| | --- |
| | pipeline_tag: text-generation |
| | tags: |
| | - labradorite |
| | - llama |
| | - llama-2 |
| | - ibm |
| | license: llama2 |
| | license_link: https://ai.meta.com/llama/license/ |
| | language: |
| | - en |
| | quantized_by: bartowski |
| | --- |
| | |
| | ## Exllama v2 Quantizations of labradorite-13b |
| |
|
| | Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.14">turboderp's ExLlamaV2 v0.0.14</a> for quantization. |
| |
|
| | <b>The "main" branch only contains the measurement.json, download one of the other branches for the model (see below)</b> |
| |
|
| | Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions. |
| |
|
| | Original model: https://huggingface.co/ibm/labradorite-13b |
| |
|
| | No GQA - VRAM requirements will be higher |
| |
|
| | | Branch | Bits | lm_head bits | VRAM (4k) | VRAM (16k) | Description | |
| | | ----- | ---- | ------- | ------ | ------ | ------------ | |
| | | [8_0](https://huggingface.co/bartowski/labradorite-13b-exl2/tree/8_0) | 8.0 | 8.0 | 16.8 GB | 26.2 GB | Near unquantized performance at vastly reduced size, **recommended**. | |
| | | [6_5](https://huggingface.co/bartowski/labradorite-13b-exl2/tree/6_5) | 6.5 | 8.0 | 14.4 GB | 24.0 GB | Near unquantized performance at vastly reduced size, **recommended**. | |
| | | [5_0](https://huggingface.co/bartowski/labradorite-13b-exl2/tree/5_0) | 5.0 | 6.0 | 12.1 GB | 21.7 GB | Slightly lower perplexity vs 6.5, can fit in 12 GB card with even lower context. | |
| | | [4_25](https://huggingface.co/bartowski/labradorite-13b-exl2/tree/4_25) | 4.25 | 6.0 | 10.9 GB | 20.5 GB | GPTQ equivalent bits per weight. | |
| | | [3_75](https://huggingface.co/bartowski/labradorite-13b-exl2/tree/3_75) | 3.75 | 6.0 | 10.1 GB | 19.7 GB | Lower quality but still generally usable. | |
| | | [3_0](https://huggingface.co/bartowski/labradorite-13b-exl2/tree/3_0) | 3.0 | 6.0 | 9.1 GB | 18.7 GB | Very low quality, not recommended unless you have to. | |
| | |
| | VRAM requirements listed for both 4k context and 16k context since without GQA the differences are massive (9.6 GB) |
| | |
| | ## Download instructions |
| | |
| | With git: |
| | |
| | ```shell |
| | git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/labradorite-13b-exl2 labradorite-13b-exl2-6_5 |
| | ``` |
| | |
| | With huggingface hub (credit to TheBloke for instructions): |
| | |
| | ```shell |
| | pip3 install huggingface-hub |
| | ``` |
| | |
| | To download the `main` (only useful if you only care about measurement.json) branch to a folder called `labradorite-13b-exl2`: |
| | |
| | ```shell |
| | mkdir labradorite-13b-exl2 |
| | huggingface-cli download bartowski/labradorite-13b-exl2 --local-dir labradorite-13b-exl2 --local-dir-use-symlinks False |
| | ``` |
| | |
| | To download from a different branch, add the `--revision` parameter: |
| | |
| | Linux: |
| | |
| | ```shell |
| | mkdir labradorite-13b-exl2-6_5 |
| | huggingface-cli download bartowski/labradorite-13b-exl2 --revision 6_5 --local-dir labradorite-13b-exl2-6_5 --local-dir-use-symlinks False |
| | ``` |
| | |
| | Windows (which apparently doesn't like _ in folders sometimes?): |
| | |
| | ```shell |
| | mkdir labradorite-13b-exl2-6.5 |
| | huggingface-cli download bartowski/labradorite-13b-exl2 --revision 6_5 --local-dir labradorite-13b-exl2-6.5 --local-dir-use-symlinks False |
| | ``` |
| | |
| | Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski |
