commit 2dbaaac5ae83313328550dd8879ed7546e43462a Author: ModelHub XC Date: Mon Jun 22 02:21:13 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: Mungert/OlympicCoder-32B-GGUF Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..0843bb1 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,84 @@ +*.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 + +OlympicCoder-32B-q4_k_s.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-f16-q4_k.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-f16-q6_k.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq3_s.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq3_m.gguf filter=lfs diff=lfs merge=lfs -text +f16/OlympicCoder-32B-F16-00002-of-00002.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq4_xs.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq1_m.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq1_s.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq2_m.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq2_s.gguf filter=lfs diff=lfs merge=lfs -text +bf16/OlympicCoder-32B-bf16-00002-of-00002.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-bf16-q6_k.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q4_1.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq2_xxs.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq3_xs.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q2_k_s.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q3_k_s.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-f16-q8_0.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q8_0.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q3_k_m.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq3_xxs.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q4_k_m.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q4_0.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq4_nl.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q5_1.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B.imatrix filter=lfs diff=lfs merge=lfs -text +bf16/OlympicCoder-32B-bf16-00001-of-00002.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q5_k_m.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q5_0.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-bf16-q4_k.gguf filter=lfs diff=lfs merge=lfs -text +f16/OlympicCoder-32B-F16-00001-of-00002.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-bf16-q8_0.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-iq2_xs.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q6_k_m.gguf filter=lfs diff=lfs merge=lfs -text +OlympicCoder-32B-q5_k_s.gguf filter=lfs diff=lfs merge=lfs -text \ No newline at end of file diff --git a/OlympicCoder-32B-bf16-q4_k.gguf b/OlympicCoder-32B-bf16-q4_k.gguf new file mode 100644 index 0000000..c6f94cd --- /dev/null +++ b/OlympicCoder-32B-bf16-q4_k.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:161112e20be3e65a2ad4a0b5aeebaba5a9e59f1217036e4ca869d14bd7685e18 +size 21888994592 diff --git a/OlympicCoder-32B-bf16-q6_k.gguf b/OlympicCoder-32B-bf16-q6_k.gguf new file mode 100644 index 0000000..d81a5d8 --- /dev/null +++ b/OlympicCoder-32B-bf16-q6_k.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:07f9ab56d315f6d29b92a1194f583df923e67885eebf68e766a088a52516d9b1 +size 28723088672 diff --git a/OlympicCoder-32B-bf16-q8_0.gguf b/OlympicCoder-32B-bf16-q8_0.gguf new file mode 100644 index 0000000..dc0c8c2 --- /dev/null +++ b/OlympicCoder-32B-bf16-q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d449581410c3be1dbc9776db4b523a5124c0d242761c3e2bca2da14aa992fbe2 +size 36280699904 diff --git a/OlympicCoder-32B-f16-q4_k.gguf b/OlympicCoder-32B-f16-q4_k.gguf new file mode 100644 index 0000000..50d1207 --- /dev/null +++ b/OlympicCoder-32B-f16-q4_k.