From bdedf27bc50a97a82da9c37e844fc402223bb860 Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Thu, 18 Jun 2026 00:08:13 +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/Llama-3.1_OpenScholar-8B-GGUF Source: Original Platform --- .gitattributes | 60 +++++++++ Llama-3.1_OpenScholar-8B-IQ2_M.gguf | 3 + Llama-3.1_OpenScholar-8B-IQ3_M.gguf | 3 + Llama-3.1_OpenScholar-8B-IQ3_XS.gguf | 3 + Llama-3.1_OpenScholar-8B-IQ4_XS.gguf | 3 + Llama-3.1_OpenScholar-8B-Q2_K.gguf | 3 + Llama-3.1_OpenScholar-8B-Q2_K_L.gguf | 3 + Llama-3.1_OpenScholar-8B-Q3_K_L.gguf | 3 + Llama-3.1_OpenScholar-8B-Q3_K_M.gguf | 3 + Llama-3.1_OpenScholar-8B-Q3_K_S.gguf | 3 + Llama-3.1_OpenScholar-8B-Q3_K_XL.gguf | 3 + Llama-3.1_OpenScholar-8B-Q4_0.gguf | 3 + Llama-3.1_OpenScholar-8B-Q4_0_4_4.gguf | 3 + Llama-3.1_OpenScholar-8B-Q4_0_4_8.gguf | 3 + Llama-3.1_OpenScholar-8B-Q4_0_8_8.gguf | 3 + Llama-3.1_OpenScholar-8B-Q4_K_L.gguf | 3 + Llama-3.1_OpenScholar-8B-Q4_K_M.gguf | 3 + Llama-3.1_OpenScholar-8B-Q4_K_S.gguf | 3 + Llama-3.1_OpenScholar-8B-Q5_K_L.gguf | 3 + Llama-3.1_OpenScholar-8B-Q5_K_M.gguf | 3 + Llama-3.1_OpenScholar-8B-Q5_K_S.gguf | 3 + Llama-3.1_OpenScholar-8B-Q6_K.gguf | 3 + Llama-3.1_OpenScholar-8B-Q6_K_L.gguf | 3 + Llama-3.1_OpenScholar-8B-Q8_0.gguf | 3 + Llama-3.1_OpenScholar-8B-f16.gguf | 3 + Llama-3.1_OpenScholar-8B.imatrix | 3 + README.md | 170 +++++++++++++++++++++++++ configuration.json | 1 + 28 files changed, 306 insertions(+) create mode 100644 .gitattributes create mode 100644 Llama-3.1_OpenScholar-8B-IQ2_M.gguf create mode 100644 Llama-3.1_OpenScholar-8B-IQ3_M.gguf create mode 100644 Llama-3.1_OpenScholar-8B-IQ3_XS.gguf create mode 100644 Llama-3.1_OpenScholar-8B-IQ4_XS.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q2_K.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q2_K_L.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q3_K_L.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q3_K_M.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q3_K_S.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q3_K_XL.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q4_0.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q4_0_4_4.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q4_0_4_8.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q4_0_8_8.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q4_K_L.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q4_K_M.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q4_K_S.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q5_K_L.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q5_K_M.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q5_K_S.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q6_K.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q6_K_L.gguf create mode 100644 Llama-3.1_OpenScholar-8B-Q8_0.gguf create mode 100644 Llama-3.1_OpenScholar-8B-f16.gguf create mode 100644 Llama-3.1_OpenScholar-8B.imatrix create mode 100644 README.md create mode 100644 configuration.json diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..75471e5 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,60 @@ +*.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 +Llama-3.1_OpenScholar-8B-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q6_K_L.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q5_K_L.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q4_K_L.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q4_0_8_8.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q4_0_4_8.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q4_0_4_4.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q3_K_XL.