legraphista's picture
Upload README.md with huggingface_hub
7f51ad9 verified
metadata
base_model: deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
inference: false
library_name: gguf
license: other
license_link: LICENSE
license_name: deepseek-license
pipeline_tag: text-generation
quantized_by: legraphista
tags:
  - quantized
  - GGUF
  - quantization
  - imat
  - imatrix
  - static
  - 16bit
  - 8bit
  - 6bit
  - 5bit
  - 4bit
  - 3bit
  - 2bit
  - 1bit

DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF

Llama.cpp imatrix quantization of deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct

Original Model: deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
Original dtype: BF16 (bfloat16)
Quantized by: llama.cpp b3166
IMatrix dataset: here


Files

IMatrix

Status: βœ… Available
Link: here

Common Quants

Filename Quant type File Size Status Uses IMatrix Is Split
DeepSeek-Coder-V2-Lite-Instruct.Q8_0.gguf Q8_0 16.70GB βœ… Available βšͺ Static πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q6_K.gguf Q6_K 14.07GB βœ… Available βšͺ Static πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q4_K.gguf Q4_K 10.36GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q3_K.gguf Q3_K 8.13GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q2_K.gguf Q2_K 6.43GB βœ… Available 🟒 IMatrix πŸ“¦ No

All Quants

Filename Quant type File Size Status Uses IMatrix Is Split
DeepSeek-Coder-V2-Lite-Instruct.BF16.gguf BF16 31.42GB βœ… Available βšͺ Static πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.FP16.gguf F16 31.42GB βœ… Available βšͺ Static πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q8_0.gguf Q8_0 16.70GB βœ… Available βšͺ Static πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q6_K.gguf Q6_K 14.07GB βœ… Available βšͺ Static πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q5_K.gguf Q5_K 11.85GB βœ… Available βšͺ Static πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q5_K_S.gguf Q5_K_S 11.14GB βœ… Available βšͺ Static πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q4_K.gguf Q4_K 10.36GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q4_K_S.gguf Q4_K_S 9.53GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ4_NL.gguf IQ4_NL 8.91GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ4_XS.gguf IQ4_XS 8.57GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q3_K.gguf Q3_K 8.13GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q3_K_L.gguf Q3_K_L 8.46GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q3_K_S.gguf Q3_K_S 7.49GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ3_M.gguf IQ3_M 7.55GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ3_S.gguf IQ3_S 7.49GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ3_XS.gguf IQ3_XS 7.12GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ3_XXS.gguf IQ3_XXS 6.96GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q2_K.gguf Q2_K 6.43GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.Q2_K_S.gguf Q2_K_S 6.46GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ2_M.gguf IQ2_M 6.33GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ2_S.gguf IQ2_S 6.01GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ2_XS.gguf IQ2_XS 5.97GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ2_XXS.gguf IQ2_XXS 5.64GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ1_M.gguf IQ1_M 5.24GB βœ… Available 🟒 IMatrix πŸ“¦ No
DeepSeek-Coder-V2-Lite-Instruct.IQ1_S.gguf IQ1_S 4.99GB βœ… Available 🟒 IMatrix πŸ“¦ No

Downloading using huggingface-cli

If you do not have hugginface-cli installed:

pip install -U "huggingface_hub[cli]"

Download the specific file you want:

huggingface-cli download legraphista/DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF --include "DeepSeek-Coder-V2-Lite-Instruct.Q8_0.gguf" --local-dir ./

If the model file is big, it has been split into multiple files. In order to download them all to a local folder, run:

huggingface-cli download legraphista/DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF --include "DeepSeek-Coder-V2-Lite-Instruct.Q8_0/*" --local-dir ./
# see FAQ for merging GGUF's

Inference

Simple chat template

<|begin▁of▁sentence|>User: {user_prompt}

Assistant: {assistant_response}<|end▁of▁sentence|>User: {next_user_prompt}

Chat template with system prompt

<|begin▁of▁sentence|>{system_prompt}

User: {user_prompt}

Assistant: {assistant_response}<|end▁of▁sentence|>User: {next_user_prompt}

Llama.cpp

llama.cpp/main -m DeepSeek-Coder-V2-Lite-Instruct.Q8_0.gguf --color -i -p "prompt here (according to the chat template)"

FAQ

Why is the IMatrix not applied everywhere?

According to this investigation, it appears that lower quantizations are the only ones that benefit from the imatrix input (as per hellaswag results).

How do I merge a split GGUF?

  1. Make sure you have gguf-split available
  2. Locate your GGUF chunks folder (ex: DeepSeek-Coder-V2-Lite-Instruct.Q8_0)
  3. Run gguf-split --merge DeepSeek-Coder-V2-Lite-Instruct.Q8_0/DeepSeek-Coder-V2-Lite-Instruct.Q8_0-00001-of-XXXXX.gguf DeepSeek-Coder-V2-Lite-Instruct.Q8_0.gguf
    • Make sure to point gguf-split to the first chunk of the split.

Got a suggestion? Ping me @legraphista!