appvoid_-_arco-plus-exl2

Quantization made by Richard Erkhov.

Github

Discord

Request more models

arco-plus - EXL2

• Model creator: https://huggingface.co/appvoid/
• Original model: https://huggingface.co/appvoid/arco-plus/

Available sizes

| Branch | Bits | Description | | ----- | ---- | ------- | ------ | ------ | ------ | ------ | ------------ | | 8_0 | 8.0 | Maximum quality that ExLlamaV2 can produce, near unquantized performance. | | 6_5 | 6.5 | Very similar to 8.0, good tradeoff of size vs performance, recommended. | | 5_0 | 5.0 | Slightly lower quality vs 6.5, but usable on 8GB cards. | | 4_25 | 4.25 | GPTQ equivalent bits per weight, slightly higher quality. | | 3_5 | 3.5 | Lower quality, only use if you have to. |

Download instructions

With git:

git clone --single-branch --branch 6_5 https://huggingface.co/appvoid_-_arco-plus-exl2 arco-plus-6_5

With huggingface hub:

pip3 install huggingface-hub

To download a specific branch, use the --revision parameter. For example, to download the 6.5 bpw branch: Linux:

huggingface-cli download appvoid_-_arco-plus-exl2 --revision 6_5 --local-dir arco-plus-6_5 --local-dir-use-symlinks False

Windows (which apparently doesn't like _ in folders sometimes?):

huggingface-cli download appvoid_-_arco-plus-exl2 --revision 6_5 --local-dir arco-plus-6.5 --local-dir-use-symlinks False

Original model description:

base_model:

• appvoid/arco
• h2oai/h2o-danube3-500m-base library_name: transformers tags:
• mergekit
• merge

---

arco+

This is an untrained passthrough model based on arco and danube as a first effort to train a small enough reasoning language model that generalizes across all kind of reasoning tasks.

Benchmarks

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
    - model: appvoid/arco
      layer_range: [0, 14]
  - sources:
    - model: h2oai/h2o-danube3-500m-base
      layer_range: [4, 16]

merge_method: passthrough
dtype: float16