From 297f2b8bcad3c99546b5984aa1d8fee995e948a5 Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Sun, 21 Jun 2026 02:51:15 +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/ibm-granite_granite-4.1-3b-GGUF Source: Original Platform --- .gitattributes | 85 +++++++++++ README.md | 180 ++++++++++++++++++++++++ configuration.json | 1 + ibm-granite_granite-4.1-3b-IQ2_M.gguf | 3 + ibm-granite_granite-4.1-3b-IQ3_M.gguf | 3 + ibm-granite_granite-4.1-3b-IQ3_XS.gguf | 3 + ibm-granite_granite-4.1-3b-IQ3_XXS.gguf | 3 + ibm-granite_granite-4.1-3b-IQ4_NL.gguf | 3 + ibm-granite_granite-4.1-3b-IQ4_XS.gguf | 3 + ibm-granite_granite-4.1-3b-Q2_K.gguf | 3 + ibm-granite_granite-4.1-3b-Q2_K_L.gguf | 3 + ibm-granite_granite-4.1-3b-Q3_K_L.gguf | 3 + ibm-granite_granite-4.1-3b-Q3_K_M.gguf | 3 + ibm-granite_granite-4.1-3b-Q3_K_S.gguf | 3 + ibm-granite_granite-4.1-3b-Q3_K_XL.gguf | 3 + ibm-granite_granite-4.1-3b-Q4_0.gguf | 3 + ibm-granite_granite-4.1-3b-Q4_1.gguf | 3 + ibm-granite_granite-4.1-3b-Q4_K_L.gguf | 3 + ibm-granite_granite-4.1-3b-Q4_K_M.gguf | 3 + ibm-granite_granite-4.1-3b-Q4_K_S.gguf | 3 + ibm-granite_granite-4.1-3b-Q5_K_L.gguf | 3 + ibm-granite_granite-4.1-3b-Q5_K_M.gguf | 3 + ibm-granite_granite-4.1-3b-Q5_K_S.gguf | 3 + ibm-granite_granite-4.1-3b-Q6_K.gguf | 3 + ibm-granite_granite-4.1-3b-Q6_K_L.gguf | 3 + ibm-granite_granite-4.1-3b-Q8_0.gguf | 3 + ibm-granite_granite-4.1-3b-bf16.gguf | 3 + ibm-granite_granite-4.1-3b-imatrix.gguf | 3 + 28 files changed, 341 insertions(+) create mode 100644 .gitattributes create mode 100644 README.md create mode 100644 configuration.json create mode 100644 ibm-granite_granite-4.1-3b-IQ2_M.gguf create mode 100644 ibm-granite_granite-4.1-3b-IQ3_M.gguf create mode 100644 ibm-granite_granite-4.1-3b-IQ3_XS.gguf create mode 100644 ibm-granite_granite-4.1-3b-IQ3_XXS.gguf create mode 100644 ibm-granite_granite-4.1-3b-IQ4_NL.gguf create mode 100644 ibm-granite_granite-4.1-3b-IQ4_XS.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q2_K.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q2_K_L.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q3_K_L.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q3_K_M.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q3_K_S.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q3_K_XL.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q4_0.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q4_1.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q4_K_L.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q4_K_M.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q4_K_S.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q5_K_L.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q5_K_M.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q5_K_S.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q6_K.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q6_K_L.gguf create mode 100644 ibm-granite_granite-4.1-3b-Q8_0.gguf create mode 100644 ibm-granite_granite-4.1-3b-bf16.gguf create mode 100644 ibm-granite_granite-4.1-3b-imatrix.gguf diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..c410f44 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,85 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bin.* filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zstandard filter=lfs diff=lfs merge=lfs -text +*.tfevents* filter=lfs diff=lfs merge=lfs -text +*.db* filter=lfs diff=lfs merge=lfs -text +*.ark* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text + +*.ggml filter=lfs diff=lfs merge=lfs -text +*.llamafile* filter=lfs diff=lfs merge=lfs -text +*.pt2 filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-IQ2_M.