bartowski/LiquidAI_LFM2-350M-Extract-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/LiquidAI_LFM2-350M-Extract-GGUF
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-10-23 13:28:14 +08:00
commit 1a8ed72425
28 changed files with 309 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

72
.gitattributes vendored Normal file
View File

@@ -0,0 +1,72 @@
*.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
LiquidAI_LFM2-350M-Extract-Q2_K_L.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q5_K_L.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-IQ2_M.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-bf16.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-IQ4_NL.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q6_K_L.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q3_K_XL.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q4_1.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q4_K_L.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text
LiquidAI_LFM2-350M-Extract-IQ3_XXS.gguf filter=lfs diff=lfs merge=lfs -text

3
LiquidAI_LFM2-350M-Extract-IQ2_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:389407faeba7d8f74c762c58e7bd3f15bfe1c413bd9aa8b2b6da7149416094bd
size 142877248

3
LiquidAI_LFM2-350M-Extract-IQ3_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:a01750ac6b36d588c1ea6dde7d161106f73abb5aaa26637c10eb8591750c8f56
size 183657024

3
LiquidAI_LFM2-350M-Extract-IQ3_XS.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:1a518bcf3a4be2314a3d4280329f6973ed37a536eba7e1f5e47791bb1c0f1d78
size 175399488

3
LiquidAI_LFM2-350M-Extract-IQ3_XXS.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:ed70f66a1ca5ff887b0fa09fce2656125680b482a70e46105a02cd5293c2aea5
size 158474816

3
LiquidAI_LFM2-350M-Extract-IQ4_NL.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:7598697359336c99ca7153c8fc584580df4e7d35ea19b5df0db972e96e5e0a14
size 219308608

3
LiquidAI_LFM2-350M-Extract-IQ4_XS.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:bd3a5fe8493b13eaa044d39b1a59bcfcaae31374cb52f4dc986936029061ca5c
size 210330176

3
LiquidAI_LFM2-350M-Extract-Q2_K.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:ff5e4dd2e7ecfa40622e4d7a6357ac1c486b990c57ff49431a21624685936426
size 160596544

3
LiquidAI_LFM2-350M-Extract-Q2_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:68b9cf4fe5bee5aa3be54531adbe7deb85aee9167c02c8c13797b65b55494da5
size 176849472

3
LiquidAI_LFM2-350M-Extract-Q3_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:c9ba5d27626660414057bc3a5ac8a74067dbe48d0901b0eef610249728f22bf8
size 203047488

3
LiquidAI_LFM2-350M-Extract-Q3_K_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:af22d155c55e85ab6022b1a1c6e0568a817919135f72d4059ab8448dcc7019a3
size 193151552

3
LiquidAI_LFM2-350M-Extract-Q3_K_S.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:273b1f3f63eee0e6c33f9aa8e69501f3033255783d011f800e23c848bdc27d22
size 181150272

3
LiquidAI_LFM2-350M-Extract-Q3_K_XL.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:4a4372c4f3478c1800d0326b19c737645a77040a07903f9e864c9e7eedae6e88
size 219300416

3
LiquidAI_LFM2-350M-Extract-Q4_0.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:b4b7eff07349e5d88015c4f16232f9b0dfa89afdfe2f14f8826f899c478caca7
size 219898432

3
LiquidAI_LFM2-350M-Extract-Q4_1.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:2dd9d95aaeea44fdbd2b602d020cd0b29bbfa177488fe685fac04fc244a28f87
size 237265472

3
LiquidAI_LFM2-350M-Extract-Q4_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:230c63a35fedd75d383af40ddb32809a00ffeb7c5b4cf538d0baa1b5f3277ec4
size 245563968

3
LiquidAI_LFM2-350M-Extract-Q4_K_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:340a7fb24b98a7dbe933169dbbb869f4d89f8c7bf27ee45d62afabfc5b376743
size 229311040

3
LiquidAI_LFM2-350M-Extract-Q4_K_S.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:74cb86c0c038e7e45a498d2dd0f2eaf519583590120da01c8d63ee6cecab4653
size 220750400

3
LiquidAI_LFM2-350M-Extract-Q5_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:23524b3a723c4504646f4fddf9da188e6e1bac03237ec350036949152780810f
size 276628032

3
LiquidAI_LFM2-350M-Extract-Q5_K_M.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:e53777a7826cc5ff741d19bb95731714da593bbe9b928b891c540acd7c5885d7
size 260375104

3
LiquidAI_LFM2-350M-Extract-Q5_K_S.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:2a1bea9bd1f2a6e75036bdd0387fefc0861e76c6f639474adf4cfbce27ddc9f5
size 255222336

3
LiquidAI_LFM2-350M-Extract-Q6_K.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:55cf3fafd382970d7bcda49ead98ac8a8c585b324d6a4491ee90bcc1817aa593
size 293380672

3
LiquidAI_LFM2-350M-Extract-Q6_K_L.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:afd79bd0efb7c60d43e47d0f35af7479878ae6891b5a10c3a4592a795ba66926
size 309633600

3
LiquidAI_LFM2-350M-Extract-Q8_0.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:2723a24792da9d64f2a1f6b9661c13b52194843150125f524b8c4e78d6d1747d
size 379216448

3
LiquidAI_LFM2-350M-Extract-bf16.gguf Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:048aa54285a04e70338b70fbae6d5943cd3d0ffca62ca90dac2cf7f0e85cc90e
size 711483680

BIN
LiquidAI_LFM2-350M-Extract-imatrix.gguf Normal file
View File

Binary file not shown.

