Large Language Models
Mike Slinn

PrivateGPT: Secure Q/A for a Corpus

Published 2024-02-12. Last modified 2024-02-14.
Time to read: 3 minutes.

This page is part of the llm collection.
PrivateGPT is a production-ready AI project that allows you to ask questions about your documents using the power of Large Language Models (LLMs), even in scenarios without an Internet connection. 100% private, no data leaves your execution environment at any point.

The project provides an API offering all the primitives required to build private, context-aware AI applications. It follows and extends the OpenAI API standard, and supports both normal and streaming responses.

The documentation is well written.

PrivateGPT relies on LlamaIndex.

Installation

The Quickstart worked on WSL/Ubuntu and macOS.

While privateGPT is computing an answer, you can check to see if the GPU is doing any of the heavy lifting, or if the CPU is doing the best it can to do all the work.

MacBook Pro GPU

On macOS you can see how hard the GPU in your Mac has been working by using the Activity Monitor app, then choosing Window / GPU History.

The image to the right shows the realtime workload for the built-in GPU in a MacBook Pro while privateGPT was computing an answer.

WSL/Ubuntu GPU

I previously wrote about getting Ubuntu / WSL working with NVIDIA CUDA. It is a mess. Best to use native Ubuntu if NVIDIA drivers are required.

Poetry and Python Virtual Environments

Python virtual environments are a good way to isolate Python programs from each other, so their dependencies do not clash. I wrote about Python virtual environments previously, and that article also discusses Poetry. The following assumes you read that material.

PrivateGPT is configured to use Poetry to manage the libraries within the venv.

Installation Commands

This section shows how I installed Private GPT v0.2.0 on WSL/Ubuntu.

First, attempt to deactivate any pre-existing venv. Do not worry if the error deactivate: command not found appears; that just means there was no active venv. It is essential that no pre-existing venv be active when attempting to run any poetry subcommand.

Shell
$ deactivate # 'deactivate: command not found' is OK :) 

Clone the privateGPT Git project:

Shell
$ git clone https://github.com/imartinez/privateGPT
$ cd privateGPT

The poetry install command creates a new venv containing the dependencies specified by the project’s poetry.lock.

