From 5aa2d3947a6a19c07622af9b08986a8382abe282 Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Thu, 30 Jul 2026 07:47:14 +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_Fallen-Gemma3-12B-v1-GGUF Source: Original Platform --- .gitattributes | 49 ++++++ README.md | 174 +++++++++++++++++++ TheDrummer_Fallen-Gemma3-12B-v1-IQ2_M.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-IQ2_S.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-IQ3_M.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XS.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XXS.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-IQ4_NL.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-IQ4_XS.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q2_K.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q2_K_L.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_L.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_M.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_S.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_XL.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q4_0.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q4_1.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_L.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_M.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_S.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_L.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_M.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_S.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q6_K.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q6_K_L.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-Q8_0.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1-bf16.gguf | 3 + TheDrummer_Fallen-Gemma3-12B-v1.imatrix | 3 + configuration.json | 1 + 29 files changed, 302 insertions(+) create mode 100644 .gitattributes create mode 100644 README.md create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-IQ2_M.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-IQ2_S.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-IQ3_M.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XS.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XXS.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-IQ4_NL.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-IQ4_XS.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q2_K.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q2_K_L.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_L.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_M.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_S.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_XL.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q4_0.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q4_1.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_L.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_M.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_S.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_L.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_M.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_S.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q6_K.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q6_K_L.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-Q8_0.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1-bf16.gguf create mode 100644 TheDrummer_Fallen-Gemma3-12B-v1.imatrix create mode 100644 configuration.json diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..2370c57 --- /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 + +TheDrummer_Fallen-Gemma3-12B-v1.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..3e85b2c --- /dev/null +++ b/README.md @@ -0,0 +1,174 @@ +--- +quantized_by: bartowski +pipeline_tag: text-generation +license: other +base_model_relation: quantized +base_model: TheDrummer/Fallen-Gemma3-12B-v1 +--- + +## Llamacpp imatrix Quantizations of Fallen-Gemma3-12B-v1 by TheDrummer + +Using llama.cpp release b4925 for quantization. + +Original model: https://huggingface.co/TheDrummer/Fallen-Gemma3-12B-v1 + +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 + +``` +user +{system_prompt} + +{prompt} +model + +model +``` + +## Download a file (not the whole branch) from below: + +| Filename | Quant type | File Size | Split | Description | +| -------- | ---------- | --------- | ----- | ----------- | +| [Fallen-Gemma3-12B-v1-bf16.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-bf16.gguf) | bf16 | 23.54GB | false | Full BF16 weights. | +| [Fallen-Gemma3-12B-v1-Q8_0.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q8_0.gguf) | Q8_0 | 12.51GB | false | Extremely high quality, generally unneeded but max available quant. | +| [Fallen-Gemma3-12B-v1-Q6_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-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*. | +| [Fallen-Gemma3-12B-v1-Q6_K.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q6_K.gguf) | Q6_K | 9.66GB | false | Very high quality, near perfect, *recommended*. | +| [Fallen-Gemma3-12B-v1-Q5_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_L.gguf) | Q5_K_L | 8.69GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. | +| [Fallen-Gemma3-12B-v1-Q5_K_M.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_M.gguf) | Q5_K_M | 8.45GB | false | High quality, *recommended*. | +| [Fallen-Gemma3-12B-v1-Q5_K_S.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_S.gguf) | Q5_K_S | 8.23GB | false | High quality, *recommended*. | +| [Fallen-Gemma3-12B-v1-Q4_1.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-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. | +| [Fallen-Gemma3-12B-v1-Q4_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_L.gguf) | Q4_K_L | 7.54GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. | +| [Fallen-Gemma3-12B-v1-Q4_K_M.