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q2_K_L.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-IQ2_M.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B-f16.gguf filter=lfs diff=lfs merge=lfs -text +Llama-3.1_OpenScholar-8B.imatrix filter=lfs diff=lfs merge=lfs -text diff --git a/Llama-3.1_OpenScholar-8B-IQ2_M.gguf b/Llama-3.1_OpenScholar-8B-IQ2_M.gguf new file mode 100644 index 0000000..851197d --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-IQ2_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6057fde113b867a0a68bfd46dff19af40024d527b507bc1892b13380302158d4 +size 2948281568 diff --git a/Llama-3.1_OpenScholar-8B-IQ3_M.gguf b/Llama-3.1_OpenScholar-8B-IQ3_M.gguf new file mode 100644 index 0000000..8042752 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f95f97f5374e8bbba66d823a28c43c3363039842605fff68094b1a5f5a69bd68 +size 3784824032 diff --git a/Llama-3.1_OpenScholar-8B-IQ3_XS.gguf b/Llama-3.1_OpenScholar-8B-IQ3_XS.gguf new file mode 100644 index 0000000..e376990 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:735b3a8c0c515816fcb91f478644749d3c9822e25175870f86f0641d144af02d +size 3518747872 diff --git a/Llama-3.1_OpenScholar-8B-IQ4_XS.gguf b/Llama-3.1_OpenScholar-8B-IQ4_XS.gguf new file mode 100644 index 0000000..c1dfe72 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:135bdf4328da919026a0e7dabfa1b02a841f391aa84c134f1ab3583a212ecd96 +size 4447663328 diff --git a/Llama-3.1_OpenScholar-8B-Q2_K.gguf b/Llama-3.1_OpenScholar-8B-Q2_K.gguf new file mode 100644 index 0000000..18bfb2e --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fa4b4bda43d7ba616b8d03592d1ef5e7d0c71ddb45b4508ea37cbdb7e92900fe +size 3179132128 diff --git a/Llama-3.1_OpenScholar-8B-Q2_K_L.gguf b/Llama-3.1_OpenScholar-8B-Q2_K_L.gguf new file mode 100644 index 0000000..b5f964e --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q2_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bded69f1ab3aafc2ae97164391a13f67e217c4e54a4a6686f07c76919e0ba80d +size 3692156128 diff --git a/Llama-3.1_OpenScholar-8B-Q3_K_L.gguf b/Llama-3.1_OpenScholar-8B-Q3_K_L.gguf new file mode 100644 index 0000000..ce0c1ec --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:00d203a93c1931a9d1ddc03e480e04c7b0cb44d0917e3b7b8bb6c9f4276cbf0e +size 4321957088 diff --git a/Llama-3.1_OpenScholar-8B-Q3_K_M.gguf b/Llama-3.1_OpenScholar-8B-Q3_K_M.gguf new file mode 100644 index 0000000..fbd715f --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:85d707d88fb251f5407862a4503a3dd9c0c67cba74663d212341761ece3df692 +size 4018918624 diff --git a/Llama-3.1_OpenScholar-8B-Q3_K_S.gguf b/Llama-3.1_OpenScholar-8B-Q3_K_S.gguf new file mode 100644 index 0000000..d35e4ed --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0a7e880aaedad98c31a47b9be6f28405e1728651a6f15629a30f4384e3fb7a12 +size 3664499936 diff --git a/Llama-3.1_OpenScholar-8B-Q3_K_XL.gguf b/Llama-3.1_OpenScholar-8B-Q3_K_XL.gguf new file mode 100644 index 0000000..540b339 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q3_K_XL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:265b371f27beedcf007b71fe20799e8dda7ced3ee29c97f8519baabd27e0cdd6 +size 4781626592 diff --git a/Llama-3.1_OpenScholar-8B-Q4_0.gguf b/Llama-3.1_OpenScholar-8B-Q4_0.gguf new file mode 100644 index 0000000..1ae8962 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b7f90f14359870e5e7d83fb66f16246c40bb3d4d9ae0bd6f5e8844b8afbd5c1f +size 4675892448 diff --git a/Llama-3.1_OpenScholar-8B-Q4_0_4_4.gguf b/Llama-3.1_OpenScholar-8B-Q4_0_4_4.gguf new file mode 100644 index 0000000..19c4aa7 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q4_0_4_4.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:56947117a3b89ea0f524bb2be5d5921879df4421286067858c04faa459e0ec3c +size 4661212384 diff --git a/Llama-3.1_OpenScholar-8B-Q4_0_4_8.gguf b/Llama-3.1_OpenScholar-8B-Q4_0_4_8.gguf new file mode 100644 index 0000000..781a454 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q4_0_4_8.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6de7180d7a4e325c62e61fd58e9385e343a4972b57a44ab5dfc2fdfe079d8724 +size 4661212384 diff --git a/Llama-3.1_OpenScholar-8B-Q4_0_8_8.gguf b/Llama-3.1_OpenScholar-8B-Q4_0_8_8.gguf new file mode 100644 index 0000000..d2d9824 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q4_0_8_8.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dd19da00e7413b4dcd54470bafb6c7accfadc760694efafb400e30c8fb947052 +size 4661212384 diff --git a/Llama-3.1_OpenScholar-8B-Q4_K_L.gguf b/Llama-3.1_OpenScholar-8B-Q4_K_L.gguf new file mode 100644 index 0000000..872a757 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q4_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:330c70a197df7659629eb3da3bb74a3273b12be654e391e740f8bea290847388 +size 5310633184 diff --git a/Llama-3.1_OpenScholar-8B-Q4_K_M.gguf b/Llama-3.1_OpenScholar-8B-Q4_K_M.gguf new file mode 100644 index 0000000..a7fa77e --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:54865fc86451959b495c494a51bb1806c8b62bf1415600f0da2966a8a1fe6c7d +size 4920734944 diff --git a/Llama-3.1_OpenScholar-8B-Q4_K_S.gguf b/Llama-3.1_OpenScholar-8B-Q4_K_S.gguf new file mode 100644 index 0000000..4135941 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4f6c5929cbdf4f91f7dbefce8d2b38e04f8af108039c8a0b976b8e59813f90ee +size 4692669664 diff --git a/Llama-3.1_OpenScholar-8B-Q5_K_L.gguf b/Llama-3.1_OpenScholar-8B-Q5_K_L.gguf new file mode 100644 index 0000000..93c033c --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q5_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:75b66fd4d56551964688f3cb00f83006646adfe8b4d4b24a44146b117f392a5a +size 6057219296 diff --git a/Llama-3.1_OpenScholar-8B-Q5_K_M.gguf b/Llama-3.1_OpenScholar-8B-Q5_K_M.gguf new file mode 100644 index 0000000..73d51a1 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c1584e2b3a7e54075b5b5cfbcc9f3094e1c41e4583c1572e6605bd25c30d5088 +size 5732988128 diff --git a/Llama-3.1_OpenScholar-8B-Q5_K_S.gguf b/Llama-3.1_OpenScholar-8B-Q5_K_S.gguf new file mode 100644 index 0000000..64a602c --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9bdf0656a6ff9dc0941535ee741f6ef9c374116903684061c63f3d1ea7238b35 +size 5599294688 diff --git a/Llama-3.1_OpenScholar-8B-Q6_K.gguf b/Llama-3.1_OpenScholar-8B-Q6_K.gguf new file mode 100644 index 0000000..dff121e --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cbcc7de869a79f1a7e440b015f7b6beffab5e9414ac5735ecd7d04e48fb07332 +size 6596007136 diff --git a/Llama-3.1_OpenScholar-8B-Q6_K_L.gguf b/Llama-3.1_OpenScholar-8B-Q6_K_L.gguf new file mode 100644 index 0000000..25056bf --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q6_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:33fcd9e20b3099567f055816ca0bf50772f32444fa122be2612cb7eaec4c419a +size 6850467040 diff --git a/Llama-3.1_OpenScholar-8B-Q8_0.gguf b/Llama-3.1_OpenScholar-8B-Q8_0.gguf new file mode 100644 index 0000000..ea3a060 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:50d74c013670c0de36876a6968846a531f5ffbe45c6a191b1eddd7536177fb9c +size 8540771552 diff --git a/Llama-3.1_OpenScholar-8B-f16.gguf b/Llama-3.1_OpenScholar-8B-f16.gguf new file mode 100644 index 0000000..b13533f --- /dev/null +++ b/Llama-3.1_OpenScholar-8B-f16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ac9e80906e24484ae385f9b84525918557693de927a52bdfe1d8e81569c0fb53 +size 16068891584 diff --git a/Llama-3.1_OpenScholar-8B.imatrix b/Llama-3.1_OpenScholar-8B.imatrix new file mode 100644 index 0000000..fed7f32 --- /dev/null +++ b/Llama-3.1_OpenScholar-8B.imatrix @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ddf3d108cec6c109e5e792aca36f13481d07a93b8db06f45ee6d2037a2a007ac +size 4988170 diff --git a/README.md b/README.md new file mode 100644 index 0000000..a61c933 --- /dev/null +++ b/README.md @@ -0,0 +1,170 @@ +--- +quantized_by: bartowski +pipeline_tag: text-generation +base_model: OpenScholar/Llama-3.1_OpenScholar-8B +license: apache-2.0 +language: +- en +--- + +## Llamacpp imatrix Quantizations of Llama-3.1_OpenScholar-8B + +Using llama.cpp release b4132 for quantization. + +Original model: https://huggingface.co/OpenScholar/Llama-3.1_OpenScholar-8B + +All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8) + +Run them in [LM Studio](https://lmstudio.ai/) + +## Prompt format + +``` +<|begin_of_text|><|start_header_id|>system<|end_header_id|> + +{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|> + +{prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|> + + +``` + +## Download a file (not the whole branch) from below: + +| Filename | Quant type | File Size | Split | Description | +| -------- | ---------- | --------- | ----- | ----------- | +| [Llama-3.1_OpenScholar-8B-f16.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-f16.gguf) | f16 | 16.07GB | false | Full F16 weights. | +| [Llama-3.1_OpenScholar-8B-Q8_0.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q8_0.gguf) | Q8_0 | 8.54GB | false | Extremely high quality, generally unneeded but max available quant. | +| [Llama-3.1_OpenScholar-8B-Q6_K_L.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q6_K_L.gguf) | Q6_K_L | 6.85GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. | +| [Llama-3.1_OpenScholar-8B-Q6_K.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q6_K.gguf) | Q6_K | 6.60GB | false | Very high quality, near perfect, *recommended*. | +| [Llama-3.1_OpenScholar-8B-Q5_K_L.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q5_K_L.gguf) | Q5_K_L | 6.06GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. | +| [Llama-3.1_OpenScholar-8B-Q5_K_M.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q5_K_M.gguf) | Q5_K_M | 5.73GB | false | High quality, *recommended*. | +| [Llama-3.1_OpenScholar-8B-Q5_K_S.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q5_K_S.gguf) | Q5_K_S | 5.60GB | false | High quality, *recommended*. | +| [Llama-3.1_OpenScholar-8B-Q4_K_L.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q4_K_L.gguf) | Q4_K_L | 5.31GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. | +| [Llama-3.1_OpenScholar-8B-Q4_K_M.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q4_K_M.gguf) | Q4_K_M | 4.92GB | false | Good quality, default size for most use cases, *recommended*. | +| [Llama-3.1_OpenScholar-8B-Q3_K_XL.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q3_K_XL.gguf) | Q3_K_XL | 4.78GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. | +| [Llama-3.1_OpenScholar-8B-Q4_K_S.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q4_K_S.gguf) | Q4_K_S | 4.69GB | false | Slightly lower quality with more space savings, *recommended*. | +| [Llama-3.1_OpenScholar-8B-Q4_0.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q4_0.gguf) | Q4_0 | 4.68GB | false | Legacy format, generally not worth using over similarly sized formats | +| [Llama-3.1_OpenScholar-8B-Q4_0_8_8.