Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* 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
+mmproj-Qwen3.8-27B-BF16.gguf filter=lfs diff=lfs merge=lfs -text
+Qwen3.8-27B-IQ1_M.gguf filter=lfs diff=lfs merge=lfs -text

CHECKSUMS.sha256

@@ -0,0 +1,18 @@
+06f619fd5f62f861c35c7d2048d97c4a7e940a179751444dbe59f3ebf711d1ea  DRY_RUN_LOG.txt
+8f7ae4307833a8729f5aedd84bd4b49473361169ba229c2846a6a4bf203cf6df  GGUF_REPORT.json
+3d8b95cd62f297c15995553e6eb9d7b888eed7baff8949644e2462aceb4ed732  INDEX.json
+bbedc3fda3305820b977265f01b8619d87570a6739de3a5582c3464840f1e57a  LICENSE
+b039a0ef227448f302e064f82dd23623a208c5dd4e099e00ad7d95077d9884fe  MMPROJ_GGUF_REPORT.json
+24119b9a81b0333ce7999b478b03ec4edaa9f6af8391e698adf231b3fc2f2525  MTP_Q4K_AUDIT.json
+01a53c23afd4ed81b79ffc697b1c4a4b83443253b6095732eee3f8fc58e96b40  MTP_n_layer_all.patch
+e298d5076e2a59d276771240891410c67892b6a0f88cd1f93a71dbf9b41e9867  PROVENANCE.json
+d3be95534adfb341b17b2fb07380853b3bc8fb38b6fdf047dd4a072629ac62d6  PROVENANCE.md
+0df42d2e8bddbc728b877c6b77bc4a3b36079076bfe0abf755abe7bf3d25cbb6  QUANTIZATION_LOG.txt
+518f096a995976518879770768658662a8f72d6f599984ceea5e826a6e6fb6f2  QUANTIZER_BUILD_PROVENANCE.txt
+d27b0cac0f4253d13df4e6692d8f7267a80c43d216f6ffe5fff6c2d150d283f6  QUANTIZER_TOOLCHAIN.json
+131cdf5c1c4b547081543382b00434e9ebf3f8eb369ef3714550086074f80bdf  Qwen3.8-27B-IQ1_M.gguf
+ce00b7a0caa6b0c33eba1bf71460804ff40b91f164f40ff4b5c491e39d8c5ec0  Qwen3.8-27B-IQ1_M.gguf.validation.json
+77db64670195a081a248cd1d84dedd9a8968f36d437c881534f8547717fbd5cc  README.md
+af0eee1cf269eb98478d5077a8bc52de6465e2a1415cb432f86c6a1529bfb284  SOURCE_DOWNLOAD_MANIFEST.json
+ee2b07a31d8c8811a852f71a748c29b54e3dc599e7cf0d7cdb5e217723e12ec2  TENSOR_TYPE_OVERRIDES.txt
+83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53  mmproj-Qwen3.8-27B-BF16.gguf

GGUF_REPORT.json

@@ -0,0 +1,45 @@
+{
+  "architecture": "qwen35",
+  "errors": [],
+  "max_tensor_end": 7870069760,
+  "ok": true,
+  "path": "/mnt/geth-vol1/qwen38_out/Qwen3.8-27B-IQ1_M.gguf",
+  "selected_metadata": {
+    "general.architecture": "qwen35",
+    "general.file_type": 31,
+    "general.name": "Qwen3.8-27B",
+    "general.quantization_version": 2,
+    "general.size_label": "27B",
+    "general.type": "model",
+    "qwen35.block_count": 65,
+    "qwen35.context_length": 262144,
+    "qwen35.embedding_length": 5120
+  },
+  "size_bytes": 7870069760,
+  "tensor_count": 866,
+  "tensor_type_bytes": {
+    "F32": 10686464,
+    "IQ1_M": 4491755520,
+    "IQ2_XXS": 129761280,
+    "Q2_K": 651141120,
+    "Q4_K": 1701642240,
+    "Q5_K": 874086400
+  },
+  "tensor_type_count": {
+    "F32": 360,
+    "IQ1_M": 408,
+    "IQ2_XXS": 16,
+    "Q2_K": 9,
+    "Q4_K": 72,
+    "Q5_K": 1
+  },
+  "tensor_type_params": {
+    "F32": 2671616,
+    "IQ1_M": 20533739520,
+    "IQ2_XXS": 503316480,
+    "Q2_K": 1984430080,
+    "Q4_K": 3025141760,
+    "Q5_K": 1271398400
+  },
+  "total_parameters": 27320697856
+}

INDEX.json

@@ -0,0 +1,110 @@
+{
+  "base_model": "Qwen/Qwen3.8-27B",
+  "created_utc": "2026-08-18T00:16:18Z",
+  "model": {
+    "filename": "Qwen3.8-27B-IQ1_M.gguf",
+    "parameters": 27320697856,
+    "sha256": "131cdf5c1c4b547081543382b00434e9ebf3f8eb369ef3714550086074f80bdf",
+    "size": 7870069760,
+    "whole_file_bpw": 2.3045003612956028
+  },
+  "projector": {
+    "filename": "mmproj-Qwen3.8-27B-BF16.gguf",
+    "sha256": "83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53",
+    "size": 931146432
+  },
+  "quantization_policy": {
+    "mtp_fallback_fraction": 0.015544012903269816,
+    "mtp_fallback_parameters": 424673280,
+    "mtp_fallback_tensors": [
+      "blk.64.attn_k.weight",
+      "blk.64.attn_output.weight",
+      "blk.64.attn_q.weight",
+      "blk.64.attn_v.weight",
+      "blk.64.ffn_down.weight",
+      "blk.64.ffn_gate.weight",
+      "blk.64.ffn_up.weight",
+      "blk.64.nextn.eh_proj.weight"
+    ],
+    "mtp_fallback_type": "Q4_K",
+    "override_rule": "blk[.]64[.].*=q4_k",
+    "override_sha256": "ee2b07a31d8c8811a852f71a748c29b54e3dc599e7cf0d7cdb5e217723e12ec2",
+    "primary_type": "IQ1_M"
+  },
+  "source_repository": "unsloth/Qwen3.8-27B-GGUF",
+  "source_revision": "f1bfb127c64f7072bdd2cad55f258b9c8b2910fe",
+  "toolchain": {
+    "llama_cpp_base_commit": "62061f91088281e65071cc38c5f69ee95c39f14e",
+    "metadata_sha256": "d27b0cac0f4253d13df4e6692d8f7267a80c43d216f6ffe5fff6c2d150d283f6",
+    "patch_sha256": "01a53c23afd4ed81b79ffc697b1c4a4b83443253b6095732eee3f8fc58e96b40",
+    "quantizer_sha256": "5d3a8456974b28569322dea7ee33941e3c9f09750e100dfce358326110b29187",
+    "run_script_sha256": "9c44f21e3d6317dad6160b79a31400b8159785cb7b5813c386655736bc321daf",
+    "upstream_fix_merge_commit": "b3ce5cedf4c007b78a45befe839fa3abada03c0b",
+    "upstream_fix_pr": 24986
+  },
+  "validation": {
+    "checks": {
+      "architecture": {
+        "ok": true,
+        "value": "qwen35"
+      },
+      "gguf_hash": {
+        "ok": true,
+        "rc": 0,
+        "tail": "sha256    03e23858f3daf5794e968a4cb48ffed5d1b1bc6f4b4565ccf4eaa76d889fe0ba  /mnt/geth-vol1/qwen38_out/Qwen3.8-27B-IQ1_M.gguf\n"
+      },
+      "gguf_structure": {
+        "ok": true,
+        "rc": 0,
+        "tail": "ms = 2, ne = (5120, 1024, 1, 1), name = blk.63.attn_v.weight, data = 0x73d35fb2b9b0\nblk.63.attn_v.weight data[:10] : 0.000000 -61206667715034480640.000000 -53298240.000000 -336439519620563392423160593572360617984.000000 -0.000000 340143099973783539667810026250043392.000000 260816666624.000000 -0.000000 -0.000000 31278922549570343796736.000000 \n\ngguf_ex_read_1: reading tensor 847 data\ngguf_ex_read_1: tensor[847]: n_dims = 2, ne = (17408, 5120, 1, 1), name = blk.63.ffn_down.weight, data = 0x73d35fdfb9b0\nblk.63.ffn_down.weight data[:10] : 1.579176 0.000000 907486167040.000000 0.000228 0.000000 174.888931 -30085307675661518343635894462840832.000000 0.000000 -3220002832384.000000 -0.000000 \n\ngguf_ex_read_1: reading tensor 848 data\ngguf_ex_read_1: tensor[848]: n_dims = 2, ne = (5120, 17408, 1, 1), name = blk.63.ffn_gate.weight, data = 0x73d3610939b0\nblk.63.ffn_gate.weight data[:10] : -139671229077918688194068480.000000 0.017403 34851.042969 -1222996110550339282620383232.000000 -0.000000 -57473286144.000000 -0.000000 0.000277 -0.000000 -192248972461473792.000000 \n\ngguf_ex_read_1: reading tensor 849 data\ngguf_ex_read_1: tensor[849]: n_dims = 2, ne = (5120, 17408, 1, 1), name = blk.63.ffn_up.weight, data = 0x73d36232b9b0\nblk.63.ffn_up.weight data[:10] : 0.000000 -178719968003268542464.000000 0.000000 3308479178082405362696192.000000 0.000000 -0.000069 -4658421132768507204730880.000000 -37492529948254162386944.000000 8719009579008.000000 -0.000000 \n\ngguf_ex_read_1: reading tensor 850 data\ngguf_ex_read_1: tensor[850]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.63.post_attention_norm.weight, data = 0x73d3635c39b0\nblk.63.post_attention_norm.weight data[:10] : 1.245117 1.146484 1.209961 1.213867 1.102051 1.255859 1.275391 1.232422 1.215820 1.231445 \n\ngguf_ex_read_1: reading tensor 851 data\ngguf_ex_read_1: tensor[851]: n_dims = 2, ne = (5120, 1024, 1, 1), name = blk.64.attn_k.weight, data = 0x73d3635c89b0\nblk.64.attn_k.weight data[:10] : 0.000000 -0.000000 -0.000000 77783911728075348258548875264.000000 -0.000000 -779807621120.000000 -116419366345506816.000000 -0.000000 -3565253426151424.000000 -0.341793 \n\ngguf_ex_read_1: reading tensor 852 data\ngguf_ex_read_1: tensor[852]: n_dims = 1, ne = (256, 1, 1, 1), name = blk.64.attn_k_norm.weight, data = 0x73d3638989b0\nblk.64.attn_k_norm.weight data[:10] : 1.308594 0.885742 1.699219 1.263672 1.605469 1.941406 1.550781 1.714844 1.310547 1.710938 \n\ngguf_ex_read_1: reading tensor 853 data\ngguf_ex_read_1: tensor[853]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.attn_norm.weight, data = 0x73d363898db0\nblk.64.attn_norm.weight data[:10] : 1.011536 1.086914 1.051514 1.046631 1.306641 1.049316 1.024048 1.145508 1.234375 1.024536 \n\ngguf_ex_read_1: reading tensor 854 data\ngguf_ex_read_1: tensor[854]: n_dims = 2, ne = (6144, 5120, 1, 1), name = blk.64.attn_output.weight, data = 0x73d36389ddb0\nblk.64.attn_output.weight data[:10] : 0.000000 -nan -0.000231 -588164655477598002671369217638400.000000 0.000000 -0.000000 -70893914414901329129066281568305152.000000 -0.000000 -0.000000 427431294489308781141138740346880.000000 \n\ngguf_ex_read_1: reading tensor 855 data\ngguf_ex_read_1: tensor[855]: n_dims = 2, ne = (5120, 12288, 1, 1), name = blk.64.attn_q.weight, data = 0x73d36497ddb0\nblk.64.attn_q.weight data[:10] : 0.000000 -0.000479 544778361270436276207616.000000 -42490689775796476996741959936580255744.000000 -0.000000 0.000209 0.000000 -1730.768677 -0.000000 -7098020793417728.000000 \n\ngguf_ex_read_1: reading tensor 856 data\ngguf_ex_read_1: tensor[856]: n_dims = 1, ne = (256, 1, 1, 1), name = blk.64.attn_q_norm.weight, data = 0x73d366b3ddb0\nblk.64.attn_q_norm.weight data[:10] : 1.244141 1.072754 1.816406 1.253906 1.605469 