From b8af0613daf6416a450dffeac81ca2a255361d58 Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Tue, 21 Jul 2026 18:56:06 +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/TheDrummer_Gemma-3-R1-12B-v1-GGUF Source: Original Platform --- .gitattributes | 47 ++++++ README.md | 165 ++++++++++++++++++++++ TheDrummer_Gemma-3-R1-12B-v1-IQ2_M.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-IQ2_S.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-IQ3_M.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-IQ3_XS.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-IQ3_XXS.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-IQ4_NL.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-IQ4_XS.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q2_K.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q2_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q3_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q3_K_M.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q3_K_S.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q3_K_XL.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q4_0.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q4_1.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q4_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q4_K_M.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q4_K_S.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q5_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q5_K_M.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q5_K_S.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q6_K.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q6_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-Q8_0.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-bf16.gguf | 3 + TheDrummer_Gemma-3-R1-12B-v1-imatrix.gguf | 3 + configuration.json | 1 + 29 files changed, 291 insertions(+) create mode 100644 .gitattributes create mode 100644 README.md create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-IQ2_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-IQ2_S.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-IQ3_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-IQ3_XS.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-IQ3_XXS.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-IQ4_NL.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-IQ4_XS.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q2_K.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q2_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q3_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q3_K_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q3_K_S.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q3_K_XL.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q4_0.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q4_1.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q4_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q4_K_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q4_K_S.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q5_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q5_K_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q5_K_S.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q6_K.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q6_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-Q8_0.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-bf16.gguf create mode 100644 TheDrummer_Gemma-3-R1-12B-v1-imatrix.gguf create mode 100644 configuration.json diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..53d7257 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,47 @@ +*.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 \ No newline at end of file diff --git a/README.md b/README.md new file mode 100644 index 0000000..fc54738 --- /dev/null +++ b/README.md @@ -0,0 +1,165 @@ +--- +quantized_by: bartowski +pipeline_tag: text-generation +base_model_relation: quantized +base_model: TheDrummer/Gemma-3-R1-12B-v1 +--- + +## Llamacpp imatrix Quantizations of Gemma-3-R1-12B-v1 by TheDrummer + +Using llama.cpp release b6139 for quantization. + +Original model: https://huggingface.co/TheDrummer/Gemma-3-R1-12B-v1 + +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/ggerganov/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 | +| -------- | ---------- | --------- | ----- | ----------- | +| [Gemma-3-R1-12B-v1-bf16.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-bf16.gguf) | bf16 | 23.54GB | false | Full BF16 weights. | +| [Gemma-3-R1-12B-v1-Q8_0.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q8_0.gguf) | Q8_0 | 12.51GB | false | Extremely high quality, generally unneeded but max available quant. | +| [Gemma-3-R1-12B-v1-Q6_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q6_K_L.gguf) | Q6_K_L | 9.90GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. | +| [Gemma-3-R1-12B-v1-Q6_K.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q6_K.gguf) | Q6_K | 9.66GB | false | Very high quality, near perfect, *recommended*. | +| [Gemma-3-R1-12B-v1-Q5_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_L.gguf) | Q5_K_L | 8.69GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. | +| [Gemma-3-R1-12B-v1-Q5_K_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_M.gguf) | Q5_K_M | 8.45GB | false | High quality, *recommended*. | +| [Gemma-3-R1-12B-v1-Q5_K_S.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_S.gguf) | Q5_K_S | 8.23GB | false | High quality, *recommended*. | +| [Gemma-3-R1-12B-v1-Q4_1.