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q4_0_8_8.gguf) | Q4_0_8_8 | 4.66GB | false | Optimized for ARM and AVX inference. Requires 'sve' support for ARM (see details below). *Don't use on Mac*. | +| [Llama-3.1_OpenScholar-8B-Q4_0_4_8.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q4_0_4_8.gguf) | Q4_0_4_8 | 4.66GB | false | Optimized for ARM inference. Requires 'i8mm' support (see details below). *Don't use on Mac*. | +| [Llama-3.1_OpenScholar-8B-Q4_0_4_4.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q4_0_4_4.gguf) | Q4_0_4_4 | 4.66GB | false | Optimized for ARM inference. Should work well on all ARM chips, not for use with GPUs. *Don't use on Mac*. | +| [Llama-3.1_OpenScholar-8B-IQ4_XS.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-IQ4_XS.gguf) | IQ4_XS | 4.45GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. | +| [Llama-3.1_OpenScholar-8B-Q3_K_L.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q3_K_L.gguf) | Q3_K_L | 4.32GB | false | Lower quality but usable, good for low RAM availability. | +| [Llama-3.1_OpenScholar-8B-Q3_K_M.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q3_K_M.gguf) | Q3_K_M | 4.02GB | false | Low quality. | +| [Llama-3.1_OpenScholar-8B-IQ3_M.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-IQ3_M.gguf) | IQ3_M | 3.78GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. | +| [Llama-3.1_OpenScholar-8B-Q2_K_L.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q2_K_L.gguf) | Q2_K_L | 3.69GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. | +| [Llama-3.1_OpenScholar-8B-Q3_K_S.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q3_K_S.gguf) | Q3_K_S | 3.66GB | false | Low quality, not recommended. | +| [Llama-3.1_OpenScholar-8B-IQ3_XS.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-IQ3_XS.gguf) | IQ3_XS | 3.52GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. | +| [Llama-3.1_OpenScholar-8B-Q2_K.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-Q2_K.gguf) | Q2_K | 3.18GB | false | Very low quality but surprisingly usable. | +| [Llama-3.1_OpenScholar-8B-IQ2_M.gguf](https://huggingface.co/bartowski/Llama-3.1_OpenScholar-8B-GGUF/blob/main/Llama-3.1_OpenScholar-8B-IQ2_M.gguf) | IQ2_M | 2.95GB | 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/Llama-3.1_OpenScholar-8B-GGUF --include "Llama-3.1_OpenScholar-8B-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/Llama-3.1_OpenScholar-8B-GGUF --include "Llama-3.1_OpenScholar-8B-Q8_0/*" --local-dir ./ +``` + +You can either specify a new local-dir (Llama-3.1_OpenScholar-8B-Q8_0) or download them all in place (./) + +
+ +## Q4_0_X_X information + + +
+ Click to view Q4_0_X_X information +These are *NOT* for Metal (Apple) or GPU (nvidia/AMD/intel) offloading, only ARM chips (and certain AVX2/AVX512 CPUs). + +If you're using an ARM chip, the Q4_0_X_X quants will have a substantial speedup. Check out Q4_0_4_4 speed comparisons [on the original pull request](https://github.com/ggerganov/llama.cpp/pull/5780#pullrequestreview-21657544660) + +To check which one would work best for your ARM chip, you can check [AArch64 SoC features](https://gpages.juszkiewicz.com.pl/arm-socs-table/arm-socs.html) (thanks EloyOn!). + +If you're using a CPU that supports AVX2 or AVX512 (typically server CPUs and AMD's latest Zen5 CPUs) and are not offloading to a GPU, the Q4_0_8_8 may offer a nice speed as well: + +
+ 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 and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide. + +The I-quants are *not* compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm. + +
+ +## 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. + +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