crumb_-_core1-base-464m-c4-exl2

Quantization made by Richard Erkhov.

Github

Discord

Request more models

core1-base-464m-c4 - EXL2

• Model creator: https://huggingface.co/crumb/
• Original model: https://huggingface.co/crumb/core1-base-464m-c4/

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/crumb_-_core1-base-464m-c4-exl2 core1-base-464m-c4-6_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 crumb_-_core1-base-464m-c4-exl2 --revision 6_5 --local-dir core1-base-464m-c4-6_5 --local-dir-use-symlinks False

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

huggingface-cli download crumb_-_core1-base-464m-c4-exl2 --revision 6_5 --local-dir core1-base-464m-c4-6.5 --local-dir-use-symlinks False

Original model description:

datasets:

• c4 language:
• en

---

CoreX models are Llama models in which the first X decoder layers are kept, and then the model is finetuned on 1 billion tokens from some dataset. Base model stems from Llama2-7b, medium from Llama2-13b, xl from Llama2-70b.