From cd14844da2776d488c63e70b000fd6a5ee9b54d7 Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Sat, 18 Jul 2026 09:24:06 +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-R1-4B-v1-GGUF Source: Original Platform --- .gitattributes | 47 +++++++ README.md | 164 +++++++++++++++++++++++ TheDrummer_Gemma-3-R1-4B-v1-IQ2_M.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-IQ3_M.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-IQ3_XS.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-IQ3_XXS.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-IQ4_NL.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-IQ4_XS.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q2_K.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q2_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q3_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q3_K_M.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q3_K_S.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q3_K_XL.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q4_0.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q4_1.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q4_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q4_K_M.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q4_K_S.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q5_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q5_K_M.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q5_K_S.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q6_K.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q6_K_L.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-Q8_0.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-bf16.gguf | 3 + TheDrummer_Gemma-3-R1-4B-v1-imatrix.gguf | 3 + configuration.json | 1 + 28 files changed, 287 insertions(+) create mode 100644 .gitattributes create mode 100644 README.md create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-IQ2_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-IQ3_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-IQ3_XS.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-IQ3_XXS.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-IQ4_NL.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-IQ4_XS.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q2_K.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q2_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q3_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q3_K_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q3_K_S.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q3_K_XL.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q4_0.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q4_1.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q4_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q4_K_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q4_K_S.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q5_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q5_K_M.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q5_K_S.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q6_K.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q6_K_L.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-Q8_0.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-v1-bf16.gguf create mode 100644 TheDrummer_Gemma-3-R1-4B-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..64387e8 --- /dev/null +++ b/README.md @@ -0,0 +1,164 @@ +--- +quantized_by: bartowski +pipeline_tag: text-generation +base_model_relation: quantized +base_model: TheDrummer/Gemma-3-R1-4B-v1 +--- + +## Llamacpp imatrix Quantizations of Gemma-3-R1-4B-v1 by TheDrummer + +Using llama.cpp release b6139 for quantization. + +Original model: https://huggingface.co/TheDrummer/Gemma-3-R1-4B-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-R1-4B-v1-bf16.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-bf16.gguf) | bf16 | 7.77GB | false | Full BF16 weights. | +| [Gemma-3-R1-4B-v1-Q8_0.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q8_0.gguf) | Q8_0 | 4.13GB | false | Extremely high quality, generally unneeded but max available quant. | +| [Gemma-3-R1-4B-v1-Q6_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q6_K_L.gguf) | Q6_K_L | 3.35GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. | +| [Gemma-3-R1-4B-v1-Q6_K.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q6_K.gguf) | Q6_K | 3.19GB | false | Very high quality, near perfect, *recommended*. | +| [Gemma-3-R1-4B-v1-Q5_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_L.gguf) | Q5_K_L | 2.99GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. | +| [Gemma-3-R1-4B-v1-Q5_K_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_M.gguf) | Q5_K_M | 2.83GB | false | High quality, *recommended*. | +| [Gemma-3-R1-4B-v1-Q5_K_S.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_S.gguf) | Q5_K_S | 2.76GB | false | High quality, *recommended*. | +| [Gemma-3-R1-4B-v1-Q4_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_L.gguf) | Q4_K_L | 2.65GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. | +| [Gemma-3-R1-4B-v1-Q4_1.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q4_1.gguf) | Q4_1 | 2.56GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. | +| [Gemma-3-R1-4B-v1-Q4_K_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_M.gguf) | Q4_K_M | 2.49GB | false | Good quality, default size for most use cases, *recommended*. | +| [Gemma-3-R1-4B-v1-Q3_K_XL.