├── .gitignore ├── LICENSE ├── README.md ├── download_model.py ├── inference.py ├── media └── inference-demo.mp4 └── requirements.txt /.gitignore: -------------------------------------------------------------------------------- 1 | # Byte-compiled / optimized / DLL files 2 | __pycache__/ 3 | *.py[cod] 4 | *$py.class 5 | 6 | # C extensions 7 | *.so 8 | 9 | # Distribution / packaging 10 | .Python 11 | build/ 12 | develop-eggs/ 13 | dist/ 14 | downloads/ 15 | eggs/ 16 | .eggs/ 17 | lib/ 18 | lib64/ 19 | parts/ 20 | sdist/ 21 | var/ 22 | wheels/ 23 | share/python-wheels/ 24 | *.egg-info/ 25 | .installed.cfg 26 | *.egg 27 | MANIFEST 28 | 29 | # PyInstaller 30 | # Usually these files are written by a python script from a template 31 | # before PyInstaller builds the exe, so as to inject date/other infos into it. 32 | *.manifest 33 | *.spec 34 | 35 | # Installer logs 36 | pip-log.txt 37 | pip-delete-this-directory.txt 38 | 39 | # Unit test / coverage reports 40 | htmlcov/ 41 | .tox/ 42 | .nox/ 43 | .coverage 44 | .coverage.* 45 | .cache 46 | nosetests.xml 47 | coverage.xml 48 | *.cover 49 | *.py,cover 50 | .hypothesis/ 51 | .pytest_cache/ 52 | cover/ 53 | 54 | # Translations 55 | *.mo 56 | *.pot 57 | 58 | # Django stuff: 59 | *.log 60 | local_settings.py 61 | db.sqlite3 62 | db.sqlite3-journal 63 | 64 | # Flask stuff: 65 | instance/ 66 | .webassets-cache 67 | 68 | # Scrapy stuff: 69 | .scrapy 70 | 71 | # Sphinx documentation 72 | docs/_build/ 73 | 74 | # PyBuilder 75 | .pybuilder/ 76 | target/ 77 | 78 | # Jupyter Notebook 79 | .ipynb_checkpoints 80 | 81 | # IPython 82 | profile_default/ 83 | ipython_config.py 84 | 85 | # pyenv 86 | # For a library or package, you might want to ignore these files since the code is 87 | # intended to run in multiple environments; otherwise, check them in: 88 | # .python-version 89 | 90 | # pipenv 91 | # According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control. 92 | # However, in case of collaboration, if having platform-specific dependencies or dependencies 93 | # having no cross-platform support, pipenv may install dependencies that don't work, or not 94 | # install all needed dependencies. 95 | #Pipfile.lock 96 | 97 | # poetry 98 | # Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control. 99 | # This is especially recommended for binary packages to ensure reproducibility, and is more 100 | # commonly ignored for libraries. 101 | # https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control 102 | #poetry.lock 103 | 104 | # pdm 105 | # Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control. 106 | #pdm.lock 107 | # pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it 108 | # in version control. 109 | # https://pdm.fming.dev/#use-with-ide 110 | .pdm.toml 111 | 112 | # PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm 113 | __pypackages__/ 114 | 115 | # Celery stuff 116 | celerybeat-schedule 117 | celerybeat.pid 118 | 119 | # SageMath parsed files 120 | *.sage.py 121 | 122 | # Environments 123 | .env 124 | .venv 125 | env/ 126 | venv/ 127 | ENV/ 128 | env.bak/ 129 | venv.bak/ 130 | 131 | # Spyder project settings 132 | .spyderproject 133 | .spyproject 134 | 135 | # Rope project settings 136 | .ropeproject 137 | 138 | # mkdocs documentation 139 | /site 140 | 141 | # mypy 142 | .mypy_cache/ 143 | .dmypy.json 144 | dmypy.json 145 | 146 | # Pyre type checker 147 | .pyre/ 148 | 149 | # pytype static type analyzer 150 | .pytype/ 151 | 152 | # Cython debug symbols 153 | cython_debug/ 154 | 155 | # PyCharm 156 | # JetBrains specific template is maintained in a separate JetBrains.gitignore that can 157 | # be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore 158 | # and can be added to the global gitignore or merged into this file. For a more nuclear 159 | # option (not recommended) you can uncomment the following to ignore the entire idea folder. 160 | #.idea/ 161 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | MIT License 2 | 3 | Copyright (c) 2023 Anton Bacaj 4 | 5 | Permission is hereby granted, free of charge, to any person obtaining a copy 6 | of this software and associated documentation files (the "Software"), to deal 7 | in the Software without restriction, including without limitation the rights 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 9 | copies of the Software, and to permit persons to whom the Software is 10 | furnished to do so, subject to the following conditions: 11 | 12 | The above copyright notice and this permission notice shall be included in all 13 | copies or substantial portions of the Software. 14 | 15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 21 | SOFTWARE. 22 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # MPT 30B inference code using CPU 2 | 3 | Run inference on the latest MPT-30B model using your CPU. This inference code uses a [ggml](https://github.com/ggerganov/ggml) quantized model. To run the model we'll use a library called [ctransformers](https://github.com/marella/ctransformers) that has bindings to ggml in python. 4 | 5 | Turn style with history on latest commit: 6 | 7 | ![Inference Chat](https://user-images.githubusercontent.com/7272343/248859199-28a82f3d-ee54-44e4-b22d-ca348ac667e3.png) 8 | 9 | Video of initial demo: 10 | 11 | [Inference Demo](https://github.com/abacaj/mpt-30B-inference/assets/7272343/486fc9b1-8216-43cc-93c3-781677235502) 12 | 13 | ## Requirements 14 | 15 | I recommend you use docker for this model, it will make everything easier for you. Minimum specs system with 32GB of ram. Recommend to use `python 3.10`. 16 | 17 | ## Tested working on 18 | 19 | Will post some numbers for these two later. 20 | 21 | - AMD Epyc 7003 series CPU 22 | - AMD Ryzen 5950x CPU 23 | 24 | ## Setup 25 | 26 | First create a venv. 27 | 28 | ```sh 29 | python -m venv env && source env/bin/activate 30 | ``` 31 | 32 | Next install dependencies. 33 | 34 | ```sh 35 | pip install -r requirements.txt 36 | ``` 37 | 38 | Next download the quantized model weights (about 19GB). 39 | 40 | ```sh 41 | python download_model.py 42 | ``` 43 | 44 | Ready to rock, run inference. 45 | 46 | ```sh 47 | python inference.py 48 | ``` 49 | 50 | Next modify inference script prompt and generation parameters. 51 | -------------------------------------------------------------------------------- /download_model.py: -------------------------------------------------------------------------------- 1 | import os 2 | from huggingface_hub import hf_hub_download 3 | 4 | 5 | def download_mpt_quant(destination_folder: str, repo_id: str, model_filename: str): 6 | local_path = os.path.abspath(destination_folder) 7 | return hf_hub_download( 8 | repo_id=repo_id, 9 | filename=model_filename, 10 | local_dir=local_path, 11 | local_dir_use_symlinks=True 12 | ) 13 | 14 | 15 | if __name__ == "__main__": 16 | """full url: https://huggingface.co/TheBloke/mpt-30B-chat-GGML/blob/main/mpt-30b-chat.ggmlv0.q4_1.bin""" 17 | 18 | repo_id = "TheBloke/mpt-30B-chat-GGML" 19 | model_filename = "mpt-30b-chat.ggmlv0.q4_1.bin" 20 | destination_folder = "models" 21 | download_mpt_quant(destination_folder, repo_id, model_filename) 22 | -------------------------------------------------------------------------------- /inference.py: -------------------------------------------------------------------------------- 1 | import os 2 | from dataclasses import dataclass, asdict 3 | from ctransformers import AutoModelForCausalLM, AutoConfig 4 | 5 | 6 | @dataclass 7 | class GenerationConfig: 8 | temperature: float 9 | top_k: int 10 | top_p: float 11 | repetition_penalty: float 12 | max_new_tokens: int 13 | seed: int 14 | reset: bool 15 | stream: bool 16 | threads: int 17 | stop: list[str] 18 | 19 | 20 | def format_prompt(system_prompt: str, user_prompt: str): 21 | """format prompt based on: https://huggingface.co/spaces/mosaicml/mpt-30b-chat/blob/main/app.py""" 22 | 23 | system_prompt = f"<|im_start|>system\n{system_prompt}<|im_end|>\n" 24 | user_prompt = f"<|im_start|>user\n{user_prompt}<|im_end|>\n" 25 | assistant_prompt = f"<|im_start|>assistant\n" 26 | 27 | return f"{system_prompt}{user_prompt}{assistant_prompt}" 28 | 29 | 30 | def generate( 31 | llm: AutoModelForCausalLM, 32 | generation_config: GenerationConfig, 33 | system_prompt: str, 34 | user_prompt: str, 35 | ): 36 | """run model inference, will return a Generator if streaming is true""" 37 | 38 | return llm( 39 | format_prompt( 40 | system_prompt, 41 | user_prompt, 42 | ), 43 | **asdict(generation_config), 44 | ) 45 | 46 | 47 | if __name__ == "__main__": 48 | config = AutoConfig.from_pretrained("mosaicml/mpt-30b-chat", context_length=8192) 49 | llm = AutoModelForCausalLM.from_pretrained( 50 | os.path.abspath("models/mpt-30b-chat.ggmlv0.q4_1.bin"), 51 | model_type="mpt", 52 | config=config, 53 | ) 54 | 55 | system_prompt = "A conversation between a user and an LLM-based AI assistant named Local Assistant. Local Assistant gives helpful and honest answers." 56 | 57 | generation_config = GenerationConfig( 58 | temperature=0.2, 59 | top_k=0, 60 | top_p=0.9, 61 | repetition_penalty=1.0, 62 | max_new_tokens=512, # adjust as needed 63 | seed=42, 64 | reset=False, # reset history (cache) 65 | stream=True, # streaming per word/token 66 | threads=int(os.cpu_count() / 2), # adjust for your CPU 67 | stop=["<|im_end|>", "|<"], 68 | ) 69 | 70 | user_prefix = "[user]: " 71 | assistant_prefix = f"[assistant]:" 72 | 73 | while True: 74 | user_prompt = input(user_prefix) 75 | generator = generate(llm, generation_config, system_prompt, user_prompt.strip()) 76 | print(assistant_prefix, end=" ", flush=True) 77 | for word in generator: 78 | print(word, end="", flush=True) 79 | print("") 80 | -------------------------------------------------------------------------------- /media/inference-demo.mp4: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/abacaj/mpt-30B-inference/2e1ee1e6f2c18cdbcd5c7e6387af5d0a33771d6a/media/inference-demo.mp4 -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | ctransformers==0.2.10 2 | transformers==4.30.2 --------------------------------------------------------------------------------