commit 71511bceb536ef2ed4e6faba94ccc49033acb12c Author: ModelHub XC Date: Mon Aug 3 00:09:13 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: bartowski/Qwen_Qwen3-1.7B-GGUF Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..53d7257 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,47 @@ +*.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 \ No newline at end of file diff --git a/Qwen_Qwen3-1.7B-IQ2_M.gguf b/Qwen_Qwen3-1.7B-IQ2_M.gguf new file mode 100644 index 0000000..ac58f22 --- /dev/null +++ b/Qwen_Qwen3-1.7B-IQ2_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d6f35c090bc6e70437df4080e21a9b0458d568123abfa3ff04e243077edc358c +size 828885408 diff --git a/Qwen_Qwen3-1.7B-IQ3_M.gguf b/Qwen_Qwen3-1.7B-IQ3_M.gguf new file mode 100644 index 0000000..d0950cf --- /dev/null +++ b/Qwen_Qwen3-1.7B-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a0abdf07593f058b056531c9168e6733ed8e64cacc28750fb84f46d1044ccab6 +size 1029366176 diff --git a/Qwen_Qwen3-1.7B-IQ3_XS.gguf b/Qwen_Qwen3-1.7B-IQ3_XS.gguf new file mode 100644 index 0000000..d1b27ee --- /dev/null +++ b/Qwen_Qwen3-1.7B-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1b166b349d5c2dc2f717a688018d1878a283c630b7729df3fdaecf76271803f0 +size 967926176 diff --git a/Qwen_Qwen3-1.7B-IQ3_XXS.gguf b/Qwen_Qwen3-1.7B-IQ3_XXS.gguf new file mode 100644 index 0000000..b914300 --- /dev/null +++ b/Qwen_Qwen3-1.7B-IQ3_XXS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c7649660c508f31f522ac7fa9e70fa03f989250ce9884cc3747f0e6c7443fbe +size 888064416 diff --git a/Qwen_Qwen3-1.7B-IQ4_NL.gguf b/Qwen_Qwen3-1.7B-IQ4_NL.gguf new file mode 100644 index 0000000..043e2dc --- /dev/null +++ b/Qwen_Qwen3-1.7B-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:66086acac62fe510a0975d60c22b27485f8a7b94bde7b908c8c38edb1891a93a +size 1229453728 diff --git a/Qwen_Qwen3-1.7B-IQ4_XS.gguf b/Qwen_Qwen3-1.7B-IQ4_XS.gguf new file mode 100644 index 0000000..d85a5f5 --- /dev/null +++ b/Qwen_Qwen3-1.7B-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:26bb1677b3dbd490e9b0e4e20b2356646f3a43819e795b39f772fdac07ba091d +size 1175689632 diff --git a/Qwen_Qwen3-1.7B-Q2_K.gguf b/Qwen_Qwen3-1.7B-Q2_K.gguf new file mode 100644 index 0000000..b64d432 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:10bfbf421f9c9a631ec1294a6f4005b76d37ea3902d90408a1121315c4b07d21 +size 879896992 diff --git a/Qwen_Qwen3-1.7B-Q2_K_L.gguf b/Qwen_Qwen3-1.7B-Q2_K_L.gguf new file mode 100644 index 0000000..d964204 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q2_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ec641bc2d026bd18a8bd9a4f136c5751aa44eed792908a4b36dca396f44feb17 +size 1183768992 diff --git a/Qwen_Qwen3-1.7B-Q3_K_L.gguf b/Qwen_Qwen3-1.7B-Q3_K_L.gguf new file mode 100644 index 0000000..ff4e7eb --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:088aa1cedfee641acdf4c51c80d02c7fd145b3bb41e7e29d155759fd19cd7a09 +size 1137205664 diff --git a/Qwen_Qwen3-1.7B-Q3_K_M.gguf b/Qwen_Qwen3-1.7B-Q3_K_M.gguf new file mode 100644 index 0000000..7e24c18 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d544d30dbb7b2f608775d69c939a9b8d89c45df2be1a5ad1390ddc1bd7f8d6f5 +size 1073242528 diff --git a/Qwen_Qwen3-1.7B-Q3_K_S.gguf b/Qwen_Qwen3-1.7B-Q3_K_S.gguf new file mode 100644 index 0000000..3e69868 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b7ecc78b5941d658c5fb055c7987d4f7fe5289da64bb92e99d8022c99ce81c3 +size 1000956320 diff --git a/Qwen_Qwen3-1.7B-Q3_K_XL.gguf b/Qwen_Qwen3-1.7B-Q3_K_XL.gguf new file mode 100644 index 0000000..283e6f5 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q3_K_XL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0d33ed05b725dacd31f0bf9056ca6ef6e75558da735cfb5f95b23d85f531f92b +size 1409474976 diff --git a/Qwen_Qwen3-1.7B-Q4_0.gguf b/Qwen_Qwen3-1.7B-Q4_0.gguf new file mode 100644 index 0000000..994bacb --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c470091d31c4ada174ee5c2547daa020e930593cbca5ca8ca385ce8ff59a2fdf +size 1231813024 diff --git a/Qwen_Qwen3-1.7B-Q4_1.gguf b/Qwen_Qwen3-1.7B-Q4_1.gguf new file mode 100644 index 0000000..f6c47ab --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c0285c63df86c805ec1837441ee4fcb44bc7baa7371061bcbe40036a0ce93dfe +size 1336981920 diff --git a/Qwen_Qwen3-1.7B-Q4_K_L.gguf b/Qwen_Qwen3-1.7B-Q4_K_L.gguf new file mode 100644 index 0000000..7ec4f45 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q4_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:82479bad9c6e9d915c7bb6dd7b99db0c550e27d5d94d32d7bd4c7d02309713cc +size 1513382304 diff --git a/Qwen_Qwen3-1.7B-Q4_K_M.gguf b/Qwen_Qwen3-1.7B-Q4_K_M.gguf new file mode 100644 index 0000000..78534f7 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:72c5c3cb38fa32d5256e2fe30d03e7a64c6c79e668ad84057e3bd66e250b24fb +size 1282439584 diff --git a/Qwen_Qwen3-1.7B-Q4_K_S.gguf b/Qwen_Qwen3-1.7B-Q4_K_S.gguf new file mode 100644 index 0000000..be67834 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9ad5c47ea747ef747d2941ae588f3770884203efb1ba517c5a5f1418af66296d +size 1235220896 diff --git a/Qwen_Qwen3-1.7B-Q5_K_L.gguf b/Qwen_Qwen3-1.7B-Q5_K_L.gguf new file mode 100644 index 0000000..5f00609 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q5_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6842a8923337690e7e90daa7450351906a3a02547638833d5c6303fb768a1422 +size 1663852960 diff --git a/Qwen_Qwen3-1.7B-Q5_K_M.gguf b/Qwen_Qwen3-1.7B-Q5_K_M.gguf new file mode 100644 index 0000000..21324a2 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4287aca1b231f27dbd20012c4bf9693c89b6c849dc02b7027096da54564d4037 +size 1471805856 diff --git a/Qwen_Qwen3-1.7B-Q5_K_S.gguf b/Qwen_Qwen3-1.7B-Q5_K_S.gguf new file mode 100644 index 0000000..4d9a969 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d5375dde435a489958730b8cde82f645506903439c705b845c7fc9e32188dc9c +size 1444510112 diff --git a/Qwen_Qwen3-1.7B-Q6_K.gguf b/Qwen_Qwen3-1.7B-Q6_K.gguf new file mode 100644 index 0000000..c7f7b75 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:95aec3c8e76caf949b5a7a3b02adbb7e307eb0aa55880f6a6f9fb5f46abe6d4f +size 1673007520 diff --git a/Qwen_Qwen3-1.7B-Q6_K_L.gguf b/Qwen_Qwen3-1.7B-Q6_K_L.gguf new file mode 100644 index 0000000..00374ff --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q6_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4d87f3f55563b5e38e000e64cd67b315fe705bc1748ea27f9bbf2a88d72e2aa6 +size 1823728032 diff --git a/Qwen_Qwen3-1.7B-Q8_0.gguf b/Qwen_Qwen3-1.7B-Q8_0.gguf new file mode 100644 index 0000000..04a4319 --- /dev/null +++ b/Qwen_Qwen3-1.7B-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:74bb7c53538ab2cc81b93f0c64da14a503159de68cff3c6770428d3850479db3 +size 2165039520 diff --git a/Qwen_Qwen3-1.7B-bf16.gguf b/Qwen_Qwen3-1.7B-bf16.gguf new file mode 100644 index 0000000..6dbd1d7 --- /dev/null +++ b/Qwen_Qwen3-1.7B-bf16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:199b4df12194e24ac097d4fcbd279ce62bd4959bed9f0d4719d05a6ab1501861 +size 4069679264 diff --git a/Qwen_Qwen3-1.7B.imatrix b/Qwen_Qwen3-1.7B.imatrix new file mode 100644 index 0000000..8f4e339 Binary files /dev/null and b/Qwen_Qwen3-1.7B.imatrix differ diff --git a/README.md b/README.md new file mode 100644 index 0000000..967039c --- /dev/null +++ b/README.md @@ -0,0 +1,170 @@ +--- +quantized_by: bartowski +pipeline_tag: text-generation +base_model: Qwen/Qwen3-1.7B +base_model_relation: quantized +--- + +## Llamacpp imatrix Quantizations of Qwen3-1.7B by Qwen + +Using llama.cpp release b5200 for quantization. + +Original model: https://huggingface.co/Qwen/Qwen3-1.7B + +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 | +| -------- | ---------- | --------- | ----- | ----------- | +| [Qwen3-1.7B-bf16.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-bf16.gguf) | bf16 | 4.07GB | false | Full BF16 weights. | +| [Qwen3-1.7B-Q8_0.