Minitron-4B-Base-GGUF

Llama.cpp static quantization of nvidia/Minitron-4B-Base

Original Model: nvidia/Minitron-4B-Base
Original dtype: BF16 (bfloat16)
Quantized by: llama.cpp b3600
IMatrix dataset: here


Files

Common Quants

Filename Quant type File Size Status Uses IMatrix Is Split
Minitron-4B-Base.Q8_0.gguf Q8_0 4.46GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q6_K.gguf Q6_K 3.45GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q4_K.gguf Q4_K 2.70GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q3_K.gguf Q3_K 2.30GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q2_K.gguf Q2_K 1.90GB βœ… Available βšͺ Static πŸ“¦ No

All Quants

Filename Quant type File Size Status Uses IMatrix Is Split
Minitron-4B-Base.BF16.gguf BF16 8.39GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.FP16.gguf F16 8.39GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q8_0.gguf Q8_0 4.46GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q6_K.gguf Q6_K 3.45GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q5_K.gguf Q5_K 3.06GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q5_K_S.gguf Q5_K_S 2.99GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q4_K.gguf Q4_K 2.70GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q4_K_S.gguf Q4_K_S 2.58GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.IQ4_NL.gguf IQ4_NL 2.58GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.IQ4_XS.gguf IQ4_XS 2.48GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q3_K.gguf Q3_K 2.30GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q3_K_L.gguf Q3_K_L 2.45GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q3_K_S.gguf Q3_K_S 2.12GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.IQ3_M.gguf IQ3_M 2.18GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.IQ3_S.gguf IQ3_S 2.12GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.IQ3_XS.gguf IQ3_XS 2.06GB βœ… Available βšͺ Static πŸ“¦ No
Minitron-4B-Base.Q2_K.gguf Q2_K 1.90GB βœ… Available βšͺ Static 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/Minitron-4B-Base-GGUF --include "Minitron-4B-Base.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/Minitron-4B-Base-GGUF --include "Minitron-4B-Base.Q8_0/*" --local-dir ./
# see FAQ for merging GGUF's

Inference

Llama.cpp

llama.cpp/main -m Minitron-4B-Base.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?

  1. Make sure you have gguf-split available
  2. Locate your GGUF chunks folder (ex: Minitron-4B-Base.Q8_0)
  3. Run gguf-split --merge Minitron-4B-Base.Q8_0/Minitron-4B-Base.Q8_0-00001-of-XXXXX.gguf Minitron-4B-Base.Q8_0.gguf
    • Make sure to point gguf-split to the first chunk of the split.

Got a suggestion? Ping me @legraphista!

Downloads last month
76
GGUF
Model size
4.19B params
Architecture
nemotron

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Inference Examples
Inference API (serverless) has been turned off for this model.

Model tree for legraphista/Minitron-4B-Base-GGUF

Quantized
(8)
this model