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q4_1.gguf) | Q4_1 | 7.56GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. | +| [Gemma-3-R1-12B-v1-Q4_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_L.gguf) | Q4_K_L | 7.54GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. | +| [Gemma-3-R1-12B-v1-Q4_K_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_M.gguf) | Q4_K_M | 7.30GB | false | Good quality, default size for most use cases, *recommended*. | +| [Gemma-3-R1-12B-v1-Q4_K_S.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_S.gguf) | Q4_K_S | 6.94GB | false | Slightly lower quality with more space savings, *recommended*. | +| [Gemma-3-R1-12B-v1-Q4_0.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q4_0.gguf) | Q4_0 | 6.91GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. | +| [Gemma-3-R1-12B-v1-IQ4_NL.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-IQ4_NL.gguf) | IQ4_NL | 6.89GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. | +| [Gemma-3-R1-12B-v1-Q3_K_XL.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_XL.gguf) | Q3_K_XL | 6.72GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. | +| [Gemma-3-R1-12B-v1-IQ4_XS.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-IQ4_XS.gguf) | IQ4_XS | 6.55GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. | +| [Gemma-3-R1-12B-v1-Q3_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_L.gguf) | Q3_K_L | 6.48GB | false | Lower quality but usable, good for low RAM availability. | +| [Gemma-3-R1-12B-v1-Q3_K_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_M.gguf) | Q3_K_M | 6.01GB | false | Low quality. | +| [Gemma-3-R1-12B-v1-IQ3_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-IQ3_M.gguf) | IQ3_M | 5.66GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. | +| [Gemma-3-R1-12B-v1-Q3_K_S.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_S.gguf) | Q3_K_S | 5.46GB | false | Low quality, not recommended. | +| [Gemma-3-R1-12B-v1-IQ3_XS.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-IQ3_XS.gguf) | IQ3_XS | 5.21GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. | +| [Gemma-3-R1-12B-v1-Q2_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q2_K_L.gguf) | Q2_K_L | 5.01GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. | +| [Gemma-3-R1-12B-v1-IQ3_XXS.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-IQ3_XXS.gguf) | IQ3_XXS | 4.78GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. | +| [Gemma-3-R1-12B-v1-Q2_K.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-Q2_K.gguf) | Q2_K | 4.77GB | false | Very low quality but surprisingly usable. | +| [Gemma-3-R1-12B-v1-IQ2_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-IQ2_M.gguf) | IQ2_M | 4.31GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. | +| [Gemma-3-R1-12B-v1-IQ2_S.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-12B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-12B-v1-IQ2_S.gguf) | IQ2_S | 4.02GB | 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/TheDrummer_Gemma-3-R1-12B-v1-GGUF --include "TheDrummer_Gemma-3-R1-12B-v1-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/TheDrummer_Gemma-3-R1-12B-v1-GGUF --include "TheDrummer_Gemma-3-R1-12B-v1-Q8_0/*" --local-dir ./ +``` + +You can either specify a new local-dir (TheDrummer_Gemma-3-R1-12B-v1-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/TheDrummer_Gemma-3-R1-12B-v1-IQ2_M.gguf b/TheDrummer_Gemma-3-R1-12B-v1-IQ2_M.gguf new file mode 100644 index 0000000..723892b --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-IQ2_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e16900b8f1e505dc04f93b836e65173349734cec8b93124c76ac2f7dc46b76ed +size 4310461120 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-IQ2_S.gguf b/TheDrummer_Gemma-3-R1-12B-v1-IQ2_S.gguf new file mode 100644 index 0000000..351f331 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-IQ2_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:90a6bd3476ada62e44131d2de4531c6432ecbf83d37b125f47f9a0d22548e0b8 +size 4020710080 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-IQ3_M.gguf b/TheDrummer_Gemma-3-R1-12B-v1-IQ3_M.gguf new file mode 100644 index 0000000..e47f86f --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19332c21e53b021533d5a7ec3d5ba7dc1ba2db2565554b23c2792e7353889aba +size 5655722560 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-IQ3_XS.gguf b/TheDrummer_Gemma-3-R1-12B-v1-IQ3_XS.gguf new file mode 100644 index 0000000..6b9f272 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:39318bf6c463ed7097228c2edf9f992adeb3f396a7585df9cd845fd80523c2a2 +size 5206166080 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-IQ3_XXS.gguf b/TheDrummer_Gemma-3-R1-12B-v1-IQ3_XXS.gguf new file mode 100644 index 0000000..bfceeff --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-IQ3_XXS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:69067b9437c783f362e60e85b111dd58625514884678af40e307d65e23da13de +size 4784900800 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-IQ4_NL.gguf b/TheDrummer_Gemma-3-R1-12B-v1-IQ4_NL.gguf new file mode 100644 index 0000000..ba0758b --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d8706eaa27512c62e34fbb1352d78d07bd4dbe3766e98e2c272a1c4b58e15386 +size 6887164480 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-IQ4_XS.gguf b/TheDrummer_Gemma-3-R1-12B-v1-IQ4_XS.gguf new file mode 100644 index 0000000..bbaa8c8 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:477c394a7ef9bc848233f9762f1413b348992587ec08a2806bb220048ab062b5 +size 6550964800 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q2_K.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q2_K.gguf new file mode 100644 index 0000000..f732b36 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b55b1e241b584b58a78e2d79c0d389730d334c0ed96852a98b87b03b4df0a700 +size 4768221760 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q2_K_L.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q2_K_L.gguf new file mode 100644 index 0000000..c9ccc42 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q2_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:83f5ac1c1d1c2692a1e427519f505b98f812775a677d8fb3aa13c2db35348ed4 +size 5012075200 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_L.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_L.gguf new file mode 100644 index 0000000..217d7f6 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6cf4bc015be26cfcbfbbdd45ea6233461afdc22077b392dd68d6bb1b2944bc96 +size 6480185920 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_M.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_M.gguf new file mode 100644 index 0000000..cdb7b6b --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6ad4f96214ce14d215cbdb6faffbd80663e80c3fcec51027891d76961f250921 +size 6008818240 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_S.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_S.gguf new file mode 100644 index 0000000..18756fa --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8100291b3b329d2f557e029bad906e00855125310100701310241984c81592b6 +size 5458315840 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_XL.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_XL.gguf new file mode 100644 index 0000000..f0535e4 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q3_K_XL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b3762bdcedfefac0c715c4a1a2d73c1dfcdc9063742e078dea1d5da87fd177d3 +size 6724039360 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q4_0.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q4_0.gguf new file mode 100644 index 0000000..c98f22f --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f8b2621d6a5b4fc32cb597415f6b45258bebda73e2f540a52caed25b99df9505 +size 6909282880 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q4_1.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q4_1.gguf new file mode 100644 index 0000000..eda84e9 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b559956e639406232c94765d76bac0b33a5d25c7e56a571ba40ba3ffa2a7bd09 +size 7559563840 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_L.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_L.gguf new file mode 100644 index 0000000..c555c6b --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fead3fe83d5731b3a1164f35847432712c9c9a389d7d6255eba0e35a8013c988 +size 7544632000 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_M.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_M.gguf new file mode 100644 index 0000000..903a9a3 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6517394bf14b85d6009e1ad8fd1fc6179fa3de3d091011cf14cacba1aee5b393 +size 7300778560 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_S.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_S.gguf new file mode 100644 index 0000000..2872048 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ddc662449c0d8abb82a6364fd8a45609722dadbc7e36b39d3cb3c4ae86f09fe8 +size 6935333440 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_L.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_L.gguf new file mode 100644 index 0000000..25bf69e --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eb31d601f678b2662138d6e8ffb800a30b54d187edc200f8ea5411bd6068bc29 +size 8688890560 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_M.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_M.gguf new file mode 100644 index 0000000..b864e9a --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:13bab8551332724fd03bdfec745187df780bd9f8661754907301e9b0fada7d38 +size 8445037120 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_S.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_S.gguf new file mode 100644 index 0000000..78523ad --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:53c23efb112dae4ccb8e8ec4620ee510316890bc87a6aca6fd0668c2bdf2caf0 +size 8231963200 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q6_K.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q6_K.gguf new file mode 100644 index 0000000..e8551a4 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0fafe7ddbbd34a127609df61b40029a1380315377b004cb8da6efdda1177b54d +size 9660811840 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q6_K_L.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q6_K_L.gguf new file mode 100644 index 0000000..a13e083 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q6_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e7fa5547515385d00b6971692be1e9e5797727138eb71e9bd29738235d61989 +size 9904665280 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-Q8_0.gguf b/TheDrummer_Gemma-3-R1-12B-v1-Q8_0.gguf new file mode 100644 index 0000000..fcfb899 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:98cec663fd4a3045a4af656399bf246bf4158fa1ce066138a7243942e4f0d504 +size 12510212800 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-bf16.gguf b/TheDrummer_Gemma-3-R1-12B-v1-bf16.gguf new file mode 100644 index 0000000..8dd8683 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-bf16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f805acb422d3236a5fe5f313d41fa78c888cc969d2dd1fb965fe673183d932eb +size 23540151712 diff --git a/TheDrummer_Gemma-3-R1-12B-v1-imatrix.gguf b/TheDrummer_Gemma-3-R1-12B-v1-imatrix.gguf new file mode 100644 index 0000000..fea8737 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-12B-v1-imatrix.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e0fbde185c035134ae4f1830085080263633bc5923fba101eb77c7fc2bca647 +size 7473632 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