metadata

base_model: deepseek-ai/DeepSeek-V2-Lite
inference: false
library_name: gguf
pipeline_tag: text-generation
quantized_by: legraphista
tags:
  - quantized
  - GGUF
  - imatrix
  - quantization
  - imat
  - imatrix
  - static

DeepSeek-V2-Lite-IMat-GGUF

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

Original Model: deepseek-ai/DeepSeek-V2-Lite
Original dtype: BF16 (bfloat16)
Quantized by: llama.cpp https://github.com/ggerganov/llama.cpp/pull/7519
IMatrix dataset: here

DeepSeek-V2-Lite-IMat-GGUF

Files

IMatrix

Status: ✅ Available
Link: here

Common Quants

Filename	Quant type	File Size	Status	Uses IMatrix	Is Split
DeepSeek-V2-Lite.Q8_0.gguf	Q8_0	16.70GB	✅ Available	⚪ No	📦 No
DeepSeek-V2-Lite.Q6_K.gguf	Q6_K	14.07GB	✅ Available	⚪ No	📦 No
DeepSeek-V2-Lite.Q4_K.gguf	Q4_K	10.36GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.Q3_K.gguf	Q3_K	8.13GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.Q2_K.gguf	Q2_K	6.43GB	✅ Available	🟢 Yes	📦 No

All Quants

Filename	Quant type	File Size	Status	Uses IMatrix	Is Split
DeepSeek-V2-Lite.FP16.gguf	F16	31.42GB	✅ Available	⚪ No	📦 No
DeepSeek-V2-Lite.BF16.gguf	BF16	31.42GB	✅ Available	⚪ No	📦 No
DeepSeek-V2-Lite.Q5_K.gguf	Q5_K	11.85GB	✅ Available	⚪ No	📦 No
DeepSeek-V2-Lite.Q5_K_S.gguf	Q5_K_S	11.14GB	✅ Available	⚪ No	📦 No
DeepSeek-V2-Lite.Q4_K_S.gguf	Q4_K_S	9.53GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.Q3_K_L.gguf	Q3_K_L	8.46GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.Q3_K_S.gguf	Q3_K_S	7.49GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.Q2_K_S.gguf	Q2_K_S	6.46GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ4_NL.gguf	IQ4_NL	8.91GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ4_XS.gguf	IQ4_XS	8.57GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ3_M.gguf	IQ3_M	7.55GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ3_S.gguf	IQ3_S	7.49GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ3_XS.gguf	IQ3_XS	7.12GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ3_XXS.gguf	IQ3_XXS	6.96GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ2_M.gguf	IQ2_M	6.33GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ2_S.gguf	IQ2_S	6.01GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ2_XS.gguf	IQ2_XS	5.97GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ2_XXS.gguf	IQ2_XXS	5.64GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ1_M.gguf	IQ1_M	5.24GB	✅ Available	🟢 Yes	📦 No
DeepSeek-V2-Lite.IQ1_S.gguf	IQ1_S	4.99GB	✅ Available	🟢 Yes	📦 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-V2-Lite-IMat-GGUF --include "DeepSeek-V2-Lite.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-V2-Lite-IMat-GGUF --include "DeepSeek-V2-Lite.Q8_0/*" --local-dir DeepSeek-V2-Lite.Q8_0
# see FAQ for merging GGUF's

Inference

Llama.cpp

llama.cpp/main -m DeepSeek-V2-Lite.Q8_0.gguf --color -i -p "prompt here"

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?

Make sure you have gguf-split available
- To get hold of gguf-split, navigate to https://github.com/ggerganov/llama.cpp/releases
- Download the appropriate zip for your system from the latest release
- Unzip the archive and you should be able to find gguf-split
Locate your GGUF chunks folder (ex: DeepSeek-V2-Lite.Q8_0)
Run gguf-split --merge DeepSeek-V2-Lite.Q8_0/DeepSeek-V2-Lite.Q8_0-00001-of-XXXXX.gguf DeepSeek-V2-Lite.Q8_0.gguf
- Make sure to point gguf-split to the first chunk of the split.

Got a suggestion? Ping me @legraphista!