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8a790add72008bfd45a49c04975227f5f203ae97e0f1a07d87f1af5771cf8f5a +size 21888994592 diff --git a/OlympicCoder-32B-f16-q6_k.gguf b/OlympicCoder-32B-f16-q6_k.gguf new file mode 100644 index 0000000..3b0851d --- /dev/null +++ b/OlympicCoder-32B-f16-q6_k.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c95beed9c6ebee1adea00bdb313642a92cf1b75936c09f2437bd96cbf504563d +size 28723088672 diff --git a/OlympicCoder-32B-f16-q8_0.gguf b/OlympicCoder-32B-f16-q8_0.gguf new file mode 100644 index 0000000..4aee907 --- /dev/null +++ b/OlympicCoder-32B-f16-q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fa72b8aee5c7efc14025324fb66abab5631a8a7c9af431d239cc2fb4811b65fe +size 36280699904 diff --git a/OlympicCoder-32B-iq1_m.gguf b/OlympicCoder-32B-iq1_m.gguf new file mode 100644 index 0000000..92bfda9 --- /dev/null +++ b/OlympicCoder-32B-iq1_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d3ed7ccd843ae1432d277d67f6432b68e0dc6ac298968edd0f46cb1d357fd8e6 +size 9220558112 diff --git a/OlympicCoder-32B-iq1_s.gguf b/OlympicCoder-32B-iq1_s.gguf new file mode 100644 index 0000000..b97cd57 --- /dev/null +++ b/OlympicCoder-32B-iq1_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eaf26c60aca9f707420d798a1ed59ef6db9ec6873eda45d26c7fde884ede1247 +size 8748698912 diff --git a/OlympicCoder-32B-iq2_m.gguf b/OlympicCoder-32B-iq2_m.gguf new file mode 100644 index 0000000..b6bba51 --- /dev/null +++ b/OlympicCoder-32B-iq2_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8a49fb737d104ab7b53eeb4f35c38f9816292efe5bdbb79ff1414e2aa656dae4 +size 12792270112 diff --git a/OlympicCoder-32B-iq2_s.gguf b/OlympicCoder-32B-iq2_s.gguf new file mode 100644 index 0000000..1daa809 --- /dev/null +++ b/OlympicCoder-32B-iq2_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5bbd966dcabb9d4f20db791a25434c0c86e323d6ba101b3704796a8e5f48120b +size 12163124512 diff --git a/OlympicCoder-32B-iq2_xs.gguf b/OlympicCoder-32B-iq2_xs.gguf new file mode 100644 index 0000000..413b953 --- /dev/null +++ b/OlympicCoder-32B-iq2_xs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:26cd6ce0c87a396ae7514ace19f6fd2eeee49195075a59950d254728709c9c17 +size 10688564512 diff --git a/OlympicCoder-32B-iq2_xxs.gguf b/OlympicCoder-32B-iq2_xxs.gguf new file mode 100644 index 0000000..3ae3434 --- /dev/null +++ b/OlympicCoder-32B-iq2_xxs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7ea1d05a0cb4f5ba9fbe049b42a33743a59e976ce2f8ffe51b587e9c23491555 +size 10006990112 diff --git a/OlympicCoder-32B-iq3_m.gguf b/OlympicCoder-32B-iq3_m.gguf new file mode 100644 index 0000000..6f6e483 --- /dev/null +++ b/OlympicCoder-32B-iq3_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:acec75aa5b0a330ee0075c30cd42072fe8fd24d52f626227e1f939257e9202d9 +size 14907444512 diff --git a/OlympicCoder-32B-iq3_s.gguf b/OlympicCoder-32B-iq3_s.gguf new file mode 100644 index 0000000..96edcca --- /dev/null +++ b/OlympicCoder-32B-iq3_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:021d8ef45d550ff3eb6ab4a256353746535d87079c4bac87d9ca19cc16b79b98 +size 14534216992 diff --git a/OlympicCoder-32B-iq3_xs.gguf b/OlympicCoder-32B-iq3_xs.gguf new file mode 100644 index 0000000..57d392c --- /dev/null +++ b/OlympicCoder-32B-iq3_xs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ecd6ff484dc0bc9e550221ddcf61eebc5859740a28e260f3d838fb5eea88631e +size 13802835232 diff --git a/OlympicCoder-32B-iq3_xxs.gguf b/OlympicCoder-32B-iq3_xxs.gguf new file mode 100644 index 0000000..01fcefe --- /dev/null +++ b/OlympicCoder-32B-iq3_xxs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9ab887ad460c533c70e5978169aceb8d6ce15afbd6d0bfb990ffc8e101f030f7 +size 13039996192 diff --git a/OlympicCoder-32B-iq4_nl.gguf b/OlympicCoder-32B-iq4_nl.gguf new file mode 100644 index 0000000..0cc52ea --- /dev/null +++ b/OlympicCoder-32B-iq4_nl.