bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-07-17 01:53:06 +08:00
commit 0dac1e79c3
28 changed files with 326 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

73
.gitattributes vendored Normal file
View File

@@ -0,0 +1,73 @@
*.7z filter=lfs diff=lfs merge=lfs -text
*.arrow filter=lfs diff=lfs merge=lfs -text
*.bin filter=lfs diff=lfs merge=lfs -text
*.bin.* filter=lfs diff=lfs merge=lfs -text
*.bz2 filter=lfs diff=lfs merge=lfs -text
*.ftz filter=lfs diff=lfs merge=lfs -text
*.gz filter=lfs diff=lfs merge=lfs -text
*.h5 filter=lfs diff=lfs merge=lfs -text
*.joblib filter=lfs diff=lfs merge=lfs -text
*.lfs.* filter=lfs diff=lfs merge=lfs -text
*.model filter=lfs diff=lfs merge=lfs -text
*.msgpack filter=lfs diff=lfs merge=lfs -text
*.onnx filter=lfs diff=lfs merge=lfs -text
*.ot filter=lfs diff=lfs merge=lfs -text
*.parquet filter=lfs diff=lfs merge=lfs -text
*.pb filter=lfs diff=lfs merge=lfs -text
*.pt filter=lfs diff=lfs merge=lfs -text
*.pth filter=lfs diff=lfs merge=lfs -text
*.rar filter=lfs diff=lfs merge=lfs -text
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
*.tar.* filter=lfs diff=lfs merge=lfs -text
*.tflite filter=lfs diff=lfs merge=lfs -text
*.tgz filter=lfs diff=lfs merge=lfs -text
*.xz filter=lfs diff=lfs merge=lfs -text
*.zip filter=lfs diff=lfs merge=lfs -text
*.zstandard filter=lfs diff=lfs merge=lfs -text
*.tfevents* filter=lfs diff=lfs merge=lfs -text
*.db* filter=lfs diff=lfs merge=lfs -text
*.ark* filter=lfs diff=lfs merge=lfs -text
**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text
**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text
**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text
*.safetensors filter=lfs diff=lfs merge=lfs -text
*.ckpt filter=lfs diff=lfs merge=lfs -text
*.ggml filter=lfs diff=lfs merge=lfs -text
*.llamafile* filter=lfs diff=lfs merge=lfs -text
*.pt2 filter=lfs diff=lfs merge=lfs -text
*.mlmodel filter=lfs diff=lfs merge=lfs -text
*.npy filter=lfs diff=lfs merge=lfs -text
*.npz filter=lfs diff=lfs merge=lfs -text
*.pickle filter=lfs diff=lfs merge=lfs -text
*.pkl filter=lfs diff=lfs merge=lfs -text
*.tar filter=lfs diff=lfs merge=lfs -text
*.wasm filter=lfs diff=lfs merge=lfs -text
*.zst filter=lfs diff=lfs merge=lfs -text
*tfevents* filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q2_K_L.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-bf16.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_K_L.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_1.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-imatrix.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q5_K_L.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ2_M.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_XL.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ3_XXS.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q6_K_L.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ4_NL.gguf filter=lfs diff=lfs merge=lfs -text
Nanbeige_Nanbeige4-3B-Thinking-2511-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text

3
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ2_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:7e81701fad657336c3675857ed07e6f0f680b96589ea31acaa46e84591557dbb
size 1512122816

3
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ3_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:e5198505c8b8edbf6617e3e7ddd5733a77aadcfdd3492ff07a192a683b9edee0
size 1908641216

3
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ3_XS.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:3fdc64f34243a18e9a27dad3cdb21c212a67f548b4d5c174b500115ceebf6110
size 1794260416

3
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ3_XXS.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:72cfb6f97845eed7b805f983c626eb250399cdab10abcb785bb45d4901c93427
size 1661585856

3
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ4_NL.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:c49cd4652cffc3da91986ac2c3dccdf9ab19afe007445f30fe83001eb29b44cf
size 2332111296

3
Nanbeige_Nanbeige4-3B-Thinking-2511-IQ4_XS.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:67136b9b3596f6dddbc8d86c7a8451bbe1419ce575e792b35ffa2e31452132b7
size 2223792576

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q2_K.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:674b74391d30ebb4e2d7e0a21ab2406e18e679c55ecfbbc27d46f69b3ba5867b
size 1623103936

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q2_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:302151033cec8d63ea7a0592f51f19bef31992b8a34de9fda59d435d420c4834
size 2038463936

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:f7cd150912768f2ddab222d8f906444d815cc0317756f6e70250aade8b9c47ac
size 2147499456

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:8580a6ee46c5fdda43bc94c469116322c2c2e2618f0a96a9838cb3cc5e87935f
size 2015608256

