bartowski/HuggingFaceTB_SmolLM3-3B-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/HuggingFaceTB_SmolLM3-3B-GGUF
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-07-23 07:18:06 +08:00
commit 2d325869f9
28 changed files with 339 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

71
.gitattributes vendored Normal file
View File

@@ -0,0 +1,71 @@
*.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
*.gguf* 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
HuggingFaceTB_SmolLM3-3B-IQ2_M.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-IQ3_XXS.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-IQ4_NL.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q2_K_L.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q3_K_XL.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q4_1.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q4_K_L.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q5_K_L.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q6_K_L.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text
HuggingFaceTB_SmolLM3-3B-bf16.gguf filter=lfs diff=lfs merge=lfs -text

3
HuggingFaceTB_SmolLM3-3B-IQ2_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:96b88ac2f3b0e7fe336c80e22feafce9f99ffc0ff8f5780f949eb1124808b7b6
size 1125355328

3
HuggingFaceTB_SmolLM3-3B-IQ3_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:ddce138598e1f963e626334658d7a082d65fce70336d82e3e3eb1039c95d72d5
size 1469357888

3
HuggingFaceTB_SmolLM3-3B-IQ3_XS.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:c4e9d926ddef9ad235269c0460be583d885c0b09ba2ce11affc38ff42c583223
size 1371414336

3
HuggingFaceTB_SmolLM3-3B-IQ3_XXS.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:356b83c016a7613e79791c2c86acea765f672afa1e18d206377e8118cff41ab7
size 1267666752

3
HuggingFaceTB_SmolLM3-3B-IQ4_NL.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:af5e551b726a81b52bc06d7d9395f82890a326e36d814f03112fb0f9f41cfafe
size 1810538304

3
HuggingFaceTB_SmolLM3-3B-IQ4_XS.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:342fe3669e6db39bf9b9b514923f2af4a7d6b069efb45cc8f583cfc31cb75553
size 1723834176

3
HuggingFaceTB_SmolLM3-3B-Q2_K.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:794509645d684d6b563b054c312f502af793e040155080d4de358e698f638354
size 1253302080

3
HuggingFaceTB_SmolLM3-3B-Q2_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:27a3fbb59204030c2146452c193c12d946df13bd614b3f2131012adb4bf61ea5
size 1316917056

3
HuggingFaceTB_SmolLM3-3B-Q3_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:5e93d6464243fb3d809fcceaba79f1ceef02637ea61ad532af4ea95bc6fb7507
size 1690214208

3
HuggingFaceTB_SmolLM3-3B-Q3_K_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:5e5c70fdc3429023ed88ead8af2a70684047e2628258b650bdb729c8437639d0
size 1571069760

3
HuggingFaceTB_SmolLM3-3B-Q3_K_S.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:40e2c50e950947363e3222353407dc407fad8c0bb03314d8b2278966b7d508aa
size 1432313664

3
HuggingFaceTB_SmolLM3-3B-Q3_K_XL.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:07660c5582bf9b26e7558f8a83723cf4a60665340db9276fee435d8586d4091a
size 1753829184

3
HuggingFaceTB_SmolLM3-3B-Q4_0.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:1e0e56d458e6b6d5ea18d0e5fe598cca252450308f1648afab49984fceae47da
size 1811455808

3
HuggingFaceTB_SmolLM3-3B-Q4_1.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:2603ac4acb987b702fb45bb439e752313a63fe02ae6f9b837cf7ef075e77a1a3
size 1981587264

3
HuggingFaceTB_SmolLM3-3B-Q4_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:e6ee0b21fff62763bae1d7ce829b5e22f92e06e7560a809b0f22d0e7347cc491
size 1978920768

3
HuggingFaceTB_SmolLM3-3B-Q4_K_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:519732558d5fa7420ab058e1b776dcfe73da78013c2fe59c7ca43c325ef89132
size 1915305792

3
HuggingFaceTB_SmolLM3-3B-Q4_K_S.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:a1df314cf8900f93d02628430c1325872bb355f9258f858b966fed2aaff2df28
size 1817616192

3
HuggingFaceTB_SmolLM3-3B-Q5_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:151d2efab4d8d7fbbdb8bd2653b9cb322ed44f937c27dd2f3e328411644a0b7d
size 2277371712

3
HuggingFaceTB_SmolLM3-3B-Q5_K_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:65a6490c29a728d1bde5b6a6858418715e773e061433386dbbc39b3f0a088f3d
size 2213756736

3
HuggingFaceTB_SmolLM3-3B-Q5_K_S.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:1bcedb1cd1b6a4a19075295fdfecffa5b324db58101b5806c10cabd3c7ffedbe
size 2157354816

3
HuggingFaceTB_SmolLM3-3B-Q6_K.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:4c7ad22be780006817f727c8a8c0fdc75d35f995c6a91e0210a931cbb20ba6fd
size 2530860864

3
HuggingFaceTB_SmolLM3-3B-Q6_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:406c5a1985fd5be91b865c6945a27ce28b5104dca338f042c0bd7f7f7cfe6175
size 2594475840

3
HuggingFaceTB_SmolLM3-3B-Q8_0.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:32056df2ec827a1f7c14b6b7f48239fe1d288aba086c811c96a61a8c9df6c733
size 3275575104

3
HuggingFaceTB_SmolLM3-3B-bf16.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:66a39ad173171c36bce86769449f8db61c062689f964a032995976d20b8245f9
size 6158339616

BIN
HuggingFaceTB_SmolLM3-3B.imatrix Normal file
View File

Binary file not shown.