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5d8c37358937e2927ec48865137f004849b6e5d6fa39cbd6b99a9423a36d6f08 +size 18682174752 diff --git a/OlympicCoder-32B-iq4_xs.gguf b/OlympicCoder-32B-iq4_xs.gguf new file mode 100644 index 0000000..fc9a4ed --- /dev/null +++ b/OlympicCoder-32B-iq4_xs.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c55d4ece18d92639fa5238d94436d26cc94e08591a59d0cb4dde732398b628a9 +size 17693154592 diff --git a/OlympicCoder-32B-q2_k_s.gguf b/OlympicCoder-32B-q2_k_s.gguf new file mode 100644 index 0000000..b2b3600 --- /dev/null +++ b/OlympicCoder-32B-q2_k_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fadeeee0708ba7c1e2019d84b46f3ddcc6158b693ceeb06b1f3e5c4c47c17427 +size 11664395552 diff --git a/OlympicCoder-32B-q3_k_m.gguf b/OlympicCoder-32B-q3_k_m.gguf new file mode 100644 index 0000000..1baaea5 --- /dev/null +++ b/OlympicCoder-32B-q3_k_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ce23eef45fb51aeaca8df6c95bc5e02a70116006ecc5aae68512309834b4d338 +size 16032369952 diff --git a/OlympicCoder-32B-q3_k_s.gguf b/OlympicCoder-32B-q3_k_s.gguf new file mode 100644 index 0000000..6c12121 --- /dev/null +++ b/OlympicCoder-32B-q3_k_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f59e1a792297cf9acc0442eb3b6f4d3cfa309881d8fd13aa84c436014e84b6ba +size 14489652512 diff --git a/OlympicCoder-32B-q4_0.gguf b/OlympicCoder-32B-q4_0.gguf new file mode 100644 index 0000000..1133adb --- /dev/null +++ b/OlympicCoder-32B-q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5470a18eef80870222c683572d495aae83b360afa7b5688392c43f77871d64da +size 18439507232 diff --git a/OlympicCoder-32B-q4_1.gguf b/OlympicCoder-32B-q4_1.gguf new file mode 100644 index 0000000..54f52f7 --- /dev/null +++ b/OlympicCoder-32B-q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9751988ef5ad604ec72184e832d25064d994e7886f471aab0ca67aa2ce074e79 +size 20487179552 diff --git a/OlympicCoder-32B-q4_k_m.gguf b/OlympicCoder-32B-q4_k_m.gguf new file mode 100644 index 0000000..2742516 --- /dev/null +++ b/OlympicCoder-32B-q4_k_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:79bb997f42d2f775604b51a3b2e06e3c674c23c872a4685800d77af536dcd98d +size 20052061472 diff --git a/OlympicCoder-32B-q4_k_s.gguf b/OlympicCoder-32B-q4_k_s.gguf new file mode 100644 index 0000000..04fc5c8 --- /dev/null +++ b/OlympicCoder-32B-q4_k_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:63ef79b83f498365bb81d3f63ae57c7745b17418b888048a9043459377fda0f4 +size 18985135392 diff --git a/OlympicCoder-32B-q5_0.gguf b/OlympicCoder-32B-q5_0.gguf new file mode 100644 index 0000000..88f9c37 --- /dev/null +++ b/OlympicCoder-32B-q5_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ed3f3accce544e972348d5777453f1ef676a5ab498a7212589bbed0cacbc1a38 +size 22534851872 diff --git a/OlympicCoder-32B-q5_1.gguf b/OlympicCoder-32B-q5_1.gguf new file mode 100644 index 0000000..3634443 --- /dev/null +++ b/OlympicCoder-32B-q5_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5a10327a135f3ad62f635bd9dbaf22f8253c21813e806348360e6576c9a8a71d +size 24582524192 diff --git a/OlympicCoder-32B-q5_k_m.gguf b/OlympicCoder-32B-q5_k_m.gguf new file mode 100644 index 0000000..f5e52dc --- /dev/null +++ b/OlympicCoder-32B-q5_k_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:957691d84a77c77807fbf2495ce05830504bc8bf25c29a5b8ca5fd5660f8f42a +size 23365561632 diff --git a/OlympicCoder-32B-q5_k_s.gguf b/OlympicCoder-32B-q5_k_s.gguf new file mode 100644 index 0000000..bdda732 --- /dev/null +++ b/OlympicCoder-32B-q5_k_s.