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q8_0.gguf) | Q8_0 | 2.17GB | false | Extremely high quality, generally unneeded but max available quant. | +| [Qwen3-1.7B-Q6_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q6_K_L.gguf) | Q6_K_L | 1.82GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. | +| [Qwen3-1.7B-Q6_K.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q6_K.gguf) | Q6_K | 1.67GB | false | Very high quality, near perfect, *recommended*. | +| [Qwen3-1.7B-Q5_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q5_K_L.gguf) | Q5_K_L | 1.66GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. | +| [Qwen3-1.7B-Q4_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q4_K_L.gguf) | Q4_K_L | 1.51GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. | +| [Qwen3-1.7B-Q5_K_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q5_K_M.gguf) | Q5_K_M | 1.47GB | false | High quality, *recommended*. | +| [Qwen3-1.7B-Q5_K_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q5_K_S.gguf) | Q5_K_S | 1.44GB | false | High quality, *recommended*. | +| [Qwen3-1.7B-Q3_K_XL.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q3_K_XL.gguf) | Q3_K_XL | 1.41GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. | +| [Qwen3-1.7B-Q4_1.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q4_1.gguf) | Q4_1 | 1.34GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. | +| [Qwen3-1.7B-Q4_K_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q4_K_M.gguf) | Q4_K_M | 1.28GB | false | Good quality, default size for most use cases, *recommended*. | +| [Qwen3-1.7B-Q4_K_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q4_K_S.gguf) | Q4_K_S | 1.24GB | false | Slightly lower quality with more space savings, *recommended*. | +| [Qwen3-1.7B-Q4_0.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q4_0.gguf) | Q4_0 | 1.23GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. | +| [Qwen3-1.7B-IQ4_NL.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-IQ4_NL.gguf) | IQ4_NL | 1.23GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. | +| [Qwen3-1.7B-IQ4_XS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-IQ4_XS.gguf) | IQ4_XS | 1.18GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. | +| [Qwen3-1.7B-Q2_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q2_K_L.gguf) | Q2_K_L | 1.18GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. | +| [Qwen3-1.7B-Q3_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q3_K_L.gguf) | Q3_K_L | 1.14GB | false | Lower quality but usable, good for low RAM availability. | +| [Qwen3-1.7B-Q3_K_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q3_K_M.gguf) | Q3_K_M | 1.07GB | false | Low quality. | +| [Qwen3-1.7B-IQ3_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-IQ3_M.gguf) | IQ3_M | 1.03GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. | +| [Qwen3-1.7B-Q3_K_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q3_K_S.gguf) | Q3_K_S | 1.00GB | false | Low quality, not recommended. | +| [Qwen3-1.7B-IQ3_XS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-IQ3_XS.gguf) | IQ3_XS | 0.97GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. | +| [Qwen3-1.7B-IQ3_XXS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-IQ3_XXS.gguf) | IQ3_XXS | 0.89GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. | +| [Qwen3-1.7B-Q2_K.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-Q2_K.gguf) | Q2_K | 0.88GB | false | Very low quality but surprisingly usable. | +| [Qwen3-1.7B-IQ2_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-1.7B-GGUF/blob/main/Qwen_Qwen3-1.7B-IQ2_M.gguf) | IQ2_M | 0.83GB | 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/Qwen_Qwen3-1.7B-GGUF --include "Qwen_Qwen3-1.7B-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_Qwen3-1.7B-GGUF --include "Qwen_Qwen3-1.7B-Q8_0/*" --local-dir ./ +``` + +You can either specify a new local-dir (Qwen_Qwen3-1.7B-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