195
README.md Normal file
View File

@@ -0,0 +1,195 @@
---
quantized_by: bartowski
pipeline_tag: text-generation
base_model: HuggingFaceTB/SmolLM3-3B
base_model_relation: quantized
license: apache-2.0
language:
- en
- fr
- es
- it
- pt
- zh
- ar
- ru
---
## Llamacpp imatrix Quantizations of SmolLM3-3B by HuggingFaceTB
Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b5856">b5856</a> for quantization.
Original model: https://huggingface.co/HuggingFaceTB/SmolLM3-3B
All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)
Run them in [LM Studio](https://lmstudio.ai/)
Run them directly with [llama.cpp](https://github.com/ggerganov/llama.cpp), or any other llama.cpp based project
## Prompt format
No chat template specified so default is used. This may be incorrect, check original model card for details.
```
<|im_start|>system
## Metadata
Knowledge Cutoff Date: June 2025
Today Date: 09 July 2025
Reasoning Mode: /think
## Custom Instructions
{system_prompt}
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
```
## What's new:
Fix chat template
## Download a file (not the whole branch) from below:
| Filename | Quant type | File Size | Split | Description |
| -------- | ---------- | --------- | ----- | ----------- |
| [SmolLM3-3B-bf16.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-bf16.gguf) | bf16 | 6.16GB | false | Full BF16 weights. |
| [SmolLM3-3B-Q8_0.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q8_0.gguf) | Q8_0 | 3.28GB | false | Extremely high quality, generally unneeded but max available quant. |
| [SmolLM3-3B-Q6_K_L.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q6_K_L.gguf) | Q6_K_L | 2.59GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. |
| [SmolLM3-3B-Q6_K.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q6_K.gguf) | Q6_K | 2.53GB | false | Very high quality, near perfect, *recommended*. |
| [SmolLM3-3B-Q5_K_L.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q5_K_L.gguf) | Q5_K_L | 2.28GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. |
| [SmolLM3-3B-Q5_K_M.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q5_K_M.gguf) | Q5_K_M | 2.21GB | false | High quality, *recommended*. |
| [SmolLM3-3B-Q5_K_S.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q5_K_S.gguf) | Q5_K_S | 2.16GB | false | High quality, *recommended*. |
| [SmolLM3-3B-Q4_K_L.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q4_K_L.gguf) | Q4_K_L | 1.98GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. |
| [SmolLM3-3B-Q4_1.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q4_1.gguf) | Q4_1 | 1.98GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. |
| [SmolLM3-3B-Q4_K_M.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q4_K_M.gguf) | Q4_K_M | 1.92GB | false | Good quality, default size for most use cases, *recommended*. |
| [SmolLM3-3B-Q4_K_S.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q4_K_S.gguf) | Q4_K_S | 1.82GB | false | Slightly lower quality with more space savings, *recommended*. |
| [SmolLM3-3B-Q4_0.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q4_0.gguf) | Q4_0 | 1.81GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. |
| [SmolLM3-3B-IQ4_NL.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-IQ4_NL.gguf) | IQ4_NL | 1.81GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. |
| [SmolLM3-3B-Q3_K_XL.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q3_K_XL.gguf) | Q3_K_XL | 1.75GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
| [SmolLM3-3B-IQ4_XS.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-IQ4_XS.gguf) | IQ4_XS | 1.72GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
| [SmolLM3-3B-Q3_K_L.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q3_K_L.gguf) | Q3_K_L | 1.69GB | false | Lower quality but usable, good for low RAM availability. |
| [SmolLM3-3B-Q3_K_M.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q3_K_M.gguf) | Q3_K_M | 1.57GB | false | Low quality. |
| [SmolLM3-3B-IQ3_M.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-IQ3_M.gguf) | IQ3_M | 1.47GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
| [SmolLM3-3B-Q3_K_S.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q3_K_S.gguf) | Q3_K_S | 1.43GB | false | Low quality, not recommended. |
| [SmolLM3-3B-IQ3_XS.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-IQ3_XS.gguf) | IQ3_XS | 1.37GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
| [SmolLM3-3B-Q2_K_L.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q2_K_L.gguf) | Q2_K_L | 1.32GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
| [SmolLM3-3B-IQ3_XXS.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-IQ3_XXS.gguf) | IQ3_XXS | 1.27GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. |
| [SmolLM3-3B-Q2_K.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-Q2_K.gguf) | Q2_K | 1.25GB | false | Very low quality but surprisingly usable. |
| [SmolLM3-3B-IQ2_M.gguf](https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF/blob/main/HuggingFaceTB_SmolLM3-3B-IQ2_M.gguf) | IQ2_M | 1.13GB | 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/HuggingFaceTB_SmolLM3-3B-GGUF --include "HuggingFaceTB_SmolLM3-3B-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/HuggingFaceTB_SmolLM3-3B-GGUF --include "HuggingFaceTB_SmolLM3-3B-Q8_0/*" --local-dir ./
```
You can either specify a new local-dir (HuggingFaceTB_SmolLM3-3B-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/ggerganov/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/ggerganov/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/ggerganov/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/ggerganov/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}