commit 5301c4198164b49b55cb4aad290cdb94dbbc9695 Author: ModelHub XC Date: Sat Aug 1 21:53:13 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: bartowski/Qwen_Qwen3-14B-GGUF Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..5532b23 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,49 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bin.* filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zstandard filter=lfs diff=lfs merge=lfs -text +*.tfevents* filter=lfs diff=lfs merge=lfs -text +*.db* filter=lfs diff=lfs merge=lfs -text +*.ark* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.gguf* filter=lfs diff=lfs merge=lfs -text +*.ggml filter=lfs diff=lfs merge=lfs -text +*.llamafile* filter=lfs diff=lfs merge=lfs -text +*.pt2 filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text + +Qwen_Qwen3-14B.imatrix filter=lfs diff=lfs merge=lfs -text \ No newline at end of file diff --git a/Qwen_Qwen3-14B-IQ2_M.gguf b/Qwen_Qwen3-14B-IQ2_M.gguf new file mode 100644 index 0000000..2a49af8 --- /dev/null +++ b/Qwen_Qwen3-14B-IQ2_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c4d9143e2cd3827a1d4cf56da13e64918bdd664f8908ae4b6d86e55cf8d4bff +size 5322941472 diff --git a/Qwen_Qwen3-14B-IQ2_S.gguf b/Qwen_Qwen3-14B-IQ2_S.gguf new file mode 100644 index 0000000..5e2eeac --- /dev/null +++ b/Qwen_Qwen3-14B-IQ2_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cfd11747daa50d47d3eeaf2af7503906c2fe966635458d8447dfb8ee40485a8b +size 4963312672 diff --git a/Qwen_Qwen3-14B-IQ2_XS.gguf b/Qwen_Qwen3-14B-IQ2_XS.gguf new file mode 100644 index 0000000..1afd3c8 --- /dev/null +++ b/Qwen_Qwen3-14B-IQ2_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7277e485b898f8fa7945f701c406401f683e5b49107e3c2f42907e829375f4d1 +size 4691589152 diff --git a/Qwen_Qwen3-14B-IQ3_M.gguf b/Qwen_Qwen3-14B-IQ3_M.gguf new file mode 100644 index 0000000..96f0a8d --- /dev/null +++ b/Qwen_Qwen3-14B-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:86b555a31a6d8bd928d5ee4b6dbf8a73ffafe419b8d7247ced73fb2be1f3c9d1 +size 6883409952 diff --git a/Qwen_Qwen3-14B-IQ3_XS.gguf b/Qwen_Qwen3-14B-IQ3_XS.gguf new file mode 100644 index 0000000..95ce847 --- /dev/null +++ b/Qwen_Qwen3-14B-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:363c51dbaec3e2944f99738568960a92488487e218f1300918af71e9e3802ce5 +size 6375301152 diff --git a/Qwen_Qwen3-14B-IQ3_XXS.gguf b/Qwen_Qwen3-14B-IQ3_XXS.gguf new file mode 100644 index 0000000..1877d61 --- /dev/null +++ b/Qwen_Qwen3-14B-IQ3_XXS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a814fb65a1600c17db566cfc8496cadcb5be0883a58b33d289a9e3eefc157c14 +size 5942666272 diff --git a/Qwen_Qwen3-14B-IQ4_NL.gguf b/Qwen_Qwen3-14B-IQ4_NL.gguf new file mode 100644 index 0000000..93db4df --- /dev/null +++ b/Qwen_Qwen3-14B-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fcb00c64571daba61218903cf0af72a4077f1edaa58823db476011c053dde86c +size 8541363232 diff --git a/Qwen_Qwen3-14B-IQ4_XS.gguf b/Qwen_Qwen3-14B-IQ4_XS.gguf new file mode 100644 index 0000000..bb1bf6f --- /dev/null +++ b/Qwen_Qwen3-14B-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:448afbe2d83a40aa38e03019c37fec002d0e42da1f6a0a8d250876915b84a00d +size 8110730272 diff --git a/Qwen_Qwen3-14B-Q2_K.gguf b/Qwen_Qwen3-14B-Q2_K.gguf new file mode 100644 index 0000000..88d88d4 --- /dev/null +++ b/Qwen_Qwen3-14B-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4496a2dc7805120ae48c381e41bc81abee89eb7a82c5f33f72868f5d973e2cb2 +size 5753984032 diff --git a/Qwen_Qwen3-14B-Q2_K_L.gguf b/Qwen_Qwen3-14B-Q2_K_L.gguf new file mode 100644 index 0000000..ef3cf69 --- /dev/null +++ b/Qwen_Qwen3-14B-Q2_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a14999ffc47a2a98d2eecf3976a2fffe0fea1a67b11d6d88242f3649913ea62d +size 6513664032 diff --git a/Qwen_Qwen3-14B-Q3_K_L.gguf b/Qwen_Qwen3-14B-Q3_K_L.gguf new file mode 100644 index 0000000..28d5758 --- /dev/null +++ b/Qwen_Qwen3-14B-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a0fa4eabfc6a0ae490e35343d214b5d236285acd7543a8325fe99ba778a02df +size 7900651552 diff --git a/Qwen_Qwen3-14B-Q3_K_M.gguf b/Qwen_Qwen3-14B-Q3_K_M.gguf new file mode 100644 index 0000000..066e2a8 --- /dev/null +++ b/Qwen_Qwen3-14B-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3c49c800881cf97b77d6a6a7d3f653e35b6134e0cb6fd6f7c379b354b48af781 +size 7321313312 diff --git a/Qwen_Qwen3-14B-Q3_K_S.gguf b/Qwen_Qwen3-14B-Q3_K_S.gguf new file mode 100644 index 0000000..8a51367 --- /dev/null +++ b/Qwen_Qwen3-14B-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:758b8e6a06f55368ea6cf151176dc0d3b931894b8630ec20cf704d2d6be279b4 +size 6657105952 diff --git a/Qwen_Qwen3-14B-Q3_K_XL.gguf b/Qwen_Qwen3-14B-Q3_K_XL.gguf new file mode 100644 index 0000000..de42f0d --- /dev/null +++ b/Qwen_Qwen3-14B-Q3_K_XL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4b34ed0ed1b710121fc93cfb2aea162a0140bae01e503512fff546e2ab6fdaab +size 8581324832 diff --git a/Qwen_Qwen3-14B-Q4_0.gguf b/Qwen_Qwen3-14B-Q4_0.gguf new file mode 100644 index 0000000..d2d13d6 --- /dev/null +++ b/Qwen_Qwen3-14B-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:639a6df3ef3231561e0a97e0d683812aa9025f401035c0c80aec7c368b02e18c +size 8543001632 diff --git a/Qwen_Qwen3-14B-Q4_1.gguf b/Qwen_Qwen3-14B-Q4_1.gguf new file mode 100644 index 0000000..177b17d --- /dev/null +++ b/Qwen_Qwen3-14B-Q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:02d8933245c909d63526421eb3c4563c73bd2b0d4b2aa2f2524b3005128094e9 +size 9389521952 diff --git a/Qwen_Qwen3-14B-Q4_K_L.gguf b/Qwen_Qwen3-14B-Q4_K_L.gguf new file mode 100644 index 0000000..0f5bf66 --- /dev/null +++ b/Qwen_Qwen3-14B-Q4_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0414094a48638a4c4862d7b822fed5ae8a1643e0c7caa799f452c01a00978eb9 +size 9579110432 diff --git a/Qwen_Qwen3-14B-Q4_K_M.gguf b/Qwen_Qwen3-14B-Q4_K_M.gguf new file mode 100644 index 0000000..c7b43e5 --- /dev/null +++ b/Qwen_Qwen3-14B-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:915913e22399475dbe6c968ac014d9f1fbe08975e489279aede9d5c7b2c98eb6 +size 9001753632 diff --git a/Qwen_Qwen3-14B-Q4_K_S.gguf b/Qwen_Qwen3-14B-Q4_K_S.gguf new file mode 100644 index 0000000..edcaedb --- /dev/null +++ b/Qwen_Qwen3-14B-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c185d2643e1e8c550c1c85991908f160f862864e8d6e7b0bc12e5f2aeef2c6b +size 8573475872 diff --git a/Qwen_Qwen3-14B-Q5_K_L.gguf b/Qwen_Qwen3-14B-Q5_K_L.gguf new file mode 100644 index 0000000..ec3f2e9 --- /dev/null +++ b/Qwen_Qwen3-14B-Q5_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e8641e156ab6386074f7d3ee37ad89f45639ba3b5e1d6e2ac4f24d3db3ae16e9 +size 10994688032 diff --git a/Qwen_Qwen3-14B-Q5_K_M.gguf b/Qwen_Qwen3-14B-Q5_K_M.gguf new file mode 100644 index 0000000..0f656b3 --- /dev/null +++ b/Qwen_Qwen3-14B-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bf19bf5c77c530762012f71215b62aa8f0bae08111ee0d537b8496479e17cfa3 +size 10514570272 diff --git a/Qwen_Qwen3-14B-Q5_K_S.gguf b/Qwen_Qwen3-14B-Q5_K_S.gguf new file mode 100644 index 0000000..c9f4ae9 --- /dev/null +++ b/Qwen_Qwen3-14B-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e13b3a83698cda6ebff869ab366f80b9fddf5d76f02f13bf8236d77d3833d10b +size 10263895072 diff --git a/Qwen_Qwen3-14B-Q6_K.gguf b/Qwen_Qwen3-14B-Q6_K.gguf new file mode 100644 index 0000000..898d95c --- /dev/null +++ b/Qwen_Qwen3-14B-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:de571a7d7c72b99de6d67e7f2b780d774c74f998e088c77133069568cad54294 +size 12121937952 diff --git a/Qwen_Qwen3-14B-Q6_K_L.gguf b/Qwen_Qwen3-14B-Q6_K_L.gguf new file mode 100644 index 0000000..5a1b162 --- /dev/null +++ b/Qwen_Qwen3-14B-Q6_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ae52c8a419483e366c2db0b4b1bf911b24a78fb14f22564f98c0b7bae0e62248 +size 12498739232 diff --git a/Qwen_Qwen3-14B-Q8_0.gguf b/Qwen_Qwen3-14B-Q8_0.gguf new file mode 100644 index 0000000..abbbe48 --- /dev/null +++ b/Qwen_Qwen3-14B-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:62e390154916e1dc6b00f63d997bda39e8f9679c209dcabb69bdff5043fac2e0 +size 15698534432 diff --git a/Qwen_Qwen3-14B-bf16.gguf b/Qwen_Qwen3-14B-bf16.gguf new file mode 100644 index 0000000..f552ca5 --- /dev/null +++ b/Qwen_Qwen3-14B-bf16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9677a58b0fa8da7771a4d8cc8080208ce02a32ab21305ded10a3798766132d3a +size 29543423776 diff --git a/Qwen_Qwen3-14B.imatrix b/Qwen_Qwen3-14B.imatrix new file mode 100644 index 0000000..d7276f1 --- /dev/null +++ b/Qwen_Qwen3-14B.imatrix @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:03eb98f405e02dc67d043f15150b362073477c9a7fb1c7d6539fbc3e2aa49787 +size 7709778 diff --git a/README.md b/README.md new file mode 100644 index 0000000..604d7ab --- /dev/null +++ b/README.md @@ -0,0 +1,172 @@ +--- +quantized_by: bartowski +pipeline_tag: text-generation +base_model: Qwen/Qwen3-14B +base_model_relation: quantized +--- + +## Llamacpp imatrix Quantizations of Qwen3-14B by Qwen + +Using llama.cpp release b5200 for quantization. + +Original model: https://huggingface.co/Qwen/Qwen3-14B + +All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8) + +Run them in [LM Studio](https://lmstudio.ai/) + +Run them directly with [llama.cpp](https://github.com/ggerganov/llama.cpp), or any other llama.cpp based project + +## Prompt format + +``` +<|im_start|>system +{system_prompt}<|im_end|> +<|im_start|>user +{prompt}<|im_end|> +<|im_start|>assistant +``` + +## Download a file (not the whole branch) from below: + +| Filename | Quant type | File Size | Split | Description | +| -------- | ---------- | --------- | ----- | ----------- | +| [Qwen3-14B-bf16.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-bf16.gguf) | bf16 | 29.54GB | false | Full BF16 weights. | +| [Qwen3-14B-Q8_0.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q8_0.gguf) | Q8_0 | 15.70GB | false | Extremely high quality, generally unneeded but max available quant. | +| [Qwen3-14B-Q6_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q6_K_L.gguf) | Q6_K_L | 12.50GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. | +| [Qwen3-14B-Q6_K.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q6_K.gguf) | Q6_K | 12.12GB | false | Very high quality, near perfect, *recommended*. | +| [Qwen3-14B-Q5_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q5_K_L.gguf) | Q5_K_L | 10.99GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. | +| [Qwen3-14B-Q5_K_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q5_K_M.gguf) | Q5_K_M | 10.51GB | false | High quality, *recommended*. | +| [Qwen3-14B-Q5_K_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q5_K_S.gguf) | Q5_K_S | 10.26GB | false | High quality, *recommended*. | +| [Qwen3-14B-Q4_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q4_K_L.gguf) | Q4_K_L | 9.58GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. | +| [Qwen3-14B-Q4_1.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q4_1.gguf) | Q4_1 | 9.39GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. | +| [Qwen3-14B-Q4_K_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q4_K_M.gguf) | Q4_K_M | 9.00GB | false | Good quality, default size for most use cases, *recommended*. | +| [Qwen3-14B-Q3_K_XL.