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_XL.gguf) | Q3_K_XL | 2.40GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. | +| [Gemma-3-R1-4B-v1-Q4_K_S.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_S.gguf) | Q4_K_S | 2.38GB | false | Slightly lower quality with more space savings, *recommended*. | +| [Gemma-3-R1-4B-v1-Q4_0.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q4_0.gguf) | Q4_0 | 2.37GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. | +| [Gemma-3-R1-4B-v1-IQ4_NL.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-IQ4_NL.gguf) | IQ4_NL | 2.36GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. | +| [Gemma-3-R1-4B-v1-IQ4_XS.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-IQ4_XS.gguf) | IQ4_XS | 2.26GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. | +| [Gemma-3-R1-4B-v1-Q3_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_L.gguf) | Q3_K_L | 2.24GB | false | Lower quality but usable, good for low RAM availability. | +| [Gemma-3-R1-4B-v1-Q3_K_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_M.gguf) | Q3_K_M | 2.10GB | false | Low quality. | +| [Gemma-3-R1-4B-v1-IQ3_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-IQ3_M.gguf) | IQ3_M | 1.99GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. | +| [Gemma-3-R1-4B-v1-Q3_K_S.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_S.gguf) | Q3_K_S | 1.94GB | false | Low quality, not recommended. | +| [Gemma-3-R1-4B-v1-Q2_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q2_K_L.gguf) | Q2_K_L | 1.89GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. | +| [Gemma-3-R1-4B-v1-IQ3_XS.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-IQ3_XS.gguf) | IQ3_XS | 1.86GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. | +| [Gemma-3-R1-4B-v1-Q2_K.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-Q2_K.gguf) | Q2_K | 1.73GB | false | Very low quality but surprisingly usable. | +| [Gemma-3-R1-4B-v1-IQ3_XXS.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-IQ3_XXS.gguf) | IQ3_XXS | 1.69GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. | +| [Gemma-3-R1-4B-v1-IQ2_M.gguf](https://huggingface.co/bartowski/TheDrummer_Gemma-3-R1-4B-v1-GGUF/blob/main/TheDrummer_Gemma-3-R1-4B-v1-IQ2_M.gguf) | IQ2_M | 1.54GB | 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 + +
+ 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-R1-4B-v1-GGUF --include "TheDrummer_Gemma-3-R1-4B-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-R1-4B-v1-GGUF --include "TheDrummer_Gemma-3-R1-4B-v1-Q8_0/*" --local-dir ./ +``` + +You can either specify a new local-dir (TheDrummer_Gemma-3-R1-4B-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-R1-4B-v1-IQ2_M.gguf b/TheDrummer_Gemma-3-R1-4B-v1-IQ2_M.gguf new file mode 100644 index 0000000..4eb23d6 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-IQ2_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c73fecf48de68fe96dd8d5491444d5792ac09a0157625496c0ea2ad3bf5b3e7c +size 1537982528 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-IQ3_M.gguf b/TheDrummer_Gemma-3-R1-4B-v1-IQ3_M.gguf new file mode 100644 index 0000000..c289557 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5e62a07d8e6259599827ece3039ee8f8818f7ad41a431874b69081938b74aa9b +size 1986803008 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-IQ3_XS.gguf b/TheDrummer_Gemma-3-R1-4B-v1-IQ3_XS.gguf new file mode 100644 index 0000000..5a5af17 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:780d58066fdc5936c04c605f5b13997b009e6c53ee89c0de0876770da91828ec +size 1863390528 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-IQ3_XXS.gguf b/TheDrummer_Gemma-3-R1-4B-v1-IQ3_XXS.gguf new file mode 100644 index 0000000..bc4ecca --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-IQ3_XXS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f221eb7d9de9af59216b645a9d84cb421896c5dc952a44cb2314f3eadb0bb85b +size 1689452608 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-IQ4_NL.gguf b/TheDrummer_Gemma-3-R1-4B-v1-IQ4_NL.gguf new file mode 100644 index 0000000..98d0f65 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f633c34e019a24f610105a62002cad8bde4be38113dd0288bec38c64ea4937dc +size 2363512128 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-IQ4_XS.gguf b/TheDrummer_Gemma-3-R1-4B-v1-IQ4_XS.gguf new file mode 100644 index 0000000..5aa2273 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a81778b332048b807bd4f414969022eb0a1ae5ce0b46f07deb9607bb3894a431 +size 2263242048 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q2_K.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q2_K.gguf new file mode 100644 index 0000000..539a175 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1d3d4d5ffbab2e036b729f008232f13e8b4adfdb45f1edc5abfaec8ad102099d +size 1729164608 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q2_K_L.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q2_K_L.gguf new file mode 100644 index 0000000..30f0451 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q2_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7ff091ad022f426d7da63bc6600f5c53a24ce1e074cb80d376a9f62f5a72d8e7 +size 1891733568 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_L.