From 9a886d0ec3b37552bedba496678d71e0af8142d1 Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Fri, 4 Sep 2026 03:16:12 +0800 Subject: [PATCH] =?UTF-8?q?=E5=88=9D=E5=A7=8B=E5=8C=96=E9=A1=B9=E7=9B=AE?= =?UTF-8?q?=EF=BC=8C=E7=94=B1ModelHub=20XC=E7=A4=BE=E5=8C=BA=E6=8F=90?= =?UTF-8?q?=E4=BE=9B=E6=A8=A1=E5=9E=8B?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Model: bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF Source: Original Platform --- .gitattributes | 49 +++++++ README.md | 178 +++++++++++++++++++++++++ configuration.json | 1 + mrm8488_Qwen3-14B-ft-limo-IQ2_M.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-IQ2_S.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-IQ2_XS.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-IQ3_M.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-IQ3_XS.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-IQ3_XXS.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-IQ4_NL.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-IQ4_XS.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q2_K.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q2_K_L.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q3_K_L.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q3_K_M.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q3_K_S.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q3_K_XL.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q4_0.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q4_1.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q4_K_L.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q4_K_M.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q4_K_S.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q5_K_L.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q5_K_M.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q5_K_S.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q6_K.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q6_K_L.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-Q8_0.gguf | 3 + mrm8488_Qwen3-14B-ft-limo-bf16.gguf | 3 + mrm8488_Qwen3-14B-ft-limo.imatrix | 3 + 30 files changed, 309 insertions(+) create mode 100644 .gitattributes create mode 100644 README.md create mode 100644 configuration.json create mode 100644 mrm8488_Qwen3-14B-ft-limo-IQ2_M.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-IQ2_S.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-IQ2_XS.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-IQ3_M.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-IQ3_XS.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-IQ3_XXS.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-IQ4_NL.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-IQ4_XS.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q2_K.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q2_K_L.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q3_K_L.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q3_K_M.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q3_K_S.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q3_K_XL.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q4_0.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q4_1.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q4_K_L.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q4_K_M.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q4_K_S.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q5_K_L.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q5_K_M.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q5_K_S.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q6_K.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q6_K_L.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-Q8_0.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo-bf16.gguf create mode 100644 mrm8488_Qwen3-14B-ft-limo.imatrix diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..dbf27fd --- /dev/null +++ b/.gitattributes @@ -0,0 +1,49 @@ +*.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 + +mrm8488_Qwen3-14B-ft-limo.imatrix filter=lfs diff=lfs merge=lfs -text \ No newline at end of file diff --git a/README.md b/README.md new file mode 100644 index 0000000..4523aee --- /dev/null +++ b/README.md @@ -0,0 +1,178 @@ +--- +quantized_by: bartowski +pipeline_tag: text-generation +tags: +- limo +datasets: +- GAIR/LIMO +base_model: mrm8488/Qwen3-14B-ft-limo +thumbnail: https://huggingface.co/mrm8488/Qwen3-14B-ft-limo/resolve/main/logo-min.png +base_model_relation: quantized +license: apache-2.0 +--- + +## Llamacpp imatrix Quantizations of Qwen3-14B-ft-limo by mrm8488 + +Using llama.cpp release b5490 for quantization. + +Original model: https://huggingface.co/mrm8488/Qwen3-14B-ft-limo + +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 + +``` +<|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 | +| -------- | ---------- | --------- | ----- | ----------- | +| [Qwen3-14B-ft-limo-bf16.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-bf16.gguf) | bf16 | 29.54GB | false | Full BF16 weights. | +| [Qwen3-14B-ft-limo-Q8_0.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q8_0.gguf) | Q8_0 | 15.70GB | false | Extremely high quality, generally unneeded but max available quant. | +| [Qwen3-14B-ft-limo-Q6_K_L.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q6_K_L.gguf) | Q6_K_L | 12.50GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. | +| [Qwen3-14B-ft-limo-Q6_K.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q6_K.gguf) | Q6_K | 12.12GB | false | Very high quality, near perfect, *recommended*. | +| [Qwen3-14B-ft-limo-Q5_K_L.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q5_K_L.gguf) | Q5_K_L | 10.99GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. | +| [Qwen3-14B-ft-limo-Q5_K_M.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q5_K_M.gguf) | Q5_K_M | 10.51GB | false | High quality, *recommended*. | +| [Qwen3-14B-ft-limo-Q5_K_S.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q5_K_S.gguf) | Q5_K_S | 10.26GB | false | High quality, *recommended*. | +| [Qwen3-14B-ft-limo-Q4_K_L.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q4_K_L.gguf) | Q4_K_L | 9.58GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. | +| [Qwen3-14B-ft-limo-Q4_1.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q4_1.gguf) | Q4_1 | 9.39GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. | +| [Qwen3-14B-ft-limo-Q4_K_M.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q4_K_M.gguf) | Q4_K_M | 9.00GB | false | Good quality, default size for most use cases, *recommended*. | +| [Qwen3-14B-ft-limo-Q3_K_XL.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q3_K_XL.gguf) | Q3_K_XL | 8.58GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. | +| [Qwen3-14B-ft-limo-Q4_K_S.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q4_K_S.gguf) | Q4_K_S | 8.57GB | false | Slightly lower quality with more space savings, *recommended*. | +| [Qwen3-14B-ft-limo-Q4_0.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q4_0.gguf) | Q4_0 | 8.54GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. | +| [Qwen3-14B-ft-limo-IQ4_NL.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-IQ4_NL.gguf) | IQ4_NL | 8.54GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. | +| [Qwen3-14B-ft-limo-IQ4_XS.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-IQ4_XS.gguf) | IQ4_XS | 8.11GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. | +| [Qwen3-14B-ft-limo-Q3_K_L.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q3_K_L.gguf) | Q3_K_L | 7.90GB | false | Lower quality but usable, good for low RAM availability. | +| [Qwen3-14B-ft-limo-Q3_K_M.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q3_K_M.gguf) | Q3_K_M | 7.32GB | false | Low quality. | +| [Qwen3-14B-ft-limo-IQ3_M.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-IQ3_M.gguf) | IQ3_M | 6.88GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. | +| [Qwen3-14B-ft-limo-Q3_K_S.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q3_K_S.gguf) | Q3_K_S | 6.66GB | false | Low quality, not recommended. | +| [Qwen3-14B-ft-limo-Q2_K_L.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q2_K_L.gguf) | Q2_K_L | 6.51GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. | +| [Qwen3-14B-ft-limo-IQ3_XS.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-IQ3_XS.gguf) | IQ3_XS | 6.38GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. | +| [Qwen3-14B-ft-limo-IQ3_XXS.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-IQ3_XXS.gguf) | IQ3_XXS | 5.94GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. | +| [Qwen3-14B-ft-limo-Q2_K.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-Q2_K.gguf) | Q2_K | 5.75GB | false | Very low quality but surprisingly usable. | +| [Qwen3-14B-ft-limo-IQ2_M.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-IQ2_M.gguf) | IQ2_M | 5.32GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. | +| [Qwen3-14B-ft-limo-IQ2_S.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-IQ2_S.gguf) | IQ2_S | 4.96GB | false | Low quality, uses SOTA techniques to be usable. | +| [Qwen3-14B-ft-limo-IQ2_XS.gguf](https://huggingface.co/bartowski/mrm8488_Qwen3-14B-ft-limo-GGUF/blob/main/mrm8488_Qwen3-14B-ft-limo-IQ2_XS.gguf) | IQ2_XS | 4.69GB | false | Low quality, uses SOTA techniques to be 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 + +
