apodex_Apodex-1.1-mini-GGUF

Llamacpp imatrix Quantizations of Apodex-1.1-mini by apodex

Using llama.cpp release b10630 for quantization.

Original model: https://huggingface.co/apodex/Apodex-1.1-mini

Model details:

• Parameter count: 36B
• Input support: text, image (with mmproj file) - details
• Speculative decoding: yes (MTP) - details
• imatrix: yes - details

How to run

Prompt format

<|im_start|>system
{system_prompt}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
<think>

Don't know which to choose? Grab Q4_K_M (21.86GB) - usually a good mix of size and performance. Download instructions available here

Available files:

Download a specific file:

hf download bartowski/apodex_Apodex-1.1-mini-GGUF --include "apodex_Apodex-1.1-mini-Q4_K_M.gguf" --local-dir ./

Downloading using the Hugging Face CLI

First, make sure you have the Hugging Face CLI installed:

pip install -U "huggingface_hub[cli]"

Download a specific file:

hf download bartowski/apodex_Apodex-1.1-mini-GGUF --include "apodex_Apodex-1.1-mini-Q4_K_M.gguf" --local-dir ./

The files marked true in the Split column above are stored as multiple parts in a folder. To download all the parts to a local folder, run:

hf download bartowski/apodex_Apodex-1.1-mini-GGUF --include "apodex_Apodex-1.1-mini-bf16/*" --local-dir ./

You can either specify a new local-dir (apodex_Apodex-1.1-mini-bf16) or download them all in place (./)

How to run

These quants run with llama.cpp - installable in one line via llama.app:

curl -LsSf https://llama.app/install.sh | sh
llama-server -hf bartowski/apodex_Apodex-1.1-mini-GGUF:Q4_K_M

llama-server includes a built-in chat web UI, served at http://localhost:8080 by default.

These quants were made with llama.cpp release b10630 - if this model's architecture is newly supported, you'll need that release or newer to run them.

They also work in: LM Studio · koboldcpp · ramalama · Jan AI · Text Generation Web UI · LoLLMs · Atomic Chat

Multimodal

This model supports image input. Alongside the quants, this repo includes the multimodal projector files mmproj-apodex_Apodex-1.1-mini-f16.gguf and mmproj-apodex_Apodex-1.1-mini-bf16.gguf, which pair with any quant above.

llama.cpp downloads the mmproj automatically when using -hf as shown above; if you're loading files manually, pass it with --mmproj.

MTP

This model has MTP (Multi-Token Prediction) layers, and they are included in these quants

MTP layers act as a built-in draft model, letting llama.cpp run speculative decoding for faster generation. To use them, add the following flag to your llama.cpp command:

--spec-type draft-mtp

Note: the MTP layers are stored at Q4_0 in the imatrix quants (except for the Q8_0 quant), since imatrix calibration does not exercise them. Q4_0 is chosen for its speed which massively benefits MTP performance.

imatrix

All quants made using imatrix option, with a calibration corpus rendered through this model's own chat template. The corpus pairs plain prose with tool-calling and reasoning conversations (corpus source data), encoded exactly as this model sees them at inference and processed with --parse-special, so chat-format special tokens contribute to the importance matrix. The corpus rendered for this model is included in this repo: apodex_Apodex-1.1-mini-calibration-v6.txt. The imatrix is available here: apodex_Apodex-1.1-mini-imatrix.gguf.

{
  "generator": "auto_quant_v2 calibration renderer",
  "recipe": "calibration-v6",
  "model": "Apodex-1.1-mini",
  "encoder": "chat_template",
  "chunk_size": 512,
  "prose_chunks": 214,
  "tool_chunks": 377,
  "total_chunks": 591,
  "tool_chunk_fraction": 0.638,
  "n_conversations": 137,
  "extension_convs_used": 0,
  "conversation_token_lengths": [
    636,
    1699,
    1329,
    1591,
    1217,
    1436,
    3252,
    898,
    1299,
    1494,
    1139,
    2168,
    953,
    1325,
    2861,
    1345,
    1192,
    1058,
    809,
    790,
    1434,
    1130,
    1479,
    1274,
    1944,
    1570,
    1726,
    980,
    1487,
    1718,
    1641,
    1292,
    1340,
    1109,
    1092,
    1784,
    1723,
    1247,
    549,
    1982,
    1483,
    1198,
    1470,
    2074,
    2162,
    1387,
    1681,
    979,
    3028,
    1174,
    2940,
    859,
    1096,
    1042,
    979,
    770,
    2562,
    987,
    1220,
    1159,
    1303,
    1243,
    998,
    1291,
    1255,
    1660,
    985,
    1624,
    2198,
    913,
    397,
    1194,
    3407,
    2908,
    781,
    1014,
    1104,
    1145,
    1370,
    1154,
    1223,
    867,
    1282,
    1143,
    1363,
    1618,
    1475,
    2138,
    945,
    730,
    2838,
    714,
    1475,
    1747,
    2019,
    1275,
    715,
    1428,
    1200,
    1801,
    1916,
    1787,
    889,
    1087,
    1109,
    2928,
    812,
    806,
    835,
    1484,
    1124,
    1649,
    840,
    424,
    393,
    2698,
    1090,
    1211,
    1902,
    2111,
    2767,
    2749,
    888,
    1052,
    912,
    1031,
    1311,
    1071,
    914,
    1429,
    902,
    782,
    1815,
    1099,
    989,
    1388,
    1551
  ],
  "warnings": []
}

Embed/output weights

Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.

ARM/AVX information

llama.cpp automatically "repacks" weights into an interleaved layout at load time for faster inference on ARM and AVX machines - details in this PR. This once required downloading special Q4_0_4_4/4_8/8_8 files; those are long gone. Online repacking now covers Q4_0, IQ4_NL, and most K-quants, so no special quant choice is needed for CPU inference.

Which file should I choose?

An older (early 2024) but still useful write-up with charts comparing quant performances is provided by Artefact2 here

The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.

If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.

If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.

Hugging Face can also do this math for you: add your hardware in your Local Apps settings and the model page will show which files fit.

Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.

If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.

If you want to get more into the weeds, you can check out this extremely useful feature chart:

llama.cpp feature matrix

But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.

These I-quants can also be used on CPU, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.

Credits

Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset.

Thank you ZeroWw for the inspiration to experiment with embed/output.

Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski