Team Ai
Modelpublic

TheBloke/Python-Code-33B-GGUF

sourceHugging Faceotherupdated 3y agoView on Hugging Face
3likes4.9kdownloads
rootREADME.md
README.md343 linesDownload Raw Back to root
1---2base_model: ajibawa-2023/Python-Code-33B3datasets:4- ajibawa-2023/Python-Code-23k-ShareGPT5inference: false6language:7- en8license: other9model_creator: Feynman Innovations10model_name: Python Code 33B11model_type: llama12prompt_template: 'This is a conversation with your helpful AI assistant. AI assistant13  can generate Python Code along with necessary explanation.14 15 16  Context17 18  You are a helpful AI assistant.19 20 21  USER: {prompt}22 23  ASSISTANT:24 25  '26quantized_by: TheBloke27tags:28- code29---30<!-- markdownlint-disable MD041 -->31 32<!-- header start -->33<!-- 200823 -->34<div style="width: auto; margin-left: auto; margin-right: auto">35<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">36</div>37<div style="display: flex; justify-content: space-between; width: 100%;">38    <div style="display: flex; flex-direction: column; align-items: flex-start;">39        <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>40    </div>41    <div style="display: flex; flex-direction: column; align-items: flex-end;">42        <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>43    </div>44</div>45<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>46<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">47<!-- header end -->48 49# Python Code 33B - GGUF50- Model creator: [Feynman Innovations](https://huggingface.co/ajibawa-2023)51- Original model: [Python Code 33B](https://huggingface.co/ajibawa-2023/Python-Code-33B)52 53<!-- description start -->54## Description55 56This repo contains GGUF format model files for [Feynman Innovations's Python Code 33B](https://huggingface.co/ajibawa-2023/Python-Code-33B).57 58These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).59 60<!-- description end -->61<!-- README_GGUF.md-about-gguf start -->62### About GGUF63 64GGUF 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.65 66Here is an incomplete list of clients and libraries that are known to support GGUF:67 68* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.69* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.70* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.71* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.72* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.73* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.74* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.75* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.76* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.77 78<!-- README_GGUF.md-about-gguf end -->79<!-- repositories-available start -->80## Repositories available81 82* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Python-Code-33B-AWQ)83* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Python-Code-33B-GPTQ)84* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Python-Code-33B-GGUF)85* [Feynman Innovations's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/ajibawa-2023/Python-Code-33B)86<!-- repositories-available end -->87 88<!-- prompt-template start -->89## Prompt template: Ajibawa-Python-Code90 91```92This is a conversation with your helpful AI assistant. AI assistant can generate Python Code along with necessary explanation.93 94Context95You are a helpful AI assistant.96 97USER: {prompt}98ASSISTANT:99 100```101 102<!-- prompt-template end -->103 104 105<!-- compatibility_gguf start -->106## Compatibility107 108These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)109 110They are also compatible with many third party UIs and libraries - please see the list at the top of this README.111 112## Explanation of quantisation methods113 114<details>115  <summary>Click to see details</summary>116 117The new methods available are:118 119* 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)120* 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.121* 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.122* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw123* 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 bpw124 125Refer to the Provided Files table below to see what files use which methods, and how.126</details>127<!-- compatibility_gguf end -->128 129<!-- README_GGUF.md-provided-files start -->130## Provided files131 132| Name | Quant method | Bits | Size | Max RAM required | Use case |133| ---- | ---- | ---- | ---- | ---- | ----- |134| [python-code-33b.Q2_K.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q2_K.gguf) | Q2_K | 2 | 13.50 GB| 16.00 GB | smallest, significant quality loss - not recommended for most purposes |135| [python-code-33b.Q3_K_S.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q3_K_S.gguf) | Q3_K_S | 3 | 14.06 GB| 16.56 GB | very small, high quality loss |136| [python-code-33b.Q3_K_M.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q3_K_M.gguf) | Q3_K_M | 3 | 15.76 GB| 18.26 GB | very small, high quality loss |137| [python-code-33b.Q3_K_L.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q3_K_L.gguf) | Q3_K_L | 3 | 17.28 GB| 19.78 GB | small, substantial quality loss |138| [python-code-33b.Q4_0.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q4_0.gguf) | Q4_0 | 4 | 18.36 GB| 20.86 GB | legacy; small, very high quality loss - prefer using Q3_K_M |139| [python-code-33b.Q4_K_S.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q4_K_S.gguf) | Q4_K_S | 4 | 18.44 GB| 20.94 GB | small, greater quality loss |140| [python-code-33b.Q4_K_M.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q4_K_M.gguf) | Q4_K_M | 4 | 19.62 GB| 22.12 GB | medium, balanced quality - recommended |141| [python-code-33b.Q5_0.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q5_0.gguf) | Q5_0 | 5 | 22.40 GB| 24.90 GB | legacy; medium, balanced quality - prefer using Q4_K_M |142| [python-code-33b.Q5_K_S.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q5_K_S.gguf) | Q5_K_S | 5 | 22.40 GB| 24.90 GB | large, low quality loss - recommended |143| [python-code-33b.Q5_K_M.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q5_K_M.gguf) | Q5_K_M | 5 | 23.05 GB| 25.55 GB | large, very low quality loss - recommended |144| [python-code-33b.Q6_K.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q6_K.gguf) | Q6_K | 6 | 26.69 GB| 29.19 GB | very large, extremely low quality loss |145| [python-code-33b.Q8_0.gguf](https://huggingface.co/TheBloke/Python-Code-33B-GGUF/blob/main/python-code-33b.Q8_0.gguf) | Q8_0 | 8 | 34.57 GB| 37.07 GB | very large, extremely low quality loss - not recommended |146 147**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.148 149 150 151<!-- README_GGUF.md-provided-files end -->152 153<!-- README_GGUF.md-how-to-download start -->154## How to download GGUF files155 156**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.157 158The following clients/libraries will automatically download models for you, providing a list of available models to choose from:159 160* LM Studio161* LoLLMS Web UI162* Faraday.dev163 164### In `text-generation-webui`165 166Under Download Model, you can enter the model repo: TheBloke/Python-Code-33B-GGUF and below it, a specific filename to download, such as: python-code-33b.Q4_K_M.gguf.167 168Then click Download.169 170### On the command line, including multiple files at once171 172I recommend using the `huggingface-hub` Python library:173 174```shell175pip3 install huggingface-hub176```177 178Then you can download any individual model file to the current directory, at high speed, with a command like this:179 180```shell181huggingface-cli download TheBloke/Python-Code-33B-GGUF python-code-33b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False182```183 184<details>185  <summary>More advanced huggingface-cli download usage</summary>186 187You can also download multiple files at once with a pattern:188 189```shell190huggingface-cli download TheBloke/Python-Code-33B-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'191```192 193For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).194 195To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:196 197```shell198pip3 install hf_transfer199```200 201And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:202 203```shell204HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Python-Code-33B-GGUF python-code-33b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False205```206 207Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.208</details>209<!-- README_GGUF.md-how-to-download end -->210 211<!-- README_GGUF.md-how-to-run start -->212## Example `llama.cpp` command213 214Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.215 216```shell217./main -ngl 32 -m python-code-33b.Q4_K_M.gguf --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "This is a conversation with your helpful AI assistant. AI assistant can generate Python Code along with necessary explanation.\n\nContext\nYou are a helpful AI assistant.\n\nUSER: {prompt}\nASSISTANT:"218```219 220Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.221 222Change `-c 2048` 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.223 224If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`225 226For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)227 228## How to run in `text-generation-webui`229 230Further instructions can be found in the text-generation-webui documentation, here: [text-generation-webui/docs/04 ‐ Model Tab.