---
license: apache-2.0
datasets:
- Azure99/blossom-chat-v3
- Azure99/blossom-math-v4
- Azure99/blossom-wizard-v3
- Azure99/blossom-orca-v3
language:
- zh
- en
quantized_by: bartowski
pipeline_tag: text-generation
---

## Exllama v2 Quantizations of blossom-v5-14b

Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.15">turboderp's ExLlamaV2 v0.0.15</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/Azure99/blossom-v5-14b

No GQA - VRAM requirements will be higher

| Branch | Bits | lm_head bits | VRAM (4k) | VRAM (16k) | Description |
| ----- | ---- | ------- | ------ | ------ | ------------ |
| [8_0](https://huggingface.co/bartowski/blossom-v5-14b-exl2/tree/8_0) | 8.0  | 8.0 | 18.5 GB | 28.8 GB | Near unquantized performance at vastly reduced size, **recommended**. |
| [6_5](https://huggingface.co/bartowski/blossom-v5-14b-exl2/tree/6_5) | 6.5  | 8.0 | 15.9 GB | 26.4 GB | Near unquantized performance at vastly reduced size, **recommended**. |
| [5_0](https://huggingface.co/bartowski/blossom-v5-14b-exl2/tree/5_0) | 5.0  | 6.0 | 13.3 GB | 23.9 GB | Slightly lower perplexity vs 6.5. |
| [4_25](https://huggingface.co/bartowski/blossom-v5-14b-exl2/tree/4_25) | 4.25 | 6.0 | 12.0 GB | 22.6 GB | GPTQ equivalent bits per weight, can fit in 12 GB card with lower context. |
| [3_75](https://huggingface.co/bartowski/blossom-v5-14b-exl2/tree/3_75) | 3.75  | 6.0 | 11.1 GB | 21.7 GB | Lower quality but still generally usable. |
| [3_0](https://huggingface.co/bartowski/blossom-v5-14b-exl2/tree/3_0) | 3.0  | 6.0 |  10.0 GB | 20.6 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/blossom-v5-14b-exl2 blossom-v5-14b-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 `blossom-v5-14b-exl2`:

```shell
mkdir blossom-v5-14b-exl2
huggingface-cli download bartowski/blossom-v5-14b-exl2 --local-dir blossom-v5-14b-exl2 --local-dir-use-symlinks False
```

To download from a different branch, add the `--revision` parameter:

Linux:

```shell
mkdir blossom-v5-14b-exl2-6_5
huggingface-cli download bartowski/blossom-v5-14b-exl2 --revision 6_5 --local-dir blossom-v5-14b-exl2-6_5 --local-dir-use-symlinks False
```

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

```shell
mkdir blossom-v5-14b-exl2-6.5
huggingface-cli download bartowski/blossom-v5-14b-exl2 --revision 6_5 --local-dir blossom-v5-14b-exl2-6.5 --local-dir-use-symlinks False
```

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