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_S.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:27b039d40b78d547f8934ba0e2f28e3f8f68c38d87a282093583f18b0531a9cb
size 1860943296

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_XL.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:7b4c38c36f9f355d5e52183d63d8153d63301aac469f7b7b6fcba4f46229b235
size 2519662016

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_0.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:cb0ad951743be35b2496f66281ca1b12e3fa0e5341bc8bf91947d1ee5fa72276
size 2333585856

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_1.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:e1316e60b72706f40bf6be612252128b1a2fd60384f557308e198ea0026c4484
size 2546127296

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:9dec7be4154583e92c4ca710f4d97a23944d6d8b91a76e14b480efcfcd689929
size 2758786496

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_K_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:4a32fe5ce083680a4be55ff8a17d925fcfe8d9adc0064fae7723dcb412f20a25
size 2443112896

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_K_S.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:53270e7202d4364b2baf0736e5477f9e235ff7070bced6be76e81da728407add
size 2340958656

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q5_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:52bfd4be7a69cf7b89dbc0a3f019c20a5f8663a76c28b6303702b8a7e37e8d6c
size 3087777216

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q5_K_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:a23e9ea20a97feac2f015a6f883bea6614baee2f98eec509d147eda174c62b49
size 2825269696

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q5_K_S.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:45f66cd304c76aec2653bfde0347bd1137284e5044b591255324e406641790ec
size 2765386176

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q6_K.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:4764452db3f8443b903895e78b9c5cd375ca868a684cc4e8b25c888c67e2647b
size 3231311296

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q6_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:bd7a42e1d9b7110fe9ff6f9f0aac9856faea2601cf522b45db3c0467c7193e66
size 3437329856

3
Nanbeige_Nanbeige4-3B-Thinking-2511-Q8_0.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:0901e8c327f5ef52ba9394fa3e0b203bc796748547137475f479b79014a01446
size 4183948736

3
Nanbeige_Nanbeige4-3B-Thinking-2511-bf16.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:473006fb9122ed1e8d14f39051ad354c736f1f46ec9032a0c5603794eecffec9
size 7871577216

3
Nanbeige_Nanbeige4-3B-Thinking-2511-imatrix.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:7af9cdf7154bc156d6f68c88a4841fe17f0e2af5db11db4ee70e78e39717a2db
size 3344064