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_M.gguf) | Q4_K_M | 7.30GB | false | Good quality, default size for most use cases, *recommended*. | +| [Fallen-Gemma3-12B-v1-Q4_K_S.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_S.gguf) | Q4_K_S | 6.94GB | false | Slightly lower quality with more space savings, *recommended*. | +| [Fallen-Gemma3-12B-v1-Q4_0.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q4_0.gguf) | Q4_0 | 6.91GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. | +| [Fallen-Gemma3-12B-v1-IQ4_NL.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-IQ4_NL.gguf) | IQ4_NL | 6.89GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. | +| [Fallen-Gemma3-12B-v1-Q3_K_XL.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-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. | +| [Fallen-Gemma3-12B-v1-IQ4_XS.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-IQ4_XS.gguf) | IQ4_XS | 6.55GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. | +| [Fallen-Gemma3-12B-v1-Q3_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_L.gguf) | Q3_K_L | 6.48GB | false | Lower quality but usable, good for low RAM availability. | +| [Fallen-Gemma3-12B-v1-Q3_K_M.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_M.gguf) | Q3_K_M | 6.01GB | false | Low quality. | +| [Fallen-Gemma3-12B-v1-IQ3_M.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_M.gguf) | IQ3_M | 5.66GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. | +| [Fallen-Gemma3-12B-v1-Q3_K_S.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_S.gguf) | Q3_K_S | 5.46GB | false | Low quality, not recommended. | +| [Fallen-Gemma3-12B-v1-IQ3_XS.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XS.gguf) | IQ3_XS | 5.21GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. | +| [Fallen-Gemma3-12B-v1-Q2_K_L.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-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. | +| [Fallen-Gemma3-12B-v1-IQ3_XXS.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XXS.gguf) | IQ3_XXS | 4.78GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. | +| [Fallen-Gemma3-12B-v1-Q2_K.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-Q2_K.gguf) | Q2_K | 4.77GB | false | Very low quality but surprisingly usable. | +| [Fallen-Gemma3-12B-v1-IQ2_M.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-v1-IQ2_M.gguf) | IQ2_M | 4.31GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. | +| [Fallen-Gemma3-12B-v1-IQ2_S.gguf](https://huggingface.co/bartowski/TheDrummer_Fallen-Gemma3-12B-v1-GGUF/blob/main/TheDrummer_Fallen-Gemma3-12B-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_Fallen-Gemma3-12B-v1-GGUF --include "TheDrummer_Fallen-Gemma3-12B-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_Fallen-Gemma3-12B-v1-GGUF --include "TheDrummer_Fallen-Gemma3-12B-v1-Q8_0/*" --local-dir ./ +``` + +You can either specify a new local-dir (TheDrummer_Fallen-Gemma3-12B-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_Fallen-Gemma3-12B-v1-IQ2_M.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-IQ2_M.gguf new file mode 100644 index 0000000..908133b --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-IQ2_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:efba304bc958fa985f839255d117a519a3216e82d70c7054a6c5af0671474a7d +size 4310460864 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-IQ2_S.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-IQ2_S.gguf new file mode 100644 index 0000000..5090e66 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-IQ2_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6120e9ed3be609edcffa64ee186f984b82a474878d036518789941f9ca7e83f4 +size 4020709824 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_M.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_M.gguf new file mode 100644 index 0000000..ba21b62 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:936a466a41e4edd2c23117bbbba23b24919d1e5bd0a80dec40907467ff7eea5e +size 5655722304 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XS.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XS.gguf new file mode 100644 index 0000000..93a4cf0 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a5047218d65874dc2340a54c4732266a3f2d90fdac4ede23c693e911cce433aa +size 5206165824 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XXS.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XXS.gguf new file mode 100644 index 0000000..7585ce2 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-IQ3_XXS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:917783afdf17dd8669cff3370f6dc777ec2ebc37900a3a80ffc95f2517c9bccf +size 4784900544 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-IQ4_NL.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-IQ4_NL.gguf new file mode 100644 index 0000000..7cb0cf3 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2bd901da54b794d96e70d31d26bd350c38e8a26a7df62c120f39f0a608b53f34 +size 6887164224 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-IQ4_XS.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-IQ4_XS.gguf new file mode 100644 index 0000000..ed6e9b7 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:64f1e12434494350bc253e5f760f6533c632d66bf9e7be96328b16eb3ee71f03 +size 6550964544 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q2_K.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q2_K.gguf new file mode 100644 index 0000000..5fb0c14 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4edb8fa538f526512658ada726416f9b8dba184c0e4dba02cbcb5a0f93ab4814 +size 4768221504 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q2_K_L.