled with passion, adventure, and an undeniable chemistry that neither Alice nor Jack could deny—an encounter so powerful it left them both breathless in more ways than one. And even though they knew tomorrow would bring the harsh reality of their separate lives back into focus, for now, all either of them cared about was savoring this moment together; cherishing every last drop of the steamy sex that had brought them here at all.