+ 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/mrm8488_Qwen3-14B-ft-limo-GGUF --include "mrm8488_Qwen3-14B-ft-limo-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/mrm8488_Qwen3-14B-ft-limo-GGUF --include "mrm8488_Qwen3-14B-ft-limo-Q8_0/*" --local-dir ./ +``` + +You can either specify a new local-dir (mrm8488_Qwen3-14B-ft-limo-Q8_0) or download them all in place (./) + +
+ +## 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. + +
+ 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 | 42.60 ± 0.31 | 105% | + +Q4_0_8_8 offers a nice bump to prompt processing and a small bump to text generation + +
+ +
+ +## Which file should I choose? + +
+ Click here for details + +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. + +
+ +## 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 diff --git a/configuration.json b/configuration.json new file mode 100644 index 0000000..bbeeda1 --- /dev/null +++ b/configuration.json @@ -0,0 +1 @@ +{"framework": "pytorch", "task": "text-generation", "allow_remote": true} \ No newline at end of file diff --git a/mrm8488_Qwen3-14B-ft-limo-IQ2_M.gguf b/mrm8488_Qwen3-14B-ft-limo-IQ2_M.gguf new file mode 100644 index 0000000..64cc62d --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-IQ2_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b98f574b8647ce535218be3e7b36332afa8260599021f7da40bbaa59b54c2440 +size 5322942112 diff --git a/mrm8488_Qwen3-14B-ft-limo-IQ2_S.gguf b/mrm8488_Qwen3-14B-ft-limo-IQ2_S.gguf new file mode 100644 index 0000000..00f6d85 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-IQ2_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:74e0f7f920c66de16e3e848c811a06d97c205384a5474fb7817ce55ab65c57b8 +size 4963313312 diff --git a/mrm8488_Qwen3-14B-ft-limo-IQ2_XS.gguf b/mrm8488_Qwen3-14B-ft-limo-IQ2_XS.gguf new file mode 100644 index 0000000..8c98f47 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-IQ2_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ad5309d79224b875009dc9e22a206275c05f604c87facf86b53e206bbf11a26 +size 4691589792 diff --git a/mrm8488_Qwen3-14B-ft-limo-IQ3_M.gguf b/mrm8488_Qwen3-14B-ft-limo-IQ3_M.gguf new file mode 100644 index 0000000..d6530b5 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f17b558b1a30d871653fe80a73a1cb93e517f1a9fa687f74e8543ecba22732cb +size 6883410592 diff --git a/mrm8488_Qwen3-14B-ft-limo-IQ3_XS.gguf b/mrm8488_Qwen3-14B-ft-limo-IQ3_XS.gguf new file mode 100644 index 0000000..433875a --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e49277d93093f3b08c265bedd59dfd9b225b9a89bbec40483a29f42a42d114c7 +size 6375301792 diff --git a/mrm8488_Qwen3-14B-ft-limo-IQ3_XXS.gguf b/mrm8488_Qwen3-14B-ft-limo-IQ3_XXS.gguf new file mode 100644 index 0000000..b83aa6f --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-IQ3_XXS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c75af3c5c7e687167f665da103ef5d00ac8c80b2c61ac34477c263aa29087509 +size 5942666912 diff --git a/mrm8488_Qwen3-14B-ft-limo-IQ4_NL.gguf b/mrm8488_Qwen3-14B-ft-limo-IQ4_NL.gguf new file mode 100644 index 0000000..dffdf56 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d81c309acbf6e906e8aff793c38119a9f7ba038bda5d79438ab703ac31bcae0c +size 8541363872 diff --git a/mrm8488_Qwen3-14B-ft-limo-IQ4_XS.gguf b/mrm8488_Qwen3-14B-ft-limo-IQ4_XS.gguf new file mode 100644 index 0000000..b648d8b --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ae75d5574466800c775815e37fb6253cb83a3420ca391efcdee6bcfc0985198b +size 8110730912 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q2_K.gguf b/mrm8488_Qwen3-14B-ft-limo-Q2_K.gguf new file mode 100644 index 0000000..d57de30 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:78bda35efa58eba25521b446c735fa52d98498bb9501cb09a8933d5a0b1384d6 +size 5753984672 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q2_K_L.gguf b/mrm8488_Qwen3-14B-ft-limo-Q2_K_L.gguf new file mode 100644 index 0000000..ed0a7fe --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q2_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b6070138e5268edd161a0d27581c621729706da89748e6ce067fa96919ecc8d4 +size 6513664672 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q3_K_L.gguf b/mrm8488_Qwen3-14B-ft-limo-Q3_K_L.gguf new file mode 100644 index 0000000..11c96ca --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:689f4ee12d8c2c5eb374a4e29dea0ebdd9b688d7ce6407b2fcb24970fbf3508b +size 7900652192 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q3_K_M.gguf b/mrm8488_Qwen3-14B-ft-limo-Q3_K_M.gguf new file mode 100644 index 0000000..a5f8fb7 