🚀 Llamacpp imatrix Quantizations of OpenCodeReasoning-Nemotron-14B by nvidia
This project provides quantized versions of the OpenCodeReasoning-Nemotron-14B model by NVIDIA, enabling efficient deployment and usage.
🚀 Quick Start
This repository contains quantized versions of the OpenCodeReasoning-Nemotron-14B model. You can run these quantized models in LM Studio or directly with llama.cpp, or any other llama.cpp based project.
✨ Features
- Quantization: Using llama.cpp release b5306 for quantization.
- Multiple Quant Types: Offers a variety of quant types to meet different needs, such as BF16, Q8_0, Q6_K_L, etc.
- Online Repacking: Some quant types support online repacking for better performance on ARM and AVX machines.
📦 Installation
Download a file (not the whole branch)
You can download the specific quantized model file from the following table:
Filename |
Quant type |
File Size |
Split |
Description |
OpenCodeReasoning-Nemotron-14B-bf16.gguf |
bf16 |
29.55GB |
false |
Full BF16 weights. |
OpenCodeReasoning-Nemotron-14B-Q8_0.gguf |
Q8_0 |
15.70GB |
false |
Extremely high quality, generally unneeded but max available quant. |
OpenCodeReasoning-Nemotron-14B-Q6_K_L.gguf |
Q6_K_L |
12.50GB |
false |
Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended. |
OpenCodeReasoning-Nemotron-14B-Q6_K.gguf |
Q6_K |
12.12GB |
false |
Very high quality, near perfect, recommended. |
OpenCodeReasoning-Nemotron-14B-Q5_K_L.gguf |
Q5_K_L |
10.99GB |
false |
Uses Q8_0 for embed and output weights. High quality, recommended. |
OpenCodeReasoning-Nemotron-14B-Q5_K_M.gguf |
Q5_K_M |
10.51GB |
false |
High quality, recommended. |
OpenCodeReasoning-Nemotron-14B-Q5_K_S.gguf |
Q5_K_S |
10.27GB |
false |
High quality, recommended. |
OpenCodeReasoning-Nemotron-14B-Q4_K_L.gguf |
Q4_K_L |
9.57GB |
false |
Uses Q8_0 for embed and output weights. Good quality, recommended. |
OpenCodeReasoning-Nemotron-14B-Q4_1.gguf |
Q4_1 |
9.39GB |
false |
Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. |
OpenCodeReasoning-Nemotron-14B-Q4_K_M.gguf |
Q4_K_M |
8.99GB |
false |
Good quality, default size for most use cases, recommended. |
OpenCodeReasoning-Nemotron-14B-Q3_K_XL.gguf |
Q3_K_XL |
8.61GB |
false |
Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
OpenCodeReasoning-Nemotron-14B-Q4_K_S.gguf |
Q4_K_S |
8.57GB |
false |
Slightly lower quality with more space savings, recommended. |
OpenCodeReasoning-Nemotron-14B-IQ4_NL.gguf |
IQ4_NL |
8.55GB |
false |
Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. |
OpenCodeReasoning-Nemotron-14B-Q4_0.gguf |
Q4_0 |
8.54GB |
false |
Legacy format, offers online repacking for ARM and AVX CPU inference. |
OpenCodeReasoning-Nemotron-14B-IQ4_XS.gguf |
IQ4_XS |
8.12GB |
false |
Decent quality, smaller than Q4_K_S with similar performance, recommended. |
OpenCodeReasoning-Nemotron-14B-Q3_K_L.gguf |
Q3_K_L |
7.92GB |
false |
Lower quality but usable, good for low RAM availability. |
OpenCodeReasoning-Nemotron-14B-Q3_K_M.gguf |
Q3_K_M |
7.34GB |
false |
Low quality. |
OpenCodeReasoning-Nemotron-14B-IQ3_M.gguf |
IQ3_M |
6.92GB |
false |
Medium-low quality, new method with decent performance comparable to Q3_K_M. |
OpenCodeReasoning-Nemotron-14B-Q3_K_S.gguf |
Q3_K_S |
6.66GB |
false |
Low quality, not recommended. |
OpenCodeReasoning-Nemotron-14B-Q2_K_L.gguf |
Q2_K_L |
6.53GB |
false |
Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
OpenCodeReasoning-Nemotron-14B-IQ3_XS.gguf |
IQ3_XS |
6.38GB |
false |
Lower quality, new method with decent performance, slightly better than Q3_K_S. |
OpenCodeReasoning-Nemotron-14B-IQ3_XXS.gguf |
IQ3_XXS |
5.95GB |
false |
Lower quality, new method with decent performance, comparable to Q3 quants. |
OpenCodeReasoning-Nemotron-14B-Q2_K.gguf |
Q2_K |
5.77GB |
false |
Very low quality but surprisingly usable. |
OpenCodeReasoning-Nemotron-14B-IQ2_M.gguf |
IQ2_M |
5.36GB |
false |
Relatively low quality, uses SOTA techniques to be surprisingly usable. |
OpenCodeReasoning-Nemotron-14B-IQ2_S.gguf |
IQ2_S |
5.00GB |
false |
Low quality, uses SOTA techniques to be usable. |
Downloading using huggingface-cli
Click to view download instructions
First, make sure you have hugginface-cli installed:
pip install -U "huggingface_hub[cli]"
Then, you can target the specific file you want:
huggingface-cli download bartowski/nvidia_OpenCodeReasoning-Nemotron-14B-GGUF --include "nvidia_OpenCodeReasoning-Nemotron-14B-Q4_K_M.gguf" --local-dir ./
If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:
huggingface-cli download bartowski/nvidia_OpenCodeReasoning-Nemotron-14B-GGUF --include "nvidia_OpenCodeReasoning-Nemotron-14B-Q8_0/*" --local-dir ./
You can either specify a new local-dir (nvidia_OpenCodeReasoning-Nemotron-14B-Q8_0) or download them all in place (./)
💻 Usage Examples
Prompt format
<|im_start|>system
{system_prompt}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
📚 Documentation
Embed/output weights
Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.