164
README.md Normal file
View File

@@ -0,0 +1,164 @@
---
quantized_by: bartowski
pipeline_tag: text-generation
base_model_relation: quantized
base_model: LiquidAI/LFM2-350M-Extract
---
## Llamacpp imatrix Quantizations of LFM2-350M-Extract by LiquidAI
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/b6683">b6683</a> for quantization.
Original model: https://huggingface.co/LiquidAI/LFM2-350M-Extract
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
No prompt format found, check original model page
## Download a file (not the whole branch) from below:
| Filename | Quant type | File Size | Split | Description |
| -------- | ---------- | --------- | ----- | ----------- |
| [LFM2-350M-Extract-bf16.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-bf16.gguf) | bf16 | 0.71GB | false | Full BF16 weights. |
| [LFM2-350M-Extract-Q8_0.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q8_0.gguf) | Q8_0 | 0.38GB | false | Extremely high quality, generally unneeded but max available quant. |
| [LFM2-350M-Extract-Q6_K_L.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q6_K_L.gguf) | Q6_K_L | 0.31GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. |
| [LFM2-350M-Extract-Q6_K.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q6_K.gguf) | Q6_K | 0.29GB | false | Very high quality, near perfect, *recommended*. |
| [LFM2-350M-Extract-Q5_K_L.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q5_K_L.gguf) | Q5_K_L | 0.28GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. |
| [LFM2-350M-Extract-Q5_K_M.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q5_K_M.gguf) | Q5_K_M | 0.26GB | false | High quality, *recommended*. |
| [LFM2-350M-Extract-Q5_K_S.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q5_K_S.gguf) | Q5_K_S | 0.26GB | false | High quality, *recommended*. |
| [LFM2-350M-Extract-Q4_K_L.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q4_K_L.gguf) | Q4_K_L | 0.25GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. |
| [LFM2-350M-Extract-Q4_1.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q4_1.gguf) | Q4_1 | 0.24GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. |
| [LFM2-350M-Extract-Q4_K_M.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q4_K_M.gguf) | Q4_K_M | 0.23GB | false | Good quality, default size for most use cases, *recommended*. |
| [LFM2-350M-Extract-Q4_K_S.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q4_K_S.gguf) | Q4_K_S | 0.22GB | false | Slightly lower quality with more space savings, *recommended*. |
| [LFM2-350M-Extract-Q4_0.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q4_0.gguf) | Q4_0 | 0.22GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. |
| [LFM2-350M-Extract-IQ4_NL.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-IQ4_NL.gguf) | IQ4_NL | 0.22GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. |
| [LFM2-350M-Extract-Q3_K_XL.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q3_K_XL.gguf) | Q3_K_XL | 0.22GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
| [LFM2-350M-Extract-IQ4_XS.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-IQ4_XS.gguf) | IQ4_XS | 0.21GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
| [LFM2-350M-Extract-Q3_K_L.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q3_K_L.gguf) | Q3_K_L | 0.20GB | false | Lower quality but usable, good for low RAM availability. |
| [LFM2-350M-Extract-Q3_K_M.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q3_K_M.gguf) | Q3_K_M | 0.19GB | false | Low quality. |
| [LFM2-350M-Extract-IQ3_M.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-IQ3_M.gguf) | IQ3_M | 0.18GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
| [LFM2-350M-Extract-Q3_K_S.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q3_K_S.gguf) | Q3_K_S | 0.18GB | false | Low quality, not recommended. |
| [LFM2-350M-Extract-IQ3_XS.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-IQ3_XS.gguf) | IQ3_XS | 0.18GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
| [LFM2-350M-Extract-Q2_K_L.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q2_K_L.gguf) | Q2_K_L | 0.18GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
| [LFM2-350M-Extract-IQ3_XXS.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-IQ3_XXS.gguf) | IQ3_XXS | 0.16GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. |
| [LFM2-350M-Extract-Q2_K.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-Q2_K.gguf) | Q2_K | 0.16GB | false | Very low quality but surprisingly usable. |
| [LFM2-350M-Extract-IQ2_M.gguf](https://huggingface.co/bartowski/LiquidAI_LFM2-350M-Extract-GGUF/blob/main/LiquidAI_LFM2-350M-Extract-IQ2_M.gguf) | IQ2_M | 0.14GB | 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/LiquidAI_LFM2-350M-Extract-GGUF --include "LiquidAI_LFM2-350M-Extract-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/LiquidAI_LFM2-350M-Extract-GGUF --include "LiquidAI_LFM2-350M-Extract-Q8_0/*" --local-dir ./
```
You can either specify a new local-dir (LiquidAI_LFM2-350M-Extract-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}