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q3_K_XL.gguf) | Q3_K_XL | 8.58GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. | +| [Qwen3-14B-Q4_K_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q4_K_S.gguf) | Q4_K_S | 8.57GB | false | Slightly lower quality with more space savings, *recommended*. | +| [Qwen3-14B-Q4_0.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q4_0.gguf) | Q4_0 | 8.54GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. | +| [Qwen3-14B-IQ4_NL.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-IQ4_NL.gguf) | IQ4_NL | 8.54GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. | +| [Qwen3-14B-IQ4_XS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-IQ4_XS.gguf) | IQ4_XS | 8.11GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. | +| [Qwen3-14B-Q3_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q3_K_L.gguf) | Q3_K_L | 7.90GB | false | Lower quality but usable, good for low RAM availability. | +| [Qwen3-14B-Q3_K_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q3_K_M.gguf) | Q3_K_M | 7.32GB | false | Low quality. | +| [Qwen3-14B-IQ3_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-IQ3_M.gguf) | IQ3_M | 6.88GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. | +| [Qwen3-14B-Q3_K_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q3_K_S.gguf) | Q3_K_S | 6.66GB | false | Low quality, not recommended. | +| [Qwen3-14B-Q2_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q2_K_L.gguf) | Q2_K_L | 6.51GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. | +| [Qwen3-14B-IQ3_XS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-IQ3_XS.gguf) | IQ3_XS | 6.38GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. | +| [Qwen3-14B-IQ3_XXS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-IQ3_XXS.gguf) | IQ3_XXS | 5.94GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. | +| [Qwen3-14B-Q2_K.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-Q2_K.gguf) | Q2_K | 5.75GB | false | Very low quality but surprisingly usable. | +| [Qwen3-14B-IQ2_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-IQ2_M.gguf) | IQ2_M | 5.32GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. | +| [Qwen3-14B-IQ2_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-IQ2_S.gguf) | IQ2_S | 4.96GB | false | Low quality, uses SOTA techniques to be usable. | +| [Qwen3-14B-IQ2_XS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-14B-GGUF/blob/main/Qwen_Qwen3-14B-IQ2_XS.gguf) | IQ2_XS | 4.69GB | false | Low quality, uses SOTA techniques to be usable. | + +## Embed/output weights + +Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to. + +## Downloading using huggingface-cli + +
+ Click to view download instructions + +First, make sure you have hugginface-cli installed: + +``` +pip install -U "huggingface_hub[cli]" +``` + +Then, you can target the specific file you want: + +``` +huggingface-cli download bartowski/Qwen_Qwen3-14B-GGUF --include "Qwen_Qwen3-14B-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/Qwen_Qwen3-14B-GGUF --include "Qwen_Qwen3-14B-Q8_0/*" --local-dir ./ +``` + +You can either specify a new local-dir (Qwen_Qwen3-14B-Q8_0) or download them all in place (./) + +