177
README.md Normal file
View File

@@ -0,0 +1,177 @@
---
quantized_by: bartowski
pipeline_tag: text-generation
license: apache-2.0
base_model: Nanbeige/Nanbeige4-3B-Thinking-2511
language:
- en
- zh
tags:
- llm
- nanbeige
base_model_relation: quantized
---
## Llamacpp imatrix Quantizations of Nanbeige4-3B-Thinking-2511 by Nanbeige
Using <a href="https://github.com/ggml-org/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggml-org/llama.cpp/releases/tag/b7340">b7340</a> for quantization.
Original model: https://huggingface.co/Nanbeige/Nanbeige4-3B-Thinking-2511
All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8) combined with a subset of combined_all_small.parquet from Ed Addario [here](https://huggingface.co/datasets/eaddario/imatrix-calibration/blob/main/combined_all_small.parquet)
Run them in [LM Studio](https://lmstudio.ai/)
Run them directly with [llama.cpp](https://github.com/ggml-org/llama.cpp), or any other llama.cpp based project
## Prompt format
```
<|im_start|>system
{system_prompt}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
```
## Download a file (not the whole branch) from below:
| Filename | Quant type | File Size | Split | Description |
| -------- | ---------- | --------- | ----- | ----------- |
| [Nanbeige4-3B-Thinking-2511-bf16.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-bf16.gguf) | bf16 | 7.87GB | false | Full BF16 weights. |
| [Nanbeige4-3B-Thinking-2511-Q8_0.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q8_0.gguf) | Q8_0 | 4.18GB | false | Extremely high quality, generally unneeded but max available quant. |
| [Nanbeige4-3B-Thinking-2511-Q6_K_L.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q6_K_L.gguf) | Q6_K_L | 3.44GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. |
| [Nanbeige4-3B-Thinking-2511-Q6_K.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q6_K.gguf) | Q6_K | 3.23GB | false | Very high quality, near perfect, *recommended*. |
| [Nanbeige4-3B-Thinking-2511-Q5_K_L.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q5_K_L.gguf) | Q5_K_L | 3.09GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. |
| [Nanbeige4-3B-Thinking-2511-Q5_K_M.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q5_K_M.gguf) | Q5_K_M | 2.83GB | false | High quality, *recommended*. |
| [Nanbeige4-3B-Thinking-2511-Q5_K_S.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q5_K_S.gguf) | Q5_K_S | 2.77GB | false | High quality, *recommended*. |
| [Nanbeige4-3B-Thinking-2511-Q4_K_L.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_K_L.gguf) | Q4_K_L | 2.76GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. |
| [Nanbeige4-3B-Thinking-2511-Q4_1.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_1.gguf) | Q4_1 | 2.55GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. |
| [Nanbeige4-3B-Thinking-2511-Q3_K_XL.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_XL.gguf) | Q3_K_XL | 2.52GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
| [Nanbeige4-3B-Thinking-2511-Q4_K_M.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_K_M.gguf) | Q4_K_M | 2.44GB | false | Good quality, default size for most use cases, *recommended*. |
| [Nanbeige4-3B-Thinking-2511-Q4_K_S.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_K_S.gguf) | Q4_K_S | 2.34GB | false | Slightly lower quality with more space savings, *recommended*. |
| [Nanbeige4-3B-Thinking-2511-Q4_0.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q4_0.gguf) | Q4_0 | 2.33GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. |
| [Nanbeige4-3B-Thinking-2511-IQ4_NL.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-IQ4_NL.gguf) | IQ4_NL | 2.33GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. |
| [Nanbeige4-3B-Thinking-2511-IQ4_XS.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-IQ4_XS.gguf) | IQ4_XS | 2.22GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
| [Nanbeige4-3B-Thinking-2511-Q3_K_L.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_L.gguf) | Q3_K_L | 2.15GB | false | Lower quality but usable, good for low RAM availability. |
| [Nanbeige4-3B-Thinking-2511-Q2_K_L.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q2_K_L.gguf) | Q2_K_L | 2.04GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
| [Nanbeige4-3B-Thinking-2511-Q3_K_M.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_M.gguf) | Q3_K_M | 2.02GB | false | Low quality. |
| [Nanbeige4-3B-Thinking-2511-IQ3_M.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-IQ3_M.gguf) | IQ3_M | 1.91GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
| [Nanbeige4-3B-Thinking-2511-Q3_K_S.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q3_K_S.gguf) | Q3_K_S | 1.86GB | false | Low quality, not recommended. |
| [Nanbeige4-3B-Thinking-2511-IQ3_XS.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-IQ3_XS.gguf) | IQ3_XS | 1.79GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
| [Nanbeige4-3B-Thinking-2511-IQ3_XXS.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-IQ3_XXS.gguf) | IQ3_XXS | 1.66GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. |
| [Nanbeige4-3B-Thinking-2511-Q2_K.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-Q2_K.gguf) | Q2_K | 1.62GB | false | Very low quality but surprisingly usable. |
| [Nanbeige4-3B-Thinking-2511-IQ2_M.gguf](https://huggingface.co/bartowski/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF/blob/main/Nanbeige_Nanbeige4-3B-Thinking-2511-IQ2_M.gguf) | IQ2_M | 1.51GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. |
## 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.
## Downloading using huggingface-cli
<details>
<summary>Click to view download instructions</summary>
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/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF --include "Nanbeige_Nanbeige4-3B-Thinking-2511-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/Nanbeige_Nanbeige4-3B-Thinking-2511-GGUF --include "Nanbeige_Nanbeige4-3B-Thinking-2511-Q8_0/*" --local-dir ./
```
You can either specify a new local-dir (Nanbeige_Nanbeige4-3B-Thinking-2511-Q8_0) or download them all in place (./)
</details>
## 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](https://github.com/ggml-org/llama.cpp/pull/9921). 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](https://github.com/ggml-org/llama.cpp/releases/tag/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](https://github.com/ggml-org/llama.cpp/pull/10541) 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.
<details>
<summary>Click to view Q4_0_X_X information (deprecated</summary>
I'm keeping this section to show the potential theoretical uplift in performance from using the Q4_0 with online repacking.
<details>
<summary>Click to view benchmarks on an AVX2 system (EPYC7702)</summary>
| 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 | 42.60 ± 0.31 | 105% |
Q4_0_8_8 offers a nice bump to prompt processing and a small bump to text generation
</details>
</details>
## Which file should I choose?
<details>
<summary>Click here for details</summary>
A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.
If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.
If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.
Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.
If you want to get more into the weeds, you can check out this extremely useful feature chart:
[llama.cpp feature matrix](https://github.com/ggml-org/llama.cpp/wiki/Feature-matrix)
But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.
These I-quants can also be used on CPU, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.
</details>
## Credits
Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset.
Thank you ZeroWw for the inspiration to experiment with embed/output.
Thank you to LM Studio for sponsoring my work.
Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski

1
configuration.json Normal file
View File

@@ -0,0 +1 @@
{"framework": "pytorch", "task": "text-generation", "allow_remote": true}