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:50c7061df7e138fd8bad708dec90b5e572c848cf0e7f01db633db75fccc342fa +size 22741658912 diff --git a/OlympicCoder-32B-q6_k_m.gguf b/OlympicCoder-32B-q6_k_m.gguf new file mode 100644 index 0000000..d66c530 --- /dev/null +++ b/OlympicCoder-32B-q6_k_m.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f371e34dad30b43ab1f3a972c274daf90e16a1591d6ba83431db22d0d21df9a7 +size 26886155552 diff --git a/OlympicCoder-32B-q8_0.gguf b/OlympicCoder-32B-q8_0.gguf new file mode 100644 index 0000000..d39984e --- /dev/null +++ b/OlympicCoder-32B-q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:474caba2f30daa80d5c9bad7aeee9f20668db75ade7cf8ac2efab9067859358d +size 34820885504 diff --git a/OlympicCoder-32B.imatrix b/OlympicCoder-32B.imatrix new file mode 100644 index 0000000..66cfc26 --- /dev/null +++ b/OlympicCoder-32B.imatrix @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4cb18168582982f986dd85e589421ced38417b42c3b298d4a664186b5530fd37 +size 14957098 diff --git a/README.md b/README.md new file mode 100644 index 0000000..f8550ef --- /dev/null +++ b/README.md @@ -0,0 +1,338 @@ +--- +license: apache-2.0 +datasets: +- open-r1/codeforces-cots +language: +- en +base_model: +- Qwen/Qwen2.5-Coder-32B-Instruct +pipeline_tag: text-generation +library_name: transformers +--- + +# OlympicCoder-32B GGUF Models + +## Ultra-Low-Bit Quantization with IQ-DynamicGate (1-2 bit) + +Our latest quantization method introduces **precision-adaptive quantization** for ultra-low-bit models (1-2 bit), with benchmark-proven improvements on **Llama-3-8B**. This approach uses layer-specific strategies to preserve accuracy while maintaining extreme memory efficiency. + +### **Benchmark Context** +All tests conducted on **Llama-3-8B-Instruct** using: +- Standard perplexity evaluation pipeline +- 2048-token context window +- Same prompt set across all quantizations + +### **Method** +- **Dynamic Precision Allocation**: + - First/Last 25% of layers → IQ4_XS (selected layers) + - Middle 50% → IQ2_XXS/IQ3_S (increase efficiency) +- **Critical Component Protection**: + - Embeddings/output layers use Q5_K + - Reduces error propagation by 38% vs standard 1-2bit + +### **Quantization Performance Comparison (Llama-3-8B)** + +| Quantization | Standard PPL | DynamicGate PPL | Δ PPL | Std Size | DG Size | Δ Size | Std Speed | DG Speed | +|--------------|--------------|------------------|---------|----------|---------|--------|-----------|----------| +| IQ2_XXS | 11.30 | 9.84 | -12.9% | 2.5G | 2.6G | +0.1G | 234s | 246s | +| IQ2_XS | 11.72 | 11.63 | -0.8% | 2.7G | 2.8G | +0.1G | 242s | 246s | +| IQ2_S | 14.31 | 9.02 | -36.9% | 2.7G | 2.9G | +0.2G | 238s | 244s | +| IQ1_M | 27.46 | 15.41 | -43.9% | 2.2G | 2.5G | +0.3G | 206s | 212s | +| IQ1_S | 53.07 | 32.00 | -39.7% | 2.1G | 2.4G | +0.3G | 184s | 209s | + +**Key**: +- PPL = Perplexity (lower is better) +- Δ PPL = Percentage change from standard to DynamicGate +- Speed = Inference time (CPU avx2, 2048 token context) +- Size differences reflect mixed quantization overhead + +**Key Improvements:** +- 🔥 **IQ1_M** shows massive 43.9% perplexity reduction (27.46 → 15.41) +- 🚀 **IQ2_S** cuts perplexity by 36.9% while adding only 0.2GB +- ⚡ **IQ1_S** maintains 39.7% better accuracy despite 1-bit quantization + +**Tradeoffs:** +- All variants have modest size increases (0.1-0.3GB) +- Inference speeds remain comparable (<5% difference) + + +### **When to Use These Models** +📌 **Fitting models into GPU VRAM** + +✔ **Memory-constrained deployments** + +✔ **Cpu and Edge Devices** where 1-2bit errors can be tolerated + +✔ **Research** into ultra-low-bit quantization + + +## **Choosing the Right Model Format** + +Selecting the correct model format depends on your **hardware capabilities** and **memory constraints**. + +### **BF16 (Brain Float 16) – Use if BF16 acceleration is available** +- A 16-bit floating-point format designed for **faster computation** while retaining good precision. +- Provides **similar dynamic range** as FP32 but with **lower memory usage**. +- Recommended if your hardware supports **BF16 acceleration** (check your device's specs). +- Ideal for **high-performance inference** with **reduced memory footprint** compared to FP32. + +📌 **Use BF16 if:** +✔ Your hardware has native **BF16 support** (e.g., newer GPUs, TPUs). +✔ You want **higher precision** while saving memory. +✔ You plan to **requantize** the model into another format. + +📌 **Avoid BF16 if:** +❌ Your hardware does **not** support BF16 (it may fall back to FP32 and run slower). +❌ You need compatibility with older devices that lack BF16 optimization. + +--- + +### **F16 (Float 16) – More widely supported than BF16** +- A 16-bit floating-point **high precision** but with less of range of values than BF16. +- Works on most devices with **FP16 acceleration support** (including many GPUs and some CPUs). +- Slightly lower numerical precision than BF16 but generally sufficient for inference. + +📌 **Use F16 if:** +✔ Your hardware supports **FP16** but **not BF16**. +✔ You need a **balance between speed, memory usage, and accuracy**. +✔ You are running on a **GPU** or another device optimized for FP16 computations. + +📌 **Avoid F16 if:** +❌ Your device lacks **native FP16 support** (it may run slower than expected). +❌ You have memory limitations. + +--- + +### **Quantized Models (Q4_K, Q6_K, Q8, etc.) – For CPU & Low-VRAM Inference** +Quantization reduces model size and memory usage while maintaining as much accuracy as possible. +- **Lower-bit models (Q4_K)** → **Best for minimal memory usage**, may have lower precision. +- **Higher-bit models (Q6_K, Q8_0)** → **Better accuracy**, requires more memory. + +📌 **Use Quantized Models if:** +✔ You are running inference on a **CPU** and need an optimized model. +✔ Your device has **low VRAM** and cannot load full-precision models. +✔ You want to reduce **memory footprint** while keeping reasonable accuracy. + +📌 **Avoid Quantized Models if:** +❌ You need **maximum accuracy** (full-precision models are better for this). +❌ Your hardware has enough VRAM for higher-precision formats (BF16/F16). + +--- + +### **Very Low-Bit Quantization (IQ3_XS, IQ3_S, IQ3_M, Q4_K, Q4_0)** +These models are optimized for **extreme memory efficiency**, making them ideal for **low-power devices** or **large-scale deployments** where memory is a critical constraint. + +- **IQ3_XS**: Ultra-low-bit quantization (3-bit) with **extreme memory efficiency**. + - **Use case**: Best for **ultra-low-memory devices** where even Q4_K is too large. + - **Trade-off**: Lower accuracy compared to higher-bit quantizations. + +- **IQ3_S**: Small block size for **maximum memory efficiency**. + - **Use case**: Best for **low-memory devices** where **IQ3_XS** is too aggressive. + +- **IQ3_M**: Medium block size for better accuracy than **IQ3_S**. + - **Use case**: Suitable for **low-memory devices** where **IQ3_S** is too limiting. + +- **Q4_K**: 4-bit quantization with **block-wise optimization** for better accuracy. + - **Use case**: Best for **low-memory devices** where **Q6_K** is too large. + +- **Q4_0**: Pure 4-bit quantization, optimized for **ARM devices**. + - **Use case**: Best for **ARM-based devices** or **low-memory environments**. + +--- + +### **Summary Table: Model Format Selection** + +| Model Format | Precision | Memory Usage | Device Requirements | Best Use Case | +|--------------|------------|---------------|----------------------|---------------| +| **BF16** | Highest | High | BF16-supported GPU/CPUs | High-speed inference with reduced memory | +| **F16** | High | High | FP16-supported devices | GPU inference when BF16 isn't available | +| **Q4_K** | Medium Low | Low | CPU or Low-VRAM devices | Best for memory-constrained environments | +| **Q6_K** | Medium | Moderate | CPU with more memory | Better accuracy while still being quantized | +| **Q8_0** | High | Moderate | CPU or GPU with enough VRAM | Best accuracy among quantized models | +| **IQ3_XS** | Very Low | Very Low | Ultra-low-memory devices | Extreme memory efficiency and low accuracy | +| **Q4_0** | Low | Low | ARM or low-memory devices | llama.cpp can optimize for ARM devices | + +--- + +## **Included Files & Details** + +### `OlympicCoder-32B-bf16.gguf` +- Model weights preserved in **BF16**. +- Use this if you want to **requantize** the model into a different format. +- Best if your device supports **BF16 acceleration**. + +### `OlympicCoder-32B-f16.gguf` +- Model weights stored in **F16**. +- Use if your device supports **FP16**, especially if BF16 is not available. + +### `OlympicCoder-32B-bf16-q8_0.gguf` +- **Output & embeddings** remain in **BF16**. +- All other layers quantized to **Q8_0**. +- Use if your device supports **BF16** and you want a quantized version. + +### `OlympicCoder-32B-f16-q8_0.gguf` +- **Output & embeddings** remain in **F16**. +- All other layers quantized to **Q8_0**. + +### `OlympicCoder-32B-q4_k.gguf` +- **Output & embeddings** quantized to **Q8_0**. +- All other layers quantized to **Q4_K**. +- Good for **CPU inference** with limited memory. + +### `OlympicCoder-32B-q4_k_s.gguf` +- Smallest **Q4_K** variant, using less memory at the cost of accuracy. +- Best for **very low-memory setups**. + +### `OlympicCoder-32B-q6_k.gguf` +- **Output & embeddings** quantized to **Q8_0**. +- All other layers quantized to **Q6_K** . + +### `OlympicCoder-32B-q8_0.gguf` +- Fully **Q8** quantized model for better accuracy. +- Requires **more memory** but offers higher precision. + +### `OlympicCoder-32B-iq3_xs.gguf` +- **IQ3_XS** quantization, optimized for **extreme memory efficiency**. +- Best for **ultra-low-memory devices**. + +### `OlympicCoder-32B-iq3_m.gguf` +- **IQ3_M** quantization, offering a **medium block size** for better accuracy. +- Suitable for **low-memory devices**. + +### `OlympicCoder-32B-q4_0.gguf` +- Pure **Q4_0** quantization, optimized for **ARM devices**. +- Best for **low-memory environments**. +- Prefer IQ4_NL for better accuracy. + +# 🚀 If you find these models useful +❤ **Please click "Like" if you find this useful!** +Help me test my **AI-Powered Network Monitor Assistant** with **quantum-ready security checks**: +👉 [Quantum Network Monitor](https://readyforquantum.com) + +💬 **How to test**: +1. Click the **chat icon** (bottom right on any page) +2. Choose an **AI assistant type**: + - `TurboLLM` (GPT-4-mini) + - `FreeLLM` (Open-source) + - `TestLLM` (Experimental CPU-only) + +### **What I’m Testing** +I’m pushing the limits of **small open-source models for AI network monitoring**, specifically: +- **Function calling** against live network services +- **How small can a model go** while still handling: + - Automated **Nmap scans** + - **Quantum-readiness checks** + - **Metasploit integration** + +🟡 **TestLLM** – Current experimental model (llama.cpp on 6 CPU threads): +- ✅ **Zero-configuration setup** +- ⏳ 30s load time (slow inference but **no API costs**) +- 🔧 **Help wanted!