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-IQ3_XXS.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-IQ4_NL.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-Q2_K_L.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-Q3_K_XL.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-Q4_1.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-Q4_K_L.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-Q5_K_L.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-Q6_K_L.gguf filter=lfs diff=lfs merge=lfs -text + +ibm-granite_granite-4.1-3b-bf16.gguf filter=lfs diff=lfs merge=lfs -text +ibm-granite_granite-4.1-3b-imatrix.gguf 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..b23dae3 --- /dev/null +++ b/README.md @@ -0,0 +1,180 @@ +--- +quantized_by: bartowski +pipeline_tag: text-generation +base_model: ibm-granite/granite-4.1-3b +license: apache-2.0 +tags: +- language +- granite-4.1 +base_model_relation: quantized +--- + +## Llamacpp imatrix Quantizations of granite-4.1-3b by ibm-granite + +Using llama.cpp release b8970 for quantization. + +Original model: https://huggingface.co/ibm-granite/granite-4.1-3b + +All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/82ae9b520227f57d79ba04add13d0d0d) + +Run them in your choice of tools: + +- [llama.cpp](https://github.com/ggml-org/llama.cpp) +- [ramalama](https://github.com/containers/ramalama) +- [LM Studio](https://lmstudio.ai/) +- [koboldcpp](https://github.com/LostRuins/koboldcpp) +- [Jan AI](https://www.jan.ai/) +- [Text Generation Web UI](https://github.com/oobabooga/text-generation-webui) +- [LoLLMs](https://github.com/ParisNeo/lollms) + +Note: if it's a newly supported model, you may need to wait for an update from the developers. + +## Prompt format + +``` +<|start_of_role|>system<|end_of_role|>{system_prompt}<|end_of_text|> +<|start_of_role|>user<|end_of_role|>{prompt}<|end_of_text|> +<|start_of_role|>assistant<|end_of_role|> +``` + +## Download a file (not the whole branch) from below: + +| Filename | Quant type | File Size | Split | Description | +| -------- | ---------- | --------- | ----- | ----------- | +| [granite-4.1-3b-bf16.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-bf16.gguf) | bf16 | 6.81GB | false | Full BF16 weights. | +| [granite-4.1-3b-Q8_0.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q8_0.gguf) | Q8_0 | 3.62GB | false | Extremely high quality, generally unneeded but max available quant. | +| [granite-4.1-3b-Q6_K_L.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q6_K_L.gguf) | Q6_K_L | 2.97GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. | +| [granite-4.1-3b-Q6_K.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q6_K.gguf) | Q6_K | 2.91GB | false | Very high quality, near perfect, *recommended*. | +| [granite-4.1-3b-Q5_K_L.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q5_K_L.gguf) | Q5_K_L | 2.56GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. | +| [granite-4.1-3b-Q5_K_M.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q5_K_M.gguf) | Q5_K_M | 2.50GB | false | High quality, *recommended*. | +| [granite-4.1-3b-Q5_K_S.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q5_K_S.gguf) | Q5_K_S | 2.40GB | false | High quality, *recommended*. | +| [granite-4.1-3b-Q4_K_L.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q4_K_L.gguf) | Q4_K_L | 2.23GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. | +| [granite-4.1-3b-Q4_1.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q4_1.gguf) | Q4_1 | 2.21GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. | +| [granite-4.1-3b-Q4_K_M.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q4_K_M.gguf) | Q4_K_M | 2.17GB | false | Good quality, default size for most use cases, *recommended*. | +| [granite-4.1-3b-Q4_K_S.