Shell
$ poetry install --with ui,local
Creating virtualenv private-gpt in /mnt/f/work/evelyn/privateGPT/.venv
Installing dependencies from lock file
Package operations: 148 installs, 1 update, 0 removals
• Installing frozenlist (1.4.1) • Installing idna (3.6) • Installing multidict (6.0.4) • Installing aiosignal (1.3.1) • Installing attrs (23.1.0) • Installing certifi (2023.11.17) • Installing charset-normalizer (3.3.2) • Installing nvidia-nvjitlink-cu12 (12.3.101) • Installing six (1.16.0) • Installing urllib3 (1.26.18) • Installing wrapt (1.16.0) • Installing yarl (1.9.4) • Installing aiohttp (3.9.1) • Installing aiohttp (3.9.1) • Installing deprecated (1.2.14) • Installing dill (0.3.7) • Installing filelock (3.13.1) • Installing fsspec (2023.12.2) • Installing markupsafe (2.1.3) • Installing mpmath (1.3.0) • Installing numpy (1.26.0) • Installing nvidia-cublas-cu12 (12.1.3.1) • Installing nvidia-cusparse-cu12 (12.1.0.106) • Installing packaging (23.2) • Installing python-dateutil (2.8.2) • Installing pytz (2023.3.post1) • Installing pyyaml (6.0.1) • Installing requests (2.31.0) • Installing tqdm (4.66.1) • Installing typing-extensions (4.9.0) • Installing tzdata (2023.3) • Installing h11 (0.14.0) • Installing huggingface-hub (0.19.4) • Installing humanfriendly (10.0) • Installing jinja2 (3.1.2) • Installing multiprocess (0.70.15) • Installing networkx (3.2.1) • Installing nvidia-cuda-cupti-cu12 (12.1.105) • Installing nvidia-cuda-nvrtc-cu12 (12.1.105) • Installing h11 (0.14.0) • Installing huggingface-hub (0.19.4) • Installing humanfriendly (10.0) • Installing jinja2 (3.1.2) • Installing multiprocess (0.70.15) • Installing networkx (3.2.1) • Installing nvidia-cuda-cupti-cu12 (12.1.105) • Installing nvidia-cuda-nvrtc-cu12 (12.1.105) • Installing nvidia-cuda-runtime-cu12 (12.1.105) • Installing nvidia-cudnn-cu12 (8.9.2.26) • Installing nvidia-cufft-cu12 (11.0.2.54) • Installing nvidia-curand-cu12 (10.3.2.106) • Installing nvidia-cusolver-cu12 (11.4.5.107) • Installing nvidia-nccl-cu12 (2.18.1) • Installing nvidia-nvtx-cu12 (12.1.105) • Installing pandas (2.1.4) • Installing protobuf (4.25.1) • Installing pyarrow (14.0.1) • Updating setuptools (68.1.2 -> 69.0.2) • Installing sniffio (1.3.0) • Installing sympy (1.12) • Installing triton (2.1.0) • Installing xxhash (3.4.1) • Installing annotated-types (0.6.0) • Installing anyio (3.7.1) • Installing coloredlogs (15.0.1) • Installing annotated-types (0.6.0) • Installing anyio (3.7.1) • Installing coloredlogs (15.0.1) • Installing datasets (2.14.4) • Installing flatbuffers (23.5.26) • Installing hpack (4.0.0) • Installing httpcore (1.0.2) • Installing hyperframe (6.0.1) • Installing jmespath (1.0.1) • Installing mypy-extensions (1.0.0) • Installing psutil (5.9.6) • Installing pydantic-core (2.14.5) • Installing regex (2023.10.3) • Installing responses (0.18.0) • Installing safetensors (0.4.1) • Installing sentencepiece (0.1.99) • Installing tokenizers (0.15.0) • Installing torch (2.1.2): • Installing accelerate (0.25.0) • Installing botocore (1.34.2) • Installing distro (1.8.0) • Installing distlib (0.3.8) • Installing click (8.1.7) • Installing dnspython (2.4.2) • Installing evaluate (0.4.1). • Installing botocore (1.34.2) • Installing accelerate (0.25.0) • Installing botocore (1.34.2) • Installing botocore (1.34.2) • Installing botocore (1.34.2) • Installing distro (1.8.0) • Installing distlib (0.3.8) • Installing click (8.1.7) • Installing dnspython (2.4.2) • Installing evaluate (0.4.1) • Installing greenlet (3.0.2) • Installing grpcio (1.60.0) • Installing h2 (4.1.0) • Installing httptools (0.6.1) • Installing httpx (0.25.2) • Installing iniconfig (2.0.0) • Installing joblib (1.3.2) • Installing marshmallow (3.20.1) • Installing onnx (1.15.0) • Installing onnxruntime (1.16.3) • Installing pillow (10.1.0) • Installing platformdirs (4.1.0) • Installing pluggy (1.3.0) • Installing pydantic (2.5.2) • Installing python-dotenv (1.0.0) • Installing scipy (1.11.4) • Installing soupsieve (2.5) • Installing starlette (0.27.0) • Installing threadpoolctl (3.2.0) • Installing transformers (4.36.1) • Installing typing-inspect (0.9.0) • Installing uvloop (0.19.0) • Installing watchfiles (0.21.0) • Installing websockets (11.0.3) • Installing aiostream (0.5.2) • Installing beautifulsoup4 (4.12.2) • Installing cfgv (3.4.0) • Installing coverage (7.3.3) • Installing dataclasses-json (0.5.14) • Installing diskcache (5.6.3) • Installing email-validator (2.1.0.post1) • Installing fastapi (0.103.2) • Installing grpcio-tools (1.60.0) • Installing identify (2.5.33) • Installing itsdangerous (2.1.2) • Installing nest-asyncio (1.5.8) • Installing cfgv (3.4.0) • Installing beautifulsoup4 (4.12.2) • Installing cfgv (3.4.0) • Installing aiostream (0.5.2) • Installing beautifulsoup4 (4.12.2) • Installing cfgv (3.4.0) • Installing cfgv (3.4.0) • Installing beautifulsoup4 (4.12.2) • Installing cfgv (3.4.0) • Installing coverage (7.3.3) • Installing dataclasses-json (0.5.14) • Installing diskcache (5.6.3) • Installing email-validator (2.1.0.post1) • Installing fastapi (0.103.2) • Installing grpcio-tools (1.60.0) • Installing identify (2.5.33) • Installing itsdangerous (2.1.2) • Installing nest-asyncio (1.5.8) • Installing nltk (3.8.1) • Installing nodeenv (1.8.0) • Installing openai (1.5.0) • Installing optimum (1.16.1) • Installing orjson (3.9.10) • Installing pathspec (0.12.1) • Installing portalocker (2.8.2) • Installing pydantic-extra-types (2.2.0) • Installing pydantic-settings (2.1.0) • Installing pytest (7.4.3) • Installing python-multipart (0.0.6) • Installing s3transfer (0.9.0) • Installing scikit-learn (1.3.2) • Installing sqlalchemy (2.0.23) • Installing tenacity (8.2.3) • Installing tiktoken (0.5.2) • Installing torchvision (0.16.2) • Installing ujson (5.9.0) • Installing uvicorn (0.24.0.post1) • Installing virtualenv (20.25.0) • Installing black (22.12.0) • Installing boto3 (1.34.2) • Installing injector (0.21.0) • Installing llama-cpp-python (0.2.23) • Installing black (22.12.0) • Installing boto3 (1.34.2) • Installing injector (0.21.0) • Installing llama-cpp-python (0.2.23) • Installing llama-index (0.9.3) • Installing mypy (1.7.1) • Installing pre-commit (2.21.0) • Installing pypdf (3.17.2) • Installing pytest-asyncio (0.21.1) • Installing pytest-cov (3.0.0) • Installing qdrant-client (1.7.0) • Installing ruff (0.1.8) • Installing sentence-transformers (2.2.2) • Installing types-pyyaml (6.0.12.12) • Installing watchdog (3.0.0)
Installing the current project: private-gpt (0.2.0)