** If you’re into **edge-device AI**, let’s collaborate! + +### **Other Assistants** +🟢 **TurboLLM** – Uses **gpt-4-mini** for: +- **Real-time network diagnostics** +- **Automated penetration testing** (Nmap/Metasploit) +- 🔑 Get more tokens by [downloading our Quantum Network Monitor Agent](https://readyforquantum.com/download/?utm_source=huggingface&utm_medium=referral&utm_campaign=huggingface_repo_readme) + +🔵 **HugLLM** – Open-source models (≈8B params): +- **2x more tokens** than TurboLLM +- **AI-powered log analysis** +- 🌐 Runs on Hugging Face Inference API + +### 💡 **Example AI Commands to Test**: +1. `"Give me info on my websites SSL certificate"` +2. `"Check if my server is using quantum safe encyption for communication"` +3. `"Run a quick Nmap vulnerability test"` +4. '"Create a cmd processor to .. (what ever you want)" Note you need to install a Quantum Network Monitor Agent to run the .net code from. This is a very flexible and powerful feature. Use with caution! + +### Final word +I fund the servers to create the models files, run the Quantum Network Monitor Service and Pay for Inference from Novita and OpenAI all from my own pocket. All of the code for creating the models and the work I have done with Quantum Network Monitor is [open source](https://github.com/Mungert69). Feel free to use what you find useful. Please support my work and consider [buying me a coffee](https://www.buymeacoffee.com/mahadeva) . +This will help me pay for the services and increase the token limits for everyone. + +Thank you :) + + + +# Model Card for OlympicCoder-32B + +OlympicCoder-32B is a code model that achieves very strong performance on competitive coding benchmarks such as LiveCodeBench andthe 2024 International Olympiad in Informatics. + +* Repository: https://github.com/huggingface/open-r1 +* Blog post: https://huggingface.co/blog/open-r1/update-3 + +## Model description + +- **Model type:** A 32B parameter model fine-tuned on a decontaminated version of the codeforces dataset. +- **Language(s) (NLP):** Primarily English +- **License:** apache-2.0 +- **Finetuned from model:** [Qwen/Qwen2.5-Coder-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct) + +## Evaluation + +We compare the performance of OlympicCoder models on two main benchmarks for competitive coding: + +* **[IOI'2024:](https://github.com/huggingface/ioi)** 6 very challenging problems from the 2024 International Olympiad in Informatics. Models are allowed up to 50 submissions per problem. +* **[LiveCodeBench:](https://livecodebench.github.io)** Python programming problems source from platforms like CodeForces and LeetCoder. We use the `v4_v5` subset of [`livecodebench/code_generation_lite`](https://huggingface.co/datasets/livecodebench/code_generation_lite), which corresponds to 268 problems. We use `lighteval` to evaluate models on LiveCodeBench using the sampling parameters described [here](https://github.com/huggingface/open-r1?tab=readme-ov-file#livecodebench). + +> [!NOTE] +> The OlympicCoder models were post-trained exclusively on C++ solutions generated by DeepSeek-R1. As a result the performance on LiveCodeBench should be considered to be partially _out-of-domain_, since this expects models to output solutions in Python. + +### IOI'24 + +![](./ioi-evals.png) + +### LiveCodeBench + +![](./lcb-evals.png) + + + +## Usage +Here's how you can run the model using the `pipeline()` function from 🤗 Transformers: + +```python +# pip install transformers +# pip install accelerate + +import torch +from transformers import pipeline + +pipe = pipeline("text-generation", model="open-r1/OlympicCoder-32B", torch_dtype=torch.bfloat16, device_map="auto") + +# We use the tokenizer's chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating +messages = [ + {"role": "user", "content": "Write a python program to calculate the 10th Fibonacci number"}, +] +prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) +outputs = pipe(prompt, max_new_tokens=8000, do_sample=True, temperature=0.7, top_k=50, top_p=0.95) +print(outputs[0]["generated_text"]) +#<|im_start|>user +#Write a python program to calculate the 10th fibonacci number<|im_end|> +#<|im_start|>assistant +#Okay, I need to write a Python program that calculates the 10th Fibonacci number. Hmm, the Fibonacci sequence starts with 0 and 1. Each subsequent number is the sum of the two preceding ones. So the sequence goes: 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, and so on. ... +``` + +> [!IMPORTANT] +> To ensure that the model consistently outputs a long chain-of-thought, we have edited the chat template to prefill the first assistant turn with a `` token. As a result, the outputs from this model will not show the opening `` token if you use the model's `generate()` method. To apply reinforcement learning with a format reward, either prepend the `` token to the model's completions or amend the chat template to remove the prefill. Check out our [blog post](https://huggingface.co/blog/open-r1/update-3#lesson-4-prefill-with-think-to-consistently-enable-long-cot) for more details. + + +## Training procedure +### Training hyper-parameters + +The following hyperparameters were used during training on 16 H100 nodes: + +- dataset: open-r1/codeforces-cots_decontaminated +- learning_rate: 4.0e-5 +- train_batch_size: 1 +- seed: 42 +- packing: false +- distributed_type: fsdp +- num_devices: 128 +- gradient_accumulation_steps: 1 +- total_train_batch_size: 16 +- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 +- lr_scheduler_type: cosine_with_min_lr +- min_lr_rate: 0.1 +- lr_scheduler_warmup_ratio: 0.03 +- num_epochs: 10.0 \ No newline at end of file diff --git a/bf16/OlympicCoder-32B-bf16-00001-of-00002.gguf b/bf16/OlympicCoder-32B-bf16-00001-of-00002.gguf new file mode 100644 index 0000000..e483b7b --- /dev/null +++ b/bf16/OlympicCoder-32B-bf16-00001-of-00002.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2ddcef0e37ec619b23f02373a501a29f67bcab9bfa1636aa86ab827d4ad549e0 +size 45902462976 diff --git a/bf16/OlympicCoder-32B-bf16-00002-of-00002.gguf b/bf16/OlympicCoder-32B-bf16-00002-of-00002.gguf new file mode 100644 index 0000000..fe47437 --- /dev/null +++ b/bf16/OlympicCoder-32B-bf16-00002-of-00002.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:97d21722f435ac27eb23be9e021858266119e3c05530a4eba4767ee39bac39f0 +size 19633507328 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/f16/OlympicCoder-32B-F16-00001-of-00002.gguf b/f16/OlympicCoder-32B-F16-00001-of-00002.gguf new file mode 100644 index 0000000..b3cac35 --- /dev/null +++ b/f16/OlympicCoder-32B-F16-00001-of-00002.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:87d0be0a75dc96a8e95583763c09597d49c92adeedcc5b6cf648546e6a74319b +size 45902462976 diff --git a/f16/OlympicCoder-32B-F16-00002-of-00002.gguf b/f16/OlympicCoder-32B-F16-00002-of-00002.gguf new file mode 100644 index 0000000..f48600c --- /dev/null +++ b/f16/OlympicCoder-32B-F16-00002-of-00002.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9f270d7304dccd6a36590478c486fa06657319d965caec2e3195100c1eda9be1 +size 19633507328