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q4_K_S.gguf) | Q4_K_S | 2.03GB | false | Slightly lower quality with more space savings, *recommended*. | +| [granite-4.1-3b-IQ4_NL.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-IQ4_NL.gguf) | IQ4_NL | 2.03GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. | +| [granite-4.1-3b-Q4_0.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q4_0.gguf) | Q4_0 | 2.02GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. | +| [granite-4.1-3b-IQ4_XS.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-IQ4_XS.gguf) | IQ4_XS | 1.94GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. | +| [granite-4.1-3b-Q3_K_XL.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q3_K_XL.gguf) | Q3_K_XL | 1.94GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. | +| [granite-4.1-3b-Q3_K_L.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q3_K_L.gguf) | Q3_K_L | 1.88GB | false | Lower quality but usable, good for low RAM availability. | +| [granite-4.1-3b-Q3_K_M.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q3_K_M.gguf) | Q3_K_M | 1.78GB | false | Low quality. | +| [granite-4.1-3b-IQ3_M.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-IQ3_M.gguf) | IQ3_M | 1.68GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. | +| [granite-4.1-3b-Q3_K_S.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q3_K_S.gguf) | Q3_K_S | 1.62GB | false | Low quality, not recommended. | +| [granite-4.1-3b-IQ3_XS.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-IQ3_XS.gguf) | IQ3_XS | 1.57GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. | +| [granite-4.1-3b-Q2_K_L.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q2_K_L.gguf) | Q2_K_L | 1.50GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. | +| [granite-4.1-3b-IQ3_XXS.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-IQ3_XXS.gguf) | IQ3_XXS | 1.46GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. | +| [granite-4.1-3b-Q2_K.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-Q2_K.gguf) | Q2_K | 1.44GB | false | Very low quality but surprisingly usable. | +| [granite-4.1-3b-IQ2_M.gguf](https://huggingface.co/bartowski/ibm-granite_granite-4.1-3b-GGUF/blob/main/ibm-granite_granite-4.1-3b-IQ2_M.gguf) | IQ2_M | 1.37GB | 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/ibm-granite_granite-4.1-3b-GGUF --include "ibm-granite_granite-4.1-3b-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/ibm-granite_granite-4.1-3b-GGUF --include "ibm-granite_granite-4.1-3b-Q8_0/*" --local-dir ./ +``` + +You can either specify a new local-dir (ibm-granite_granite-4.1-3b-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/ggml-org/llama.cpp/pull/9921). If you use Q4_0 and your hardware would benefit from repacking weights, it will do it automatically on the fly. + +As of llama.cpp build [b4282](https://github.com/ggml-org/llama.cpp/releases/tag/b4282) you will not be able to run the Q4_0_X_X files and will instead need to use Q4_0. + +Additionally, if you want to get slightly better quality for , you can use IQ4_NL thanks to [this PR](https://github.com/ggml-org/llama.cpp/pull/10541) which will also repack the weights for ARM, though only the 4_4 for now. The loading time may be slower but it will result in an overall speed incrase. + +
+ 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/ggml-org/llama.cpp/wiki/Feature-matrix) + +But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size. + +These I-quants can also be used on CPU, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide. + +
+ +## 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/ibm-granite_granite-4.1-3b-IQ2_M.gguf b/ibm-granite_granite-4.1-3b-IQ2_M.gguf new file mode 100644 index 0000000..f342451 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-IQ2_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6302ae7879bf27592b5098f2546c4df126b904f651f8c127b1c07588ca508736 +size 1374878688 diff --git a/ibm-granite_granite-4.1-3b-IQ3_M.gguf b/ibm-granite_granite-4.1-3b-IQ3_M.gguf new file mode 100644 index 0000000..c71617f --- /dev/null +++ b/ibm-granite_granite-4.1-3b-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d9aaecf34d70d9f7ba248acabb4fb17b6cc91350c2ba2c54faa26e9bdd4b5e0f +size 1680276448 diff --git a/ibm-granite_granite-4.1-3b-IQ3_XS.gguf b/ibm-granite_granite-4.1-3b-IQ3_XS.gguf new file mode 100644 index 0000000..201a412 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:427222a36e8d2342c33c261e3d0389246e94976463e0ea4fbbfa8a6a79b8d667 +size 1565998048 diff --git a/ibm-granite_granite-4.1-3b-IQ3_XXS.gguf b/ibm-granite_granite-4.1-3b-IQ3_XXS.gguf new file mode 100644 index 0000000..73204eb --- /dev/null +++ b/ibm-granite_granite-4.1-3b-IQ3_XXS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:61e6c3e7714ee159dcc15b21f317e4b88325c37f0ed35ba7692bfc95bafe437f +size 1464294368 diff --git a/ibm-granite_granite-4.1-3b-IQ4_NL.gguf b/ibm-granite_granite-4.1-3b-IQ4_NL.gguf new file mode 100644 index 0000000..7e6a2f0 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c493540614f0d88f15ff6d7729278ee8e37c5905a5dc770b58831b2b6b041237 +size 2031918048 diff --git a/ibm-granite_granite-4.1-3b-IQ4_XS.gguf b/ibm-granite_granite-4.1-3b-IQ4_XS.gguf new file mode 100644 index 0000000..ab4a47b --- /dev/null +++ b/ibm-granite_granite-4.1-3b-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:74d1d979776834e9f70b38a4b31c4319dddea35d7ee41d7e7affbe4c35872e17 +size 1939348448 diff --git a/ibm-granite_granite-4.1-3b-Q2_K.gguf b/ibm-granite_granite-4.1-3b-Q2_K.gguf new file mode 100644 index 0000000..cd64665 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e439ae7bc575481f624443a5587284d983550a8b6a784d0fd64ee2513b4bfc5c +size 1435950048 diff --git a/ibm-granite_granite-4.1-3b-Q2_K_L.gguf b/ibm-granite_granite-4.1-3b-Q2_K_L.gguf new file mode 100644 index 0000000..076552e --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q2_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:90d5bae2e041e08a7c4f491fba5a05df2adaa7501d19f10fd9aae8dd559f8db2 +size 1498168288 diff --git a/ibm-granite_granite-4.1-3b-Q3_K_L.gguf b/ibm-granite_granite-4.1-3b-Q3_K_L.gguf new file mode 100644 index 0000000..4a61b22 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9a2a0e1257595dc2abe3fd42eabdc5ccbe4cd24407952c73fadd3c86950e0d33 +size 1882618848 diff --git a/ibm-granite_granite-4.1-3b-Q3_K_M.gguf b/ibm-granite_granite-4.1-3b-Q3_K_M.gguf new file mode 100644 index 0000000..ca6e9c5 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6ee47b2e8c2e52108a1e93005b44fd78ce8245679ec7ae5fbcfec1655c7f844e +size 1779727328 diff --git a/ibm-granite_granite-4.1-3b-Q3_K_S.gguf b/ibm-granite_granite-4.1-3b-Q3_K_S.gguf new file mode 100644 index 0000000..db5ae61 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a11e38e6d2d39c06e217c33a9dd0fc9ab7ee0fcebf9c7aef37075b062eabcb62 +size 1618017248 diff --git a/ibm-granite_granite-4.1-3b-Q3_K_XL.gguf b/ibm-granite_granite-4.1-3b-Q3_K_XL.gguf new file mode 100644 index 0000000..023fcc3 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q3_K_XL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d158e2d393b4dce569b99a9243750c946fce14090b198aac2ae4dbac0cef193 +size 1944837088 diff --git a/ibm-granite_granite-4.1-3b-Q4_0.gguf b/ibm-granite_granite-4.1-3b-Q4_0.gguf new file mode 100644 index 0000000..4d56e07 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ccfd6ac2fc8bca6cb2fa1b8cb312aaa9200f13fe65b0cd745e8316e0093013fc +size 2024954848 diff --git a/ibm-granite_granite-4.1-3b-Q4_1.gguf b/ibm-granite_granite-4.1-3b-Q4_1.gguf new file mode 100644 index 0000000..331e2fc --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e21b9270252ab4c1060e5bb0a69e5ffe5c9a903cc0a5fe03718a0c5cd7b71547 +size 2206817248 diff --git a/ibm-granite_granite-4.1-3b-Q4_K_L.gguf b/ibm-granite_granite-4.1-3b-Q4_K_L.gguf new file mode 100644 index 0000000..4e54446 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q4_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e3ce8ab2f27fde22f37e8cbfa0fc47ad7947df6fc527ebe2b6a15317f29518a1 +size 2234833888 diff --git a/ibm-granite_granite-4.1-3b-Q4_K_M.gguf b/ibm-granite_granite-4.1-3b-Q4_K_M.gguf new file mode 100644 index 0000000..d2d8887 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d5c3d3d2442cd398331548df3e2f7059225bf097dfe03d9135586e31e43f9465 +size 2172615648 diff --git a/ibm-granite_granite-4.1-3b-Q4_K_S.gguf b/ibm-granite_granite-4.1-3b-Q4_K_S.gguf new file mode 100644 index 0000000..a0712f4 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ae3b44a037b04efbf49635e179ee92776aa8f418905cfad68e18656f4b1b45a1 +size 2034785248 diff --git a/ibm-granite_granite-4.1-3b-Q5_K_L.gguf b/ibm-granite_granite-4.1-3b-Q5_K_L.gguf new file mode 100644 index 0000000..fbcf4ed --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q5_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e463625f45d98043478efdf7fc8fa692f192b9861cc5af9e3d6f407fdd94a4e9 +size 2558417888 diff --git a/ibm-granite_granite-4.1-3b-Q5_K_M.gguf b/ibm-granite_granite-4.1-3b-Q5_K_M.gguf new file mode 100644 index 0000000..d086a95 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:49ed190d9f14d358906b833e9a4610a44065c4627c755942017f8af6510ca013 +size 2496199648 diff --git a/ibm-granite_granite-4.1-3b-Q5_K_S.gguf b/ibm-granite_granite-4.1-3b-Q5_K_S.gguf new file mode 100644 index 0000000..21137f3 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e1397c95644345202fe28fe49b1743545afd7b196fff974e9b8694b882c5db6 +size 2395233248 diff --git a/ibm-granite_granite-4.1-3b-Q6_K.gguf b/ibm-granite_granite-4.1-3b-Q6_K.gguf new file mode 100644 index 0000000..53c5f1b --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0371b947583f27dfca124efeda87e27856df85980c14fbfb28c1435023e71718 +size 2909895648 diff --git a/ibm-granite_granite-4.1-3b-Q6_K_L.gguf b/ibm-granite_granite-4.1-3b-Q6_K_L.gguf new file mode 100644 index 0000000..d15b6f4 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q6_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d8c7e777fbeac672728820bf9d0ee1f283fd2baeaaa3eccfab6020a4335fc0e +size 2972113888 diff --git a/ibm-granite_granite-4.1-3b-Q8_0.gguf b/ibm-granite_granite-4.1-3b-Q8_0.gguf new file mode 100644 index 0000000..011b252 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:69b3aef4cc576457417db283504dd67a721cdc3e3591538472011b3283ec725c +size 3619691488 diff --git a/ibm-granite_granite-4.1-3b-bf16.gguf b/ibm-granite_granite-4.1-3b-bf16.gguf new file mode 100644 index 0000000..694eca9 --- /dev/null +++ b/ibm-granite_granite-4.1-3b-bf16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:020b35c0e924b482333c1f5dc08cfcfb7d0c8d810598cb4a8ebfc6ee86829c64 +size 6809656000 diff --git a/ibm-granite_granite-4.1-3b-imatrix.gguf b/ibm-granite_granite-4.1-3b-imatrix.gguf new file mode 100644 index 0000000..aa7268d --- /dev/null +++ b/ibm-granite_granite-4.1-3b-imatrix.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:66bbaf7bfe7ab17fb162336a71e5668b2f179ae27c4dfa9034513850f1a40478 +size 3811424