Now download the embedding model and tokenizer. The first line of scripts/setup is the shebang for running Python:

scripts/setup shebang
#!/usr/bin/env python3

This means we do not need to specify python3 on the command line in order to run the setup script.

Shell
$ poetry run scripts/setup
13:40:58.082 [INFO    ] private_gpt.settings.settings_loader - Starting application with profiles=['default']
Downloading embedding BAAI/bge-small-en-v1.5
Fetching 13 files: 100%|███████████████████████████████████████████████████| 13/13 [00:00<00:00, 67.85it/s]
Embedding model downloaded!
Downloading LLM mistral-7b-instruct-v0.2.Q4_K_M.gguf
LLM model downloaded!
Downloading tokenizer mistralai/Mistral-7B-Instruct-v0.2
Tokenizer downloaded!
Setup done 

The default model is downloaded from HuggingFace, TheBloke/Mistral-7B-Instruct-v0.2-GGUF (4.37 GB). The default model and maximum number of tokens is defined in settings.yaml, which at this point looks like:

settings.yaml fragment
llm:
  mode: local
  # Should be matching the selected model
  max_new_tokens: 512
  context_window: 3900
  tokenizer: mistralai/Mistral-7B-Instruct-v0.2

While it is possible to edit settings.yaml, it is better to edit the overlay, settings-local.yaml, then set the PGPT_PROFILES environment variable to use the overlaid settings.

The privateGPT docs also talk about a file called settings-cuda.yaml, which I imagine would contain configuration specific to NVIDIA GPUs. However, that file was not provided, and can be ignored. I found that privateGPT used my NVIDIA RTX 3060 video card automagically.

Command-Line Subshell

If you want to debug PrivateGPT from the command-line, you could launch a Poetry subshell, so the dependencies managed by Poetry are available. Use the poetry shell command for this.

Shell
$ poetry shell
Creating virtualenv private-gpt in /mnt/f/work/evelyn/privateGPT/.venv
  Spawning shell within /mnt/f/work/privateGPT/.venv
  . /mnt/f/work/privateGPT/.venv/bin/activate
  . /mnt/f/work/privateGPT/.venv/bin/activate 
$ irb irb(main):001> Ctrl-D
$ Ctrl-D

Of course, you could also use poetry run for this purpose:-

Shell
$ poetry run irb
irb(main):001> 

Running PrivateGPT

Finally I launched the privateGPT API server and the gradio UI using the overlaid settings:

Shell
$ PGPT_PROFILES=local poetry run python -m private_gpt
11:15:00.269 [INFO    ] private_gpt.settings.settings_loader - Starting application with profiles=['default']
11:15:20.216 [INFO    ] private_gpt.components.llm.llm_component - Initializing the LLM in mode=local
llama_model_loader: loaded meta data with 24 key-value pairs and 291 tensors from /mnt/f/work/evelyn/privateGPT/models/mistral-7b-instruct-v0.2.Q4_K_M.gguf (version GGUF V3 (latest))
llama_model_loader: - tensor    0:                token_embd.weight q4_K     [  4096, 32000,     1,     1 ]
llama_model_loader: - tensor    1:              blk.0.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor    2:              blk.0.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor    3:              blk.0.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor    4:         blk.0.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor    5:            blk.0.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor    6:              blk.0.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor    7:            blk.0.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor    8:           blk.0.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor    9:            blk.0.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   10:              blk.1.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   11:              blk.1.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   12:              blk.1.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   13:         blk.1.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   14:            blk.1.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   15:              blk.1.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   16:            blk.1.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   17:           blk.1.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   18:            blk.1.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   19:              blk.2.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   20:              blk.2.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   21:              blk.2.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   22:         blk.2.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   23:            blk.2.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   24:              blk.2.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   25:            blk.2.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   26:           blk.2.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   27:            blk.2.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   28:              blk.3.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   29:              blk.3.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   30:              blk.3.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   31:         blk.3.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   32:            blk.3.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   33:              blk.3.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   34:            blk.3.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   35:           blk.3.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   36:            blk.3.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   37:              blk.4.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   38:              blk.4.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   39:              blk.4.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   40:         blk.4.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   41:            blk.4.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   42:              blk.4.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   43:            blk.4.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   44:           blk.4.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   45:            blk.4.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   46:              blk.5.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   47:              blk.5.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   48:              blk.5.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   49:         blk.5.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   50:            blk.5.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   51:              blk.5.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   52:            blk.5.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   53:           blk.5.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   54:            blk.5.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   55:              blk.6.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   56:              blk.6.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   57:              blk.6.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   58:         blk.6.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   59:            blk.6.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   60:              blk.6.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   61:            blk.6.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   62:           blk.6.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   63:            blk.6.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   64:              blk.7.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   65:              blk.7.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   66:              blk.7.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   67:         blk.7.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   68:            blk.7.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   69:              blk.7.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   70:            blk.7.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   71:           blk.7.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   72:            blk.7.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   73:              blk.8.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   74:              blk.8.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   75:              blk.8.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   76:         blk.8.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   77:            blk.8.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   78:              blk.8.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   79:            blk.8.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   80:           blk.8.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   81:            blk.8.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   82:              blk.9.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   83:              blk.9.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   84:              blk.9.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   85:         blk.9.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   86:            blk.9.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   87:              blk.9.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   88:            blk.9.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   89:           blk.9.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   90:            blk.9.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   91:             blk.10.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   92:             blk.10.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   93:             blk.10.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor   94:        blk.10.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor   95:           blk.10.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   96:             blk.10.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor   97:           blk.10.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor   98:          blk.10.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor   99:           blk.10.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  100:             blk.11.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  101:             blk.11.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  102:             blk.11.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  103:        blk.11.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  104:           blk.11.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  105:             blk.11.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  106:           blk.11.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  107:          blk.11.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  108:           blk.11.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  109:             blk.12.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  110:             blk.12.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  111:             blk.12.