metadata
license: other
license_name: deepseek
license_link: LICENSE
quantized_by: bartowski
pipeline_tag: text-generation
Exllama v2 Quantizations of deepseek-coder-7b-instruct-v1.5
Using turboderp's ExLlamaV2 v0.0.12 for quantization.
The "main" branch only contains the measurement.json, download one of the other branches for the model (see below)
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/deepseek-ai/deepseek-coder-7b-instruct-v1.5
No GQA - VRAM requirements will be higher
| Branch | Bits | lm_head bits | Size (4k) | Size (16k) | Description |
|---|---|---|---|---|---|
| 8_0 | 8.0 | 8.0 | 9.4 GB | 15.6 GB | Maximum quality that ExLlamaV2 can produce, near unquantized performance. |
| 6_5 | 6.5 | 8.0 | 8.6 GB | 14.8 GB | Near unquantized performance at vastly reduced size, recommended. |
| 5_0 | 5.0 | 6.0 | 7.2 GB | 13.4 GB | Slightly lower quality vs 6.5, but usable on 8GB cards with 4k context. |
| 4_25 | 4.25 | 6.0 | 6.5 GB | 12.7 GB | GPTQ equivalent bits per weight. |
| 3_5 | 3.5 | 6.0 | 5.9 GB | 12.1 GB | Lower quality, not recommended. |
Download instructions
With git:
git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/deepseek-coder-7b-instruct-v1.5-exl2 deepseek-coder-7b-instruct-v1.5-exl2-6_5
With huggingface hub (credit to TheBloke for instructions):
pip3 install huggingface-hub
To download the main (only useful if you only care about measurement.json) branch to a folder called deepseek-coder-7b-instruct-v1.5-exl2:
mkdir deepseek-coder-7b-instruct-v1.5-exl2
huggingface-cli download bartowski/deepseek-coder-7b-instruct-v1.5-exl2 --local-dir deepseek-coder-7b-instruct-v1.5-exl2 --local-dir-use-symlinks False
To download from a different branch, add the --revision parameter:
Linux:
mkdir deepseek-coder-7b-instruct-v1.5-exl2-6_5
huggingface-cli download bartowski/deepseek-coder-7b-instruct-v1.5-exl2 --revision 6_5 --local-dir deepseek-coder-7b-instruct-v1.5-exl2-6_5 --local-dir-use-symlinks False
Windows (which apparently doesn't like _ in folders sometimes?):
mkdir deepseek-coder-7b-instruct-v1.5-exl2-6.5
huggingface-cli download bartowski/deepseek-coder-7b-instruct-v1.5-exl2 --revision 6_5 --local-dir deepseek-coder-7b-instruct-v1.5-exl2-6.5 --local-dir-use-symlinks False
Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski