commit 6d6684ce9db79cfe8924c9b9d4e47729e4f607e8 Author: ModelHub XC Date: Mon Apr 13 11:33:36 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: s3dev-ai/Mistral-Small-3.1-24B-Instruct-2503-gguf Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..15cb732 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,42 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt 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 +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz 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 +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl 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 +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* 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 +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text +Mistral-Small-3.1-24B-Instruct-2503-BF16.gguf filter=lfs diff=lfs merge=lfs -text +Mistral-Small-3.1-24B-Instruct-2503-F32.gguf filter=lfs diff=lfs merge=lfs -text +Mistral-Small-3.1-24B-Instruct-2503-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text +Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Mistral-Small-3.1-24B-Instruct-2503-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Mistral-Small-3.1-24B-Instruct-2503-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text +Mistral-Small-3.1-24B-Instruct-2503-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text diff --git a/Mistral-Small-3.1-24B-Instruct-2503-BF16.gguf b/Mistral-Small-3.1-24B-Instruct-2503-BF16.gguf new file mode 100644 index 0000000..1228475 --- /dev/null +++ b/Mistral-Small-3.1-24B-Instruct-2503-BF16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff48641c1ad66d82d19c829a529cb91ac9e31faa5dcd26ab9945c60275a2acff +size 47153520544 diff --git a/Mistral-Small-3.1-24B-Instruct-2503-F32.gguf b/Mistral-Small-3.1-24B-Instruct-2503-F32.gguf new file mode 100644 index 0000000..0868deb --- /dev/null +++ b/Mistral-Small-3.1-24B-Instruct-2503-F32.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2e5c191a5c3d9eb2763ac9d9b6d07fb8387fb547d1ce4ac1fd4a1e76fa3f8e94 +size 94297497504 diff --git a/Mistral-Small-3.1-24B-Instruct-2503-Q2_K.gguf b/Mistral-Small-3.1-24B-Instruct-2503-Q2_K.gguf new file mode 100644 index 0000000..dd83071 --- /dev/null +++ b/Mistral-Small-3.1-24B-Instruct-2503-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3519950ecb65ecd287ee1fad0c1a388ba80abcb7f0503f296d24f162bcf921e4 +size 8890326944 diff --git a/Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf b/Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf new file mode 100644 index 0000000..3645ea7 --- /dev/null +++ b/Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d6817796bc933559eedac5a2accc73c8e4f80114b94ac79895c66d47c04255e1 +size 14333910944 diff --git a/Mistral-Small-3.1-24B-Instruct-2503-Q5_K_M.gguf b/Mistral-Small-3.1-24B-Instruct-2503-Q5_K_M.gguf new file mode 100644 index 0000000..b103a62 --- /dev/null +++ b/Mistral-Small-3.1-24B-Instruct-2503-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:93df130a393dc813ec4ab21e477f68b99bc98e1a3134e14865d88264b68e86d7 +size 16763985824 diff --git a/Mistral-Small-3.1-24B-Instruct-2503-Q6_K.gguf b/Mistral-Small-3.1-24B-Instruct-2503-Q6_K.gguf new file mode 100644 index 0000000..f1a792f --- /dev/null +++ b/Mistral-Small-3.1-24B-Instruct-2503-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2e61611be0486baa5ab240f32fb3e0e6c13e256b4eba06d65622c7b7ae68d68e +size 19345940384 diff --git a/Mistral-Small-3.1-24B-Instruct-2503-Q8_0.gguf b/Mistral-Small-3.1-24B-Instruct-2503-Q8_0.gguf new file mode 100644 index 0000000..9d80981 --- /dev/null +++ b/Mistral-Small-3.1-24B-Instruct-2503-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bfbfea11a9cab94d11266938064f7927f5533ddd99ad51940621383281d4be0a +size 25054781344 diff --git a/README.md b/README.md new file mode 100644 index 0000000..c287f85 --- /dev/null +++ b/README.md @@ -0,0 +1,77 @@ +--- +base_model: + - mistralai/Mistral-Small-3.1-24B-Instruct-2503 +language: + - en +model_creator: Mistral AI +model_name: Mistral-Small-3.1-24B-Instruct-2503 +model_type: llama +quantized_by: s3dev-ai +tags: + - text-generation +--- + +# Overview + +This model repository provides various quantisations of the following [base model](https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503), in GGUF format. +- mistralai/Mistral-Small-3.1-24B-Instruct-2503 + +# Model Description + +For a full model description, please refer to the [base model's](https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503) card. + +This model, and subsequent quantisations, have been converted directly from the author's base model *unaltered*. + +## How are the GGUF files created? +After cloning the author's original base model repository, [`llama.cpp`](https://github.com/ggml-org/llama.cpp) is used to convert the model to GGUF format, using `--outtype=f32` to preserve the original model's 32-bit fidelity. + +Finally, for each subsequent quantisation level, `llama.cpp`'s `llama-quantize` executable is called using the F32 GGUF file as the source file. + +# Quantisation +The purpose of this repository is to provide *unaltered* quantisations of the author's base model. This section is designed to help the user visualise the difference in quantisation levels, in efforts to assist in model (quantisation) selection. + +## Comparison Statistics +To aid a user in model/quantisation selection, the team has created the following statistics specifically for comparing the similarity scores across quantisation runs. + +The dataset against which each run was conducted is composed of 175 question/answer pairs, divided amongst 7 topics, specifically designed to test a quantisation's processing ability. The test dataset was created by Mistral Large (via [Le Chat](https://chat.mistral.ai/chat)) using prompts explicitly stating the requirement for the question/answer pairs to be designed for Mistral model quantisation testing. + +The similarity scores used by these statistics were calculated as the cosine similarity between the embedding of the 'gold standard' answer provided in the dataset, and the embedding of the response from the quantised model. The embedding model used in these tests is the [all-MiniLM-L6-v2 Q8_0](https://huggingface.co/s3dev-ai/all-MiniLM-L6-v2-gguf). + +### Range +The range graph below illustrates how the range of similarity scores varies amongst the quantisation levels. Included in the range stats are the: + +- Minimum scores +- Maximum scores +- Mean scores +- Score distribution (KDE) +- Outliers + + +
+ Stats Graph: Range +
+ +### Mean +The mean graph below illustrates how the mean similarity scores (when grouped by 'topic') vary amongst the quantisation levels. + + +
+ Stats Graph: Mean +
+ +### Standard Deviation +The standard deviation graph below illustrates the how spread of similarity scores vary amongst the quantisation levels, when grouped by the test dataset's 'topic' categories. + + +
+ Stats Graph: StdDev +
+ +### Kernel Density Estimate +The KDE graph below illustrates the how distribution of similarity scores vary amongst the quantisation levels. + + +
+ Stats Graph: KDE +
+ diff --git a/imgs/small-3.1-kde.png b/imgs/small-3.1-kde.png new file mode 100644 index 0000000..ef49115 Binary files /dev/null and b/imgs/small-3.1-kde.png differ diff --git a/imgs/small-3.1-mean.png b/imgs/small-3.1-mean.png new file mode 100644 index 0000000..f0bac83 Binary files /dev/null and b/imgs/small-3.1-mean.png differ diff --git a/imgs/small-3.1-range.png b/imgs/small-3.1-range.png new file mode 100644 index 0000000..fe6a350 Binary files /dev/null and b/imgs/small-3.1-range.png differ diff --git a/imgs/small-3.1-stddev.png b/imgs/small-3.1-stddev.png new file mode 100644 index 0000000..412b07e Binary files /dev/null and b/imgs/small-3.1-stddev.png differ diff --git a/sha256/Mistral-Small-3.1-24B-Instruct-2503-BF16.sha256 b/sha256/Mistral-Small-3.1-24B-Instruct-2503-BF16.sha256 new file mode 100644 index 0000000..7064c1b --- /dev/null +++ b/sha256/Mistral-Small-3.1-24B-Instruct-2503-BF16.sha256 @@ -0,0 +1 @@ +ff48641c1ad66d82d19c829a529cb91ac9e31faa5dcd26ab9945c60275a2acff ./Mistral-Small-3.1-24B-Instruct-2503-BF16.gguf diff --git a/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q2_K.sha256 b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q2_K.sha256 new file mode 100644 index 0000000..f98f31c --- /dev/null +++ b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q2_K.sha256 @@ -0,0 +1 @@ +3519950ecb65ecd287ee1fad0c1a388ba80abcb7f0503f296d24f162bcf921e4 ./Mistral-Small-3.1-24B-Instruct-2503-Q2_K.gguf diff --git a/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.sha256 b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.sha256 new file mode 100644 index 0000000..684c5ea --- /dev/null +++ b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.sha256 @@ -0,0 +1 @@ +d6817796bc933559eedac5a2accc73c8e4f80114b94ac79895c66d47c04255e1 ./Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf diff --git a/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q5_K_M.sha256 b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q5_K_M.sha256 new file mode 100644 index 0000000..0108198 --- /dev/null +++ b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q5_K_M.sha256 @@ -0,0 +1 @@ +93df130a393dc813ec4ab21e477f68b99bc98e1a3134e14865d88264b68e86d7 ./Mistral-Small-3.1-24B-Instruct-2503-Q5_K_M.gguf diff --git a/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q6_K.sha256 b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q6_K.sha256 new file mode 100644 index 0000000..5f6fa27 --- /dev/null +++ b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q6_K.sha256 @@ -0,0 +1 @@ +2e61611be0486baa5ab240f32fb3e0e6c13e256b4eba06d65622c7b7ae68d68e ./Mistral-Small-3.1-24B-Instruct-2503-Q6_K.gguf diff --git a/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q8_0.sha256 b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q8_0.sha256 new file mode 100644 index 0000000..7334e5a --- /dev/null +++ b/sha256/Mistral-Small-3.1-24B-Instruct-2503-Q8_0.sha256 @@ -0,0 +1 @@ +bfbfea11a9cab94d11266938064f7927f5533ddd99ad51940621383281d4be0a ./Mistral-Small-3.1-24B-Instruct-2503-Q8_0.gguf