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q2_K_L.gguf new file mode 100644 index 0000000..07d25f4 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q2_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2595881f8db9bbc02c5215d7a2f573878810f757cf35c0ecebe07080d2807534 +size 5012074944 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_L.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_L.gguf new file mode 100644 index 0000000..e555475 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9b3833dcb37615298051028f61fcb79bb888a3a223576c33849881cdae7ad16e +size 6480185664 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_M.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_M.gguf new file mode 100644 index 0000000..0327187 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:763e92a177593240b4b8f72effd083583e390d2ef172a62d65cd4dafe47e717d +size 6008817984 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_S.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_S.gguf new file mode 100644 index 0000000..45cad62 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a4078be6eaffe14c6dccd9ed4cebae5a63674b77e5ee6fa0d2eb6c7855c2367a +size 5458315584 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_XL.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_XL.gguf new file mode 100644 index 0000000..22ce43e --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q3_K_XL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:483adf73ecec482fc7484e69fdd17afcd0d97b79e2b9b56d24c6fe8baa4bf9ed +size 6724039104 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q4_0.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_0.gguf new file mode 100644 index 0000000..e23c55d --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:31f84c32fb8c9eebf12a95266c56830ba1ccd40da779227c289a6faf3440c6e6 +size 6909282624 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q4_1.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_1.gguf new file mode 100644 index 0000000..966fe41 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ec5c56b660aaff41674782c2b50cc5f465f81ca3e83d00d40e7b0ca8d91e6ff +size 7559563584 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_L.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_L.gguf new file mode 100644 index 0000000..cd89231 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b78a3f1b58d7c70c265a4b04df6d506d7e8309bbdb679965f023bca77ca6e9cd +size 7544631744 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_M.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_M.gguf new file mode 100644 index 0000000..097f139 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8b5ff6cf6cd68688fa50c29e7b3c15c3f31c5c4794fff2dd71c9ca5a3d05cff3 +size 7300778304 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_S.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_S.gguf new file mode 100644 index 0000000..3d69fa5 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f67089ef26e5abdbcba235aa6d88ef86fd8e8a2c960cd9e027eb816ad334d8e8 +size 6935333184 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_L.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_L.gguf new file mode 100644 index 0000000..985e738 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:366b05d9961c344c197e20822d0b4464b10268188d3d34a1a95809d3f57f9887 +size 8688890304 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_M.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_M.gguf new file mode 100644 index 0000000..64e066e --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f0c55602c741d25c727937b7f4dca4f633955dc6fc5a18659b3c3adfcbaee9c2 +size 8445036864 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_S.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_S.gguf new file mode 100644 index 0000000..8521bbc --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c13a4b5abce99f782b044953e2e78e90b616e72fa01aee296e633ecae6ed567 +size 8231962944 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q6_K.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q6_K.gguf new file mode 100644 index 0000000..445c08c --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a8c9efa9f36e658054a62038af468fe974ce93567fbb4bbc651fad1e7bd6a903 +size 9660811584 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q6_K_L.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q6_K_L.gguf new file mode 100644 index 0000000..f7a1402 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q6_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7ba2a888a24e0e04e6d2fb9e461bd4789e31cc96b1c64ed65ed0281111b77b47 +size 9904665024 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-Q8_0.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-Q8_0.gguf new file mode 100644 index 0000000..c9bd23f --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dc9c2c4afa1a425bd54c37df00d21a00c37f7e0891d06e77f6f125f94f508fe1 +size 12510212544 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1-bf16.gguf b/TheDrummer_Fallen-Gemma3-12B-v1-bf16.gguf new file mode 100644 index 0000000..80d6c0a --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1-bf16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:24f23b0fc07e102e656bdf51b117cf43dc91f250e48e2059b8c4c6eab34e7e41 +size 23540151456 diff --git a/TheDrummer_Fallen-Gemma3-12B-v1.imatrix b/TheDrummer_Fallen-Gemma3-12B-v1.imatrix new file mode 100644 index 0000000..ff51309 --- /dev/null +++ b/TheDrummer_Fallen-Gemma3-12B-v1.imatrix @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:df2ffb905be26a14eb635e3617dfc104c909453b65a9b56886320cf92e6098fd +size 7433114 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