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/04%20%E2%80%90%20Model%20Tab.md#llamacpp).231 232## How to run from Python code233 234You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.235 236### How to load this model in Python code, using ctransformers237 238#### First install the package239 240Run one of the following commands, according to your system:241 242```shell243# Base ctransformers with no GPU acceleration244pip install ctransformers245# Or with CUDA GPU acceleration246pip install ctransformers[cuda]247# Or with AMD ROCm GPU acceleration (Linux only)248CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers249# Or with Metal GPU acceleration for macOS systems only250CT_METAL=1 pip install ctransformers --no-binary ctransformers251```252 253#### Simple ctransformers example code254 255```python256from ctransformers import AutoModelForCausalLM257 258# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.259llm = AutoModelForCausalLM.from_pretrained("TheBloke/Python-Code-33B-GGUF", model_file="python-code-33b.Q4_K_M.gguf", model_type="llama", gpu_layers=50)260 261print(llm("AI is going to"))262```263 264## How to use with LangChain265 266Here are guides on using llama-cpp-python and ctransformers with LangChain:267 268* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)269* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)270 271<!-- README_GGUF.md-how-to-run end -->272 273<!-- footer start -->274<!-- 200823 -->275## Discord276 277For further support, and discussions on these models and AI in general, join us at:278 279[TheBloke AI's Discord server](https://discord.gg/theblokeai)280 281## Thanks, and how to contribute282 283Thanks to the [chirper.ai](https://chirper.ai) team!284 285Thanks to Clay from [gpus.llm-utils.org](llm-utils)!286 287I'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.288 289If 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.290 291Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.292 293* Patreon: https://patreon.com/TheBlokeAI294* Ko-Fi: https://ko-fi.com/TheBlokeAI295 296**Special thanks to**: Aemon Algiz.297 298**Patreon special mentions**: Brandon Frisco, LangChain4j, Spiking Neurons AB, transmissions 11, Joseph William Delisle, Nitin Borwankar, Willem Michiel, Michael Dempsey, vamX, Jeffrey Morgan, zynix, jjj, Omer Bin Jawed, Sean Connelly, jinyuan sun, Jeromy Smith, Shadi, Pawan Osman, Chadd, Elijah Stavena, Illia Dulskyi, Sebastain Graf, Stephen Murray, terasurfer, Edmond Seymore, Celu Ramasamy, Mandus, Alex, biorpg, Ajan Kanaga, Clay Pascal, Raven Klaugh, 阿明, K, ya boyyy, usrbinkat, Alicia Loh, John Villwock, ReadyPlayerEmma, Chris Smitley, Cap'n Zoog, fincy, GodLy, S_X, sidney chen, Cory Kujawski, OG, Mano Prime, AzureBlack, Pieter, Kalila, Spencer Kim, Tom X Nguyen, Stanislav Ovsiannikov, Michael Levine, Andrey, Trailburnt, Vadim, Enrico Ros, Talal Aujan, Brandon Phillips, Jack West, Eugene Pentland, Michael Davis, Will Dee, webtim, Jonathan Leane, Alps Aficionado, Rooh Singh, Tiffany J. Kim, theTransient, Luke @flexchar, Elle, Caitlyn Gatomon, Ari Malik, subjectnull, Johann-Peter Hartmann, Trenton Dambrowitz, Imad Khwaja, Asp the Wyvern, Emad Mostaque, Rainer Wilmers, Alexandros Triantafyllidis, Nicholas, Pedro Madruga, SuperWojo, Harry Royden McLaughlin, James Bentley, Olakabola, David Ziegler, Ai Maven, Jeff Scroggin, Nikolai Manek, Deo Leter, Matthew Berman, Fen Risland, Ken Nordquist, Manuel Alberto Morcote, Luke Pendergrass, TL, Fred von Graf, Randy H, Dan Guido, NimbleBox.ai, Vitor Caleffi, Gabriel Tamborski, knownsqashed, Lone Striker, Erik Bjäreholt, John Detwiler, Leonard Tan, Iucharbius299 300 301Thank you to all my generous patrons and donaters!302 303And thank you again to a16z for their generous grant.304 305<!-- footer end -->306 307<!-- original-model-card start -->308# Original model card: Feynman Innovations's Python Code 33B309 310 311**Python-Code-33B**312 313Large Language Models (LLMs) are good with code generations. Sometimes LLMs do make mistakes in code generation. How about if they can give detailed explanation along with the code.314This is what I have tried over here. The base Llama-2 model was used for training purpose. It is trained on around 23000+ set of codes. Each set having 2 conversations.315This data was generated using GPT-3.5, GPT-4 etc. This conversation is in Vicuna/ShareGPT format. Each set, along with code, has detailed explanation.316I have released the [data](https://huggingface.co/datasets/ajibawa-2023/Python-Code-23k-ShareGPT).317 318**Training:**319Entire dataset was trained on Azure 4 x A100 80GB. For 3 epoch, training took 42 hours. DeepSpeed codebase was used for training purpose. This was trained on Llama-1 by Meta.320 321 322**GPTQ GGML & AWQ**323 324GPTQ: TBA325 326GGUF: TBA327 328AWQ: TBA329 330 331**Example Prompt:**332```333This is a conversation with your helpful AI assistant. AI assistant can generate Python Code along with necessary explanation.334 335Context336You are a helpful AI assistant.337 338USER: <prompt>339ASSISTANT:340```341 342<!-- original-model-card end -->343