+ +## ARM/AVX information + +Previously, you would download Q4_0_4_4/4_8/8_8, and these would have their weights interleaved in memory in order to improve performance on ARM and AVX machines by loading up more data in one pass. + +Now, however, there is something called "online repacking" for weights. details in [this PR](https://github.com/ggerganov/llama.cpp/pull/9921). If you use Q4_0 and your hardware would benefit from repacking weights, it will do it automatically on the fly. + +As of llama.cpp build [b4282](https://github.com/ggerganov/llama.cpp/releases/tag/b4282) you will not be able to run the Q4_0_X_X files and will instead need to use Q4_0. + +Additionally, if you want to get slightly better quality for , you can use IQ4_NL thanks to [this PR](https://github.com/ggerganov/llama.cpp/pull/10541) which will also repack the weights for ARM, though only the 4_4 for now. The loading time may be slower but it will result in an overall speed incrase. + +
+ Click to view Q4_0_X_X information (deprecated + +I'm keeping this section to show the potential theoretical uplift in performance from using the Q4_0 with online repacking. + +
+ Click to view benchmarks on an AVX2 system (EPYC7702) + +| model | size | params | backend | threads | test | t/s | % (vs Q4_0) | +| ------------------------------ | ---------: | ---------: | ---------- | ------: | ------------: | -------------------: |-------------: | +| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp512 | 204.03 ± 1.03 | 100% | +| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp1024 | 282.92 ± 0.19 | 100% | +| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp2048 | 259.49 ± 0.44 | 100% | +| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg128 | 39.12 ± 0.27 | 100% | +| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg256 | 39.31 ± 0.69 | 100% | +| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg512 | 40.52 ± 0.03 | 100% | +| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp512 | 301.02 ± 1.74 | 147% | +| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp1024 | 287.23 ± 0.20 | 101% | +| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp2048 | 262.77 ± 1.81 | 101% | +| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg128 | 18.80 ± 0.99 | 48% | +| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg256 | 24.46 ± 3.04 | 83% | +| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg512 | 36.32 ± 3.59 | 90% | +| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp512 | 271.71 ± 3.53 | 133% | +| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp1024 | 279.86 ± 45.63 | 100% | +| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp2048 | 320.77 ± 5.00 | 124% | +| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg128 | 43.51 ± 0.05 | 111% | +| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg256 | 43.35 ± 0.09 | 110% | +| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg512 | 42.60 ± 0.31 | 105% | + +Q4_0_8_8 offers a nice bump to prompt processing and a small bump to text generation + +
+ +
+ +## Which file should I choose? + +
+ Click here for details + +A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9) + +The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have. + +If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM. + +If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total. + +Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'. + +If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M. + +If you want to get more into the weeds, you can check out this extremely useful feature chart: + +[llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix) + +But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size. + +These I-quants can also be used on CPU, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide. + +
+ +## Credits + +Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset. + +Thank you ZeroWw for the inspiration to experiment with embed/output. + +Thank you to LM Studio for sponsoring my work. + +Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski diff --git a/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