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  112:        blk.12.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  113:           blk.12.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  114:             blk.12.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  115:           blk.12.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  116:          blk.12.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  117:           blk.12.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  118:             blk.13.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  119:             blk.13.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  120:             blk.13.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  121:        blk.13.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  122:           blk.13.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  123:             blk.13.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  124:           blk.13.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  125:          blk.13.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  126:           blk.13.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  127:             blk.14.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  128:             blk.14.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  129:             blk.14.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  130:        blk.14.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  131:           blk.14.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  132:             blk.14.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  133:           blk.14.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  134:          blk.14.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  135:           blk.14.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  136:             blk.15.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  137:             blk.15.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  138:             blk.15.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  139:        blk.15.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  140:           blk.15.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  141:             blk.15.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  142:           blk.15.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  143:          blk.15.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  144:           blk.15.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  145:             blk.16.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  146:             blk.16.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  147:             blk.16.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  148:        blk.16.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  149:           blk.16.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  150:             blk.16.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  151:           blk.16.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  152:          blk.16.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  153:           blk.16.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  154:             blk.17.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  155:             blk.17.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  156:             blk.17.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  157:        blk.17.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  158:           blk.17.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  159:             blk.17.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  160:           blk.17.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  161:          blk.17.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  162:           blk.17.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  163:             blk.18.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  164:             blk.18.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  165:             blk.18.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  166:        blk.18.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  167:           blk.18.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  168:             blk.18.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  169:           blk.18.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  170:          blk.18.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  171:           blk.18.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  172:             blk.19.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  173:             blk.19.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  174:             blk.19.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  175:        blk.19.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  176:           blk.19.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  177:             blk.19.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  178:           blk.19.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  179:          blk.19.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  180:           blk.19.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  181:             blk.20.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  182:             blk.20.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  183:             blk.20.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  184:        blk.20.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  185:           blk.20.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  186:             blk.20.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  187:           blk.20.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  188:          blk.20.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  189:           blk.20.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  190:             blk.21.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  191:             blk.21.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  192:             blk.21.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  193:        blk.21.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  194:           blk.21.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  195:             blk.21.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  196:           blk.21.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  197:          blk.21.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  198:           blk.21.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  199:             blk.22.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  200:             blk.22.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  201:             blk.22.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  202:        blk.22.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  203:           blk.22.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  204:             blk.22.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  205:           blk.22.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  206:          blk.22.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  207:           blk.22.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  208:             blk.23.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  209:             blk.23.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  210:             blk.23.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  211:        blk.23.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  212:           blk.23.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  213:             blk.23.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  214:           blk.23.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  215:          blk.23.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  216:           blk.23.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  217:             blk.24.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  218:             blk.24.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  219:             blk.24.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  220:        blk.24.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  221:           blk.24.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  222:             blk.24.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  223:           blk.24.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  224:          blk.24.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  225:           blk.24.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  226:             blk.25.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  227:             blk.25.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  228:             blk.25.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  229:        blk.25.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  230:           blk.25.