ARM/AVX information
Previously, you would download Q4_0_4_4/4_8/8_8, and these would have their weights interleaved in memory in order to improve performance on ARM and AVX machines by loading up more data in one pass.
Now, however, there is something called "online repacking" for weights. details in this PR. If you use Q4_0 and your hardware would benefit from repacking weights, it will do it automatically on the fly.
As of llama.cpp build b4282 you will not be able to run the Q4_0_X_X files and will instead need to use Q4_0.
Additionally, if you want to get slightly better quality for , you can use IQ4_NL thanks to this PR which will also repack the weights for ARM, though only the 4_4 for now. The loading time may be slower but it will result in an overall speed incrase.
Click to view Q4_0_X_X information (deprecated)
I'm keeping this section to show the potential theoretical uplift in performance from using the Q4_0 with online repacking.
Click to view benchmarks on an AVX2 system (EPYC7702)
model |
size |
params |
backend |
threads |
test |
t/s |
% (vs Q4_0) |
qwen2 3B Q4_0 |
1.70 GiB |
3.09 B |
CPU |
64 |
pp512 |
204.03 ± 1.03 |
100% |
qwen2 3B Q4_0 |
1.70 GiB |
3.09 B |
CPU |
64 |
pp1024 |
282.92 ± 0.19 |
100% |
qwen2 3B Q4_0 |
1.70 GiB |
3.09 B |
CPU |
64 |
pp2048 |
259.49 ± 0.44 |
100% |
qwen2 3B Q4_0 |
1.70 GiB |
3.09 B |
CPU |
64 |
tg128 |
39.12 ± 0.27 |
100% |
qwen2 3B Q4_0 |
1.70 GiB |
3.09 B |
CPU |
64 |
tg256 |
39.31 ± 0.69 |
100% |
qwen2 3B Q4_0 |
1.70 GiB |
3.09 B |
CPU |
64 |
tg512 |
40.52 ± 0.03 |
100% |
qwen2 3B Q4_K_M |
1.79 GiB |
3.09 B |
CPU |
64 |
pp512 |
301.02 ± 1.74 |
147% |
qwen2 3B Q4_K_M |
1.79 GiB |
3.09 B |
CPU |
64 |
pp1024 |
287.23 ± 0.20 |
101% |
qwen2 3B Q4_K_M |
1.79 GiB |
3.09 B |
CPU |
64 |
pp2048 |
262.77 ± 1.81 |
101% |
qwen2 3B Q4_K_M |
1.79 GiB |
3.09 B |
CPU |
64 |
tg128 |
18.80 ± 0.99 |
48% |
qwen2 3B Q4_K_M |
1.79 GiB |
3.09 B |
CPU |
64 |
tg256 |
24.46 ± 3.04 |
83% |
qwen2 3B Q4_K_M |
1.79 GiB |
3.09 B |
CPU |
64 |
tg512 |
36.32 ± 3.59 |
90% |
qwen2 3B Q4_0_8_8 |
1.69 GiB |
3.09 B |
CPU |
64 |
pp512 |
271.71 ± 3.53 |
133% |
qwen2 3B Q4_0_8_8 |
1.69 GiB |
3.09 B |
CPU |
64 |
pp1024 |
279.86 ± 45.63 |
100% |
qwen2 3B Q4_0_8_8 |
1.69 GiB |
3.09 B |
CPU |
64 |
pp2048 |
320.77 ± 5.00 |
124% |
qwen2 3B Q4_0_8_8 |
1.69 GiB |
3.09 B |
CPU |
64 |
tg128 |
43.51 ± 0.05 |
111% |
qwen2 3B Q4_0_8_8 |
1.69 GiB |
3.09 B |
CPU |
64 |
tg256 |
43.35 ± 0.09 |
110% |
qwen2 3B Q4_0_8_8 |
1.69 GiB |
3.09 B |
CPU |
64 |
tg512 |
43.35 ± 0.09 |
110% |
📄 License
This project is licensed under the Apache-2.0 license.