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:63e982bce79456cc6a7510492d0cc3489681baf2ab1efa3eb6339ccb24c0527a +size 7321313952 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q3_K_S.gguf b/mrm8488_Qwen3-14B-ft-limo-Q3_K_S.gguf new file mode 100644 index 0000000..29be82b --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bd4043b0d4d2ac6ac334ebdee82506e213859f8ef9f10032b1fe824938033385 +size 6657106592 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q3_K_XL.gguf b/mrm8488_Qwen3-14B-ft-limo-Q3_K_XL.gguf new file mode 100644 index 0000000..46498ea --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q3_K_XL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:243c03ba504b1bddb0c4c6ef6684013a4b1268c1ff64b4d31abe1683088f6509 +size 8581325472 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q4_0.gguf b/mrm8488_Qwen3-14B-ft-limo-Q4_0.gguf new file mode 100644 index 0000000..71216e7 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aa66924af9a094ad7cd0056d2da70e579e200fab58b005462e811c7f2823545f +size 8543002272 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q4_1.gguf b/mrm8488_Qwen3-14B-ft-limo-Q4_1.gguf new file mode 100644 index 0000000..83f00b5 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:77695d4c39dcb38498eb21cd1ab4004bc81c6bdf160fd03a5d107566fa32c758 +size 9389522592 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q4_K_L.gguf b/mrm8488_Qwen3-14B-ft-limo-Q4_K_L.gguf new file mode 100644 index 0000000..1216b5f --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q4_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6c53e28ad5dbada4921703015acb5852c4565d0d16be4239545e645ffe61f291 +size 9579111072 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q4_K_M.gguf b/mrm8488_Qwen3-14B-ft-limo-Q4_K_M.gguf new file mode 100644 index 0000000..9072773 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19d6dfd4a470cb293ad5e96bd94689fa2d12d1024eac548479c2e64f967d5f00 +size 9001754272 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q4_K_S.gguf b/mrm8488_Qwen3-14B-ft-limo-Q4_K_S.gguf new file mode 100644 index 0000000..3420b90 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:988bf52b121d5158a834773b3bb547ed2801cff08784a9201a7385f98fa2c8dc +size 8573476512 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q5_K_L.gguf b/mrm8488_Qwen3-14B-ft-limo-Q5_K_L.gguf new file mode 100644 index 0000000..add4608 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q5_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:76d648fdce87abec81f3ff1d54a21078367b15b1860d63cc4ca05d553751b7ea +size 10994688672 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q5_K_M.gguf b/mrm8488_Qwen3-14B-ft-limo-Q5_K_M.gguf new file mode 100644 index 0000000..8be72b0 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:33d690e326daab01576ae921210ccc8651f87b476acc88caf01183161adb40d5 +size 10514570912 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q5_K_S.gguf b/mrm8488_Qwen3-14B-ft-limo-Q5_K_S.gguf new file mode 100644 index 0000000..4446835 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a865054dc6f11bf7a1442f03eb7004f91b2aeef6abfa641418500c8854a66d7d +size 10263895712 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q6_K.gguf b/mrm8488_Qwen3-14B-ft-limo-Q6_K.gguf new file mode 100644 index 0000000..2065156 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8df4f633dc566d70fc7c43ab3fd5959b09ef162107bb507a971817c884de2464 +size 12121938592 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q6_K_L.gguf b/mrm8488_Qwen3-14B-ft-limo-Q6_K_L.gguf new file mode 100644 index 0000000..49586d2 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q6_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:867e14ca4708b54a7106ef4755976f5a38799a86ddfba1681361d63b043d0090 +size 12498739872 diff --git a/mrm8488_Qwen3-14B-ft-limo-Q8_0.gguf b/mrm8488_Qwen3-14B-ft-limo-Q8_0.gguf new file mode 100644 index 0000000..fd55030 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f698c797cbcf4978f7d3d93a2584b6885916869a12a894bf4c8f2b2b94aa8e53 +size 15698535072 diff --git a/mrm8488_Qwen3-14B-ft-limo-bf16.gguf b/mrm8488_Qwen3-14B-ft-limo-bf16.gguf new file mode 100644 index 0000000..83dd3f1 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo-bf16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:83d76b3b533d78ae5f209ce608b156a079996386651ebf1195a87c98c0099b85 +size 29543424384 diff --git a/mrm8488_Qwen3-14B-ft-limo.imatrix b/mrm8488_Qwen3-14B-ft-limo.imatrix new file mode 100644 index 0000000..f21a6b5 --- /dev/null +++ b/mrm8488_Qwen3-14B-ft-limo.imatrix @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f4e758e82af9cf03504bbe28ee14afbd0f5133b26372282680dceeac4a5750f1 +size 7709778