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  231:             blk.25.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  232:           blk.25.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  233:          blk.25.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  234:           blk.25.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  235:             blk.26.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  236:             blk.26.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  237:             blk.26.attn_v.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  238:        blk.26.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  239:           blk.26.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  240:             blk.26.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  241:           blk.26.ffn_down.weight q4_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  242:          blk.26.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  243:           blk.26.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  244:             blk.27.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  245:             blk.27.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  246:             blk.27.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  247:        blk.27.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  248:           blk.27.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  249:             blk.27.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  250:           blk.27.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  251:          blk.27.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  252:           blk.27.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  253:             blk.28.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  254:             blk.28.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  255:             blk.28.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  256:        blk.28.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  257:           blk.28.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  258:             blk.28.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  259:           blk.28.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  260:          blk.28.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  261:           blk.28.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  262:             blk.29.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  263:             blk.29.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  264:             blk.29.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  265:        blk.29.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  266:           blk.29.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  267:             blk.29.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  268:           blk.29.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  269:          blk.29.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  270:           blk.29.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  271:             blk.30.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  272:             blk.30.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  273:             blk.30.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  274:        blk.30.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  275:           blk.30.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  276:             blk.30.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  277:           blk.30.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  278:          blk.30.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  279:           blk.30.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  280:             blk.31.attn_q.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  281:             blk.31.attn_k.weight q4_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  282:             blk.31.attn_v.weight q6_K     [  4096,  1024,     1,     1 ]
llama_model_loader: - tensor  283:        blk.31.attn_output.weight q4_K     [  4096,  4096,     1,     1 ]
llama_model_loader: - tensor  284:           blk.31.ffn_gate.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  285:             blk.31.ffn_up.weight q4_K     [  4096, 14336,     1,     1 ]
llama_model_loader: - tensor  286:           blk.31.ffn_down.weight q6_K     [ 14336,  4096,     1,     1 ]
llama_model_loader: - tensor  287:          blk.31.attn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  288:           blk.31.ffn_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  289:               output_norm.weight f32      [  4096,     1,     1,     1 ]
llama_model_loader: - tensor  290:                    output.weight q6_K     [  4096, 32000,     1,     1 ]
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv   0:                       general.architecture str              = llama
llama_model_loader: - kv   1:                               general.name str              = mistralai_mistral-7b-instruct-v0.2
llama_model_loader: - kv   2:                       llama.context_length u32              = 32768
llama_model_loader: - kv   3:                     llama.embedding_length u32              = 4096
llama_model_loader: - kv   4:                          llama.block_count u32              = 32
llama_model_loader: - kv   5:                  llama.feed_forward_length u32              = 14336
llama_model_loader: - kv   6:                 llama.rope.dimension_count u32              = 128
llama_model_loader: - kv   7:                 llama.attention.head_count u32              = 32
llama_model_loader: - kv   8:              llama.attention.head_count_kv u32              = 8
llama_model_loader: - kv   9:     llama.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  10:                       llama.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  11:                          general.file_type u32              = 15
llama_model_loader: - kv  12:                       tokenizer.ggml.model str              = llama
llama_model_loader: - kv  13:                      tokenizer.ggml.tokens arr[str,32000]   = ["<unk>", "<s>", "</s>", "<0x00>", "<...
llama_model_loader: - kv  14:                      tokenizer.ggml.scores arr[f32,32000]   = [0.000000, 0.000000, 0.000000, 0.0000...
llama_model_loader: - kv  15:                  tokenizer.ggml.token_type arr[i32,32000]   = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ...
llama_model_loader: - kv  16:                tokenizer.ggml.bos_token_id u32              = 1
llama_model_loader: - kv  17:                tokenizer.ggml.eos_token_id u32              = 2
llama_model_loader: - kv  18:            tokenizer.ggml.unknown_token_id u32              = 0
llama_model_loader: - kv  19:            tokenizer.ggml.padding_token_id u32              = 0
llama_model_loader: - kv  20:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  21:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  22:                    tokenizer.chat_template str              = {{ bos_token }}{% for message in mess...
llama_model_loader: - kv  23:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:   65 tensors
llama_model_loader: - type q4_K:  193 tensors
llama_model_loader: - type q6_K:   33 tensors
llm_load_vocab: special tokens definition check successful ( 259/32000 ).
llm_load_print_meta: format           = GGUF V3 (latest)
llm_load_print_meta: arch             = llama
llm_load_print_meta: vocab type       = SPM
llm_load_print_meta: n_vocab          = 32000
llm_load_print_meta: n_merges         = 0
llm_load_print_meta: n_ctx_train      = 32768
llm_load_print_meta: n_embd           = 4096
llm_load_print_meta: n_head           = 32
llm_load_print_meta: n_head_kv        = 8
llm_load_print_meta: n_layer          = 32
llm_load_print_meta: n_rot            = 128
llm_load_print_meta: n_gqa            = 4
llm_load_print_meta: f_norm_eps       = 0.0e+00
llm_load_print_meta: f_norm_rms_eps   = 1.0e-05
llm_load_print_meta: f_clamp_kqv      = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: n_ff             = 14336
llm_load_print_meta: n_expert         = 0
llm_load_print_meta: n_expert_used    = 0
llm_load_print_meta: rope scaling     = linear
llm_load_print_meta: freq_base_train  = 1000000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_yarn_orig_ctx  = 32768
llm_load_print_meta: rope_finetuned   = unknown
llm_load_print_meta: model type       = 7B
llm_load_print_meta: model ftype      = mostly Q4_K - Medium
llm_load_print_meta: model params     = 7.24 B
llm_load_print_meta: model size       = 4.07 GiB (4.83 BPW)
llm_load_print_meta: general.name     = mistralai_mistral-7b-instruct-v0.2
llm_load_print_meta: BOS token        = 1 '<s>'
llm_load_print_meta: EOS token        = 2 '</s>'
llm_load_print_meta: UNK token        = 0 '<unk>'
llm_load_print_meta: PAD token        = 0 '<unk>'
llm_load_print_meta: LF token         = 13 '<0x0A>'
llm_load_tensors: ggml ctx size =    0.12 MiB
llm_load_tensors: mem required  = 4165.48 MiB 

I pointed my web browser at localhost:8001 and saw:

After uploading this PDF, I asked a question that the PDF had the answer to. Private GPT gave the the page number that it got the answer from. The reference was duplicated.

I uploaded another PDF, and asked the same question again.

The previous explanation and reference was ignored, and unfortunately only the new document was searched. Again, the reference was duplicated. I wonder if it is possible to suppress the introductory phrase "According to the context provided".

Quiet Launch Script

This bash script swallows most of the spurious messages that privateGPT spews all over the floor each time it runs.

#!/bin/bash

function help {
  echo "$(basename "$0") - Run privateGPT
  OPTIONS:
    -h display this help information
    -v show all messages

  If you want to monitor your GPU activity, run one of the following:
    nvtop
    watch -d gpustat
  "
  exit 1
}

cd "$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )" || exit

# Disable any activated virtual environment
# See https://mslinn.com/blog/2021/04/09/python-venvs.html#pvenv
VENV="$( dirname "$( which virtualenv)" )"
if [ ! "$VENV" == /usr/bin ]; then source "$VENV/activate" && deactivate 2>/; fi

# export PGPT_PROFILES=local,logger
export PGPT_PROFILES=local

unset HIDDEN VERBOSE
if [ "$1" == -v ]; then VERBOSE=true; shift; fi
if [ "$1" == -h ]; then help; fi

echo "privateGPT serving on http://localhost:8001 (Press CTRL+C to quit)"
if [ "$VERBOSE" ]; then
  poetry run python -m private_gpt
else
  export UVICORN_LOG_LEVEL=WARN
  # Set up filter for grep to suppress noisy stderr
  for X in 'Initializing the LLM in mode=local' \
           'Initializing the embedding model in mode=' \
           llama_index.indices.loading \
           'llama_new_context_with_model\:' \
           'llama_build_graph\:' \
           'llm_load_print_meta\:' \
           'llm_load_tensors\:' \
           'llm_load_vocab\:' \
           'llama_model_loader\:' \
           'Mounting the gradio UI, at path=' \
           uvicorn.access \
           uvicorn.error; do
    if [ "$HIDDEN" ]; then
      HIDDEN="$HIDDEN\|$X"
    else
      HIDDEN="$X"
    fi
  done
  # echo "$HIDDEN"
  SUPPRESS_WARNINGS=-Wignore
  poetry run python $SUPPRESS_WARNINGS -m private_gpt 2> >(grep -v "$HIDDEN")
fi

Changing the Apology

The PrivateGPT’s default apology for not knowing the answer to a question was constructed from static text, followed by a directory of documents in the corpus. My test corpus only had 2 PDFs, as shown:

I find it strange that a message stating no answer was to be had cited specific page numbers for each document in the corpus.

To change the prompt, edit settings.yaml and change:

settings.yaml
default_query_system_prompt: >
  You can only answer questions about the provided context.
  If you know the answer but it is not based in the provided context, don't provide
  the answer, just state the answer is not in the context provided.

Into this:

settings.yaml
default_query_system_prompt: >
  You can only answer questions about the provided context.
  If you know the answer but it is not based in the provided context, don't provide
  the answer, just state the following as the response: "Sorry, I do not know.".

Chunking

PrivateGPT splits documents in chunks (LlamaIndex calls them nodes). You can get a sense of the chunks by typing the following incantation; some sample output is shown. Page_label is the page number shown in the document.

Shell
$ curl localhost:8001/v1/ingest/list | jq
... {
  "object": "ingest.document",
  "doc_id": "d0d4aef0-c28d-48e0-ae92-ec16ec2c8806",
  "doc_metadata": {
    "page_label": "141",
    "file_name": "l11manual_en.pdf"
  }
},
{
  "object": "ingest.document",
  "doc_id": "6a76e36b-0e3f-453d-886b-bd070ddfe383",
  "doc_metadata": {
    "page_label": "27",
    "file_name": "l11manual_en.pdf"
  }
},
{
  "object": "ingest.document",
  "doc_id": "adc8d60f-9b4b-43ea-8fe3-b8a740e8058b",
  "doc_metadata": {
    "page_label": "59",
    "file_name": "l11manual_en.pdf"
  }
}... 

The above privateGPT API is documented here, and the List Ingested call is documented here.

Videos

References



* indicates a required field.

Please select the following to receive Mike Slinn’s newsletter:

You can unsubscribe at any time by clicking the link in the footer of emails.

Mike Slinn uses Mailchimp as his marketing platform. By clicking below to subscribe, you acknowledge that your information will be transferred to Mailchimp for processing. Learn more about Mailchimp’s privacy practices.