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_L.gguf new file mode 100644 index 0000000..ad1a7b2 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fb97a305ed7b5b2c44a98b13dbefbf017ceb3fe4b5ac5016ddf0a7b1e302fc21 +size 2236085568 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_M.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_M.gguf new file mode 100644 index 0000000..3570dc0 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e23384b013fd7d7a4769f2fda63e81c0fc1f5d2c4e592c9034ec3f85b0ef4f43 +size 2098459968 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_S.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_S.gguf new file mode 100644 index 0000000..577f08f --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:407335bcc087e8ac9ebc2c02c320868b13f0c6475b2e8d1a8472017bc41bca11 +size 1937364288 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_XL.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_XL.gguf new file mode 100644 index 0000000..51db05a --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q3_K_XL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e464098b91afe316bbe04f8c3948a954d4e511f585a9bd50a63ad1ad8905ea8b +size 2398654528 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q4_0.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q4_0.gguf new file mode 100644 index 0000000..9e3b76a --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fe7bf82ecde40fe0fc1eb8d499589676b9a30585a352ad54a67e77edff4e6298 +size 2370065728 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q4_1.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q4_1.gguf new file mode 100644 index 0000000..7972cd2 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:31054d37772600780e1a91b22d980ad995602ac7d133190208749bdb455622c2 +size 2564052288 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_L.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_L.gguf new file mode 100644 index 0000000..bf80229 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e291fd612d630390f4f5365fb4fccfb58886a431ad849f95df3ba05444a03d7e +size 2652463168 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_M.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_M.gguf new file mode 100644 index 0000000..1f9c001 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:72a7dc5bddbdf6bbea0d47aea8573d6baa191f4ddebd75547091c991678bcd08 +size 2489894208 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_S.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_S.gguf new file mode 100644 index 0000000..c5a1006 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:36274277f077ea30bc50a3b7a5aa2f7be206c849484c37463e437d9a46b14189 +size 2377930048 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_L.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_L.gguf new file mode 100644 index 0000000..0d779bb --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b8836459220db3301ec5574dc3f8a77f37264613c55a7ccf21eaef3c0c4fb2da +size 2992267328 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_M.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_M.gguf new file mode 100644 index 0000000..b855e69 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ce2a08687bb6009fd99633e82f1eda1899ba43790390194e345a01ec8356399f +size 2829698368 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_S.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_S.gguf new file mode 100644 index 0000000..0c5a14e --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b35e60995b6339b856f08654dce8d38a4705fcbc193d8c8766dbce435fba16e +size 2764592448 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q6_K.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q6_K.gguf new file mode 100644 index 0000000..8c5d59c --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:03447a9a331a934592dfdea2414931110448c0850b3d88d2d2eff84e580c623a +size 3190740288 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q6_K_L.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q6_K_L.gguf new file mode 100644 index 0000000..c7dc48d --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q6_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:59d3377e6589dcf7881ad8cde4eebf40bc34aafe31d5e63a259b8611227d429a +size 3353309248 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-Q8_0.gguf b/TheDrummer_Gemma-3-R1-4B-v1-Q8_0.gguf new file mode 100644 index 0000000..f36aa3a --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:58bd8e78955f42012d7d39e3458bd10f792bd8bbcabf95bceda8c113b248483b +size 4130402368 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-bf16.gguf b/TheDrummer_Gemma-3-R1-4B-v1-bf16.gguf new file mode 100644 index 0000000..49a6fa1 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-bf16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fce7b7cb85c09b03e3bbf1e4e389bae495386a7ec90f0da1a9b6277a5ff0859e +size 7767803680 diff --git a/TheDrummer_Gemma-3-R1-4B-v1-imatrix.gguf b/TheDrummer_Gemma-3-R1-4B-v1-imatrix.gguf new file mode 100644 index 0000000..46f20c5 --- /dev/null +++ b/TheDrummer_Gemma-3-R1-4B-v1-imatrix.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3e250c8ae830c08880b7a7db72df4f13c75b58dce81b7574390aecc82ec3b537 +size 3448608 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