1.941406 1.494141 1.863281 1.298828 1.550781 \n\ngguf_ex_read_1: reading tensor 857 data\ngguf_ex_read_1: tensor[857]: n_dims = 2, ne = (5120, 1024, 1, 1), name = blk.64.attn_v.weight, data = 0x73d366b3e1b0\nblk.64.attn_v.weight data[:10] : 0.000000 -0.000000 -0.000000 -0.000000 5447.940430 -14293357591483383808.000000 0.000000 5937133812701839162740289447133184.000000 -0.000000 75039989832789001467068416.000000 \n\ngguf_ex_read_1: reading tensor 858 data\ngguf_ex_read_1: tensor[858]: n_dims = 2, ne = (17408, 5120, 1, 1), name = blk.64.ffn_down.weight, data = 0x73d366e0e1b0\nblk.64.ffn_down.weight data[:10] : 0.000000 -10061968781293743379290446102528.000000 -0.000000 -0.000000 0.000000 225828.859375 0.000000 -1020286891632021366355436645646336.000000 0.000000 51912808.000000 \n\ngguf_ex_read_1: reading tensor 859 data\ngguf_ex_read_1: tensor[859]: n_dims = 2, ne = (5120, 17408, 1, 1), name = blk.64.ffn_gate.weight, data = 0x73d369dde1b0\nblk.64.ffn_gate.weight data[:10] : 0.000000 -39567656087807327807204908597248.000000 -8563729381619400704.000000 -100886693044704566552415174727269613568.000000 1195184917292318819663597264699392.000000 -0.000016 -207488335099243366305018937344.000000 67749950533943957651456.000000 -0.000000 457139.312500 \n\ngguf_ex_read_1: reading tensor 860 data\ngguf_ex_read_1: tensor[860]: n_dims = 2, ne = (5120, 17408, 1, 1), name = blk.64.ffn_up.weight, data = 0x73d36cdae1b0\nblk.64.ffn_up.weight data[:10] : 0.000000 -0.000000 2250904178786304.000000 -13267505927444837330583552.000000 -0.000000 -0.000240 -0.000000 270797587349504.000000 18843456574607247956150831967895552.000000 944346309187725626185875456.000000 \n\ngguf_ex_read_1: reading tensor 861 data\ngguf_ex_read_1: tensor[861]: n_dims = 2, ne = (10240, 5120, 1, 1), name = blk.64.nextn.eh_proj.weight, data = 0x73d36fd7e1b0\nblk.64.nextn.eh_proj.weight data[:10] : 0.000000 -0.000000 -144442774580488843072372736.000000 8119155712.000000 76078228243781219482940852409089392640.000000 17060836274248755462719215239168.000000 -0.000000 0.197302 -0.000000 256557.281250 \n\ngguf_ex_read_1: reading tensor 862 data\ngguf_ex_read_1: tensor[862]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.nextn.enorm.weight, data = 0x73d37199e1b0\nblk.64.nextn.enorm.weight data[:10] : 0.410156 0.511719 0.728516 0.554688 0.621094 0.697266 0.390625 0.658203 0.509766 0.691406 \n\ngguf_ex_read_1: reading tensor 863 data\ngguf_ex_read_1: tensor[863]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.nextn.hnorm.weight, data = 0x73d3719a31b0\nblk.64.nextn.hnorm.weight data[:10] : 0.762695 0.875000 0.784180 0.835938 0.935059 0.822266 0.773438 0.822266 0.830078 0.865234 \n\ngguf_ex_read_1: reading tensor 864 data\ngguf_ex_read_1: tensor[864]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.nextn.shared_head_norm.weight, data = 0x73d3719a81b0\nblk.64.nextn.shared_head_norm.weight data[:10] : 2.312500 1.972656 2.117188 2.093750 1.593750 2.257812 2.296875 2.085938 2.203125 2.015625 \n\ngguf_ex_read_1: reading tensor 865 data\ngguf_ex_read_1: tensor[865]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.post_attention_norm.weight, data = 0x73d3719ad1b0\nblk.64.post_attention_norm.weight data[:10] : 1.273438 1.210938 1.263672 1.249023 1.425781 1.261719 1.289062 1.310547 1.371094 1.204102 \n\ngguf_ex_read_1: ctx_data size: 7859392080\n"
+      },
+      "magic": {
+        "ok": true,
+        "value": "GGUF"
+      },
+      "mmproj_hash": {
+        "ok": true,
+        "rc": 0,
+        "tail": "sha256    c03a8b045fd65613ca56acc139c6112eef41e74f7d7290e9b29657c087098230  /mnt/geth-vol1/qwen38_assets/mmproj-Qwen3.8-27B-BF16.gguf\n"
+      },
+      "mmproj_structure": {
+        "ok": true,
+        "rc": 0,
+        "tail": "sor 309 data\ngguf_ex_read_1: tensor[309]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.8.ln1.weight, data = 0x7877fde59430\nv.blk.8.ln1.weight data[:10] : 1.000000 0.796875 0.843750 0.648438 1.007812 0.847656 0.957031 0.839844 1.171875 0.941406 \n\ngguf_ex_read_1: reading tensor 310 data\ngguf_ex_read_1: tensor[310]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.8.ln2.bias, data = 0x7877fde5a630\nv.blk.8.ln2.bias data[:10] : 0.210938 -0.180664 -0.176758 0.008301 0.048584 0.043945 0.041992 0.069336 -0.021484 0.056396 \n\ngguf_ex_read_1: reading tensor 311 data\ngguf_ex_read_1: tensor[311]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.8.ln2.weight, data = 0x7877fde5b830\nv.blk.8.ln2.weight data[:10] : 1.132812 1.078125 0.949219 0.832031 1.101562 1.078125 1.062500 1.179688 1.304688 1.062500 \n\ngguf_ex_read_1: reading tensor 312 data\ngguf_ex_read_1: tensor[312]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.attn_out.bias, data = 0x7877fde5ca30\nv.blk.9.attn_out.bias data[:10] : 0.028198 -0.052002 0.011719 0.073730 0.086426 0.068359 -0.188477 0.204102 -0.037842 0.077148 \n\ngguf_ex_read_1: reading tensor 313 data\ngguf_ex_read_1: tensor[313]: n_dims = 2, ne = (1152, 1152, 1, 1), name = v.blk.9.attn_out.weight, data = 0x7877fde5dc30\nv.blk.9.attn_out.weight data[:10] : 0.001108 0.012679 0.003055 0.004982 -0.010421 -0.014663 0.019498 0.000313 0.018583 0.010787 \n\ngguf_ex_read_1: reading tensor 314 data\ngguf_ex_read_1: tensor[314]: n_dims = 1, ne = (3456, 1, 1, 1), name = v.blk.9.attn_qkv.bias, data = 0x7877fe0e5c30\nv.blk.9.attn_qkv.bias data[:10] : 0.098145 -0.263672 -0.648438 -0.132812 -0.222656 -0.182617 -5.500000 -0.042969 -0.056885 0.024170 \n\ngguf_ex_read_1: reading tensor 315 data\ngguf_ex_read_1: tensor[315]: n_dims = 2, ne = (1152, 3456, 1, 1), name = v.blk.9.attn_qkv.weight, data = 0x7877fe0e9230\nv.blk.9.attn_qkv.weight data[:10] : 0.029020 -0.014846 -0.002666 0.000901 0.001493 0.042171 -0.016447 0.013137 -0.018095 -0.021391 \n\ngguf_ex_read_1: reading tensor 316 data\ngguf_ex_read_1: tensor[316]: n_dims = 1, ne = (4304, 1, 1, 1), name = v.blk.9.ffn_up.bias, data = 0x7877fe881230\nv.blk.9.ffn_up.bias data[:10] : -1.101562 -1.773438 -1.843750 -2.281250 -2.421875 -0.218750 -1.265625 -1.953125 -0.832031 -1.523438 \n\ngguf_ex_read_1: reading tensor 317 data\ngguf_ex_read_1: tensor[317]: n_dims = 2, ne = (1152, 4304, 1, 1), name = v.blk.9.ffn_up.weight, data = 0x7877fe885570\nv.blk.9.ffn_up.weight data[:10] : -0.002498 -0.005683 -0.001531 0.002743 0.003376 -0.011672 0.004371 0.027128 -0.011061 0.023100 \n\ngguf_ex_read_1: reading tensor 318 data\ngguf_ex_read_1: tensor[318]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ffn_down.bias, data = 0x7877ff1fa570\nv.blk.9.ffn_down.bias data[:10] : -0.194336 -0.386719 -0.068359 -0.149414 0.188477 -0.074707 0.335938 0.082520 0.063965 -0.017578 \n\ngguf_ex_read_1: reading tensor 319 data\ngguf_ex_read_1: tensor[319]: n_dims = 2, ne = (4304, 1152, 1, 1), name = v.blk.9.ffn_down.weight, data = 0x7877ff1fb770\nv.blk.9.ffn_down.weight data[:10] : 0.005439 0.005607 -0.010055 -0.009353 0.004051 0.005958 0.008925 0.006217 0.002064 0.001108 \n\ngguf_ex_read_1: reading tensor 320 data\ngguf_ex_read_1: tensor[320]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ln1.bias, data = 0x7877ffb70770\nv.blk.9.ln1.bias data[:10] : -0.038086 -0.055420 -0.062500 -0.006104 0.062988 0.094727 0.161133 -0.009949 -0.194336 -0.058594 \n\ngguf_ex_read_1: reading tensor 321 data\ngguf_ex_read_1: tensor[321]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ln1.weight, data = 0x7877ffb71970\nv.blk.9.ln1.weight data[:10] : 1.078125 0.968750 0.890625 0.726562 1.164062 0.972656 1.093750 1.007812 1.421875 1.039062 \n\ngguf_ex_read_1: reading tensor 322 data\ngguf_ex_read_1: tensor[322]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ln2.bias, data = 0x7877ffb72b70\nv.blk.9.ln2.bias data[:10] : -0.223633 -0.011475 -0.066406 -0.148438 -0.049316 0.095215 0.494141 -0.287109 -0.175781 -0.194336 \n\ngguf_ex_read_1: reading tensor 323 data\ngguf_ex_read_1: tensor[323]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ln2.weight, data = 0x7877ffb73d70\nv.blk.9.ln2.weight data[:10] : 1.171875 1.195312 1.000000 0.914062 1.203125 1.234375 1.039062 1.117188 1.421875 1.203125 \n\ngguf_ex_read_1: reading tensor 324 data\ngguf_ex_read_1: tensor[324]: n_dims = 1, ne = (4608, 1, 1, 1), name = mm.0.bias, data = 0x7877ffb74f70\nmm.0.bias data[:10] : 0.024536 0.007996 0.020142 0.022705 0.024170 0.017456 0.020386 0.025879 0.035156 0.018311 \n\ngguf_ex_read_1: reading tensor 325 data\ngguf_ex_read_1: tensor[325]: n_dims = 2, ne = (4608, 4608, 1, 1), name = mm.0.weight, data = 0x7877ffb79770\nmm.0.weight data[:10] : -0.020231 -0.012984 -0.011489 0.003628 -0.013350 -0.005119 0.003124 0.018949 -0.001943 -0.014998 \n\ngguf_ex_read_1: reading tensor 326 data\ngguf_ex_read_1: tensor[326]: n_dims = 1, ne = (5120, 1, 1, 1), name = mm.2.bias, data = 0x7878023f9770\nmm.2.bias data[:10] : 0.009338 0.000195 -0.007538 0.005615 -0.010071 -0.008667 0.028687 -0.018799 0.015076 0.000706 \n\ngguf_ex_read_1: reading tensor 327 data\ngguf_ex_read_1: tensor[327]: n_dims = 2, ne = (4608, 5120, 1, 1), name = mm.2.weight, data = 0x7878023fe770\nmm.2.weight data[:10] : 0.006339 0.027616 -0.002262 0.002880 -0.004585 -0.004310 0.003974 0.006446 0.011275 -0.022978 \n\ngguf_ex_read_1: reading tensor 328 data\ngguf_ex_read_1: tensor[328]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.post_ln.bias, data = 0x7878050fe770\nv.post_ln.bias data[:10] : 0.010437 0.021973 -0.063477 0.167969 -0.085449 -0.233398 -0.154297 -0.080078 -0.121582 -0.065430 \n\ngguf_ex_read_1: reading tensor 329 data\ngguf_ex_read_1: tensor[329]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.post_ln.weight, data = 0x7878050ff970\nv.post_ln.weight data[:10] : 1.335938 1.273438 1.257812 1.343750 1.437500 1.210938 1.320312 1.351562 1.203125 1.390625 \n\ngguf_ex_read_1: reading tensor 330 data\ngguf_ex_read_1: tensor[330]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.patch_embd.bias, data = 0x787805100b70\nv.patch_embd.bias data[:10] : 0.107910 -0.341797 0.000954 -0.163086 0.081055 -0.076172 0.043213 -0.318359 0.045166 -0.392578 \n\ngguf_ex_read_1: reading tensor 331 data\ngguf_ex_read_1: tensor[331]: n_dims = 4, ne = (16, 16, 3, 1152), name = v.patch_embd.weight, data = 0x787805101d70\nv.patch_embd.weight data[:10] : 0.008179 -0.007172 -0.004852 -0.011902 -0.004822 0.010376 -0.008911 0.000607 -0.001587 0.024658 \n\ngguf_ex_read_1: reading tensor 332 data\ngguf_ex_read_1: tensor[332]: n_dims = 4, ne = (16, 16, 3, 1152), name = v.patch_embd.weight.1, data = 0x787805461d70\nv.patch_embd.weight.1 data[:10] : 0.009277 -0.006104 -0.003723 -0.010681 -0.003418 0.012024 -0.007751 0.001305 -0.000683 0.025879 \n\ngguf_ex_read_1: reading tensor 333 data\ngguf_ex_read_1: tensor[333]: n_dims = 2, ne = (1152, 2304, 1, 1), name = v.position_embd.weight, data = 0x7878057c1d70\nv.position_embd.weight data[:10] : -0.010620 -0.004089 -0.074707 0.092285 -0.141602 0.002914 -0.005249 0.100586 -0.010498 0.298828 \n\ngguf_ex_read_1: ctx_data size: 931249488\n"
+      },
+      "projector_identity": {
+        "actual_sha256": "83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53",
+        "actual_size_bytes": 931146432,
+        "expected_sha256": "83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53",
+        "expected_size_bytes": 931146432,
+        "ok": true
+      },
+      "text_one_token": {
+        "ok": true,
+        "rc": 0,
+        "soft": true,
+        "tail": "warning: no usable GPU found, --gpu-layers option will be ignored\nwarning: one possible reason is that llama.cpp was compiled without GPU support\nwarning: consult docs/build.md for compilation instructions\n--no-conversation is not supported by llama-cli\nplease use llama-completion instead\n\nLoading model... \n\n\n\u2584\u2584 \u2584\u2584\n\u2588\u2588 \u2588\u2588\n\u2588\u2588 \u2588\u2588  \u2580\u2580\u2588\u2584 \u2588\u2588\u2588\u2584\u2588\u2588\u2588\u2584  \u2580\u2580\u2588\u2584    \u2584\u2588\u2588\u2588\u2588 \u2588\u2588\u2588\u2588\u2584 \u2588\u2588\u2588\u2588\u2584\n\u2588\u2588 \u2588\u2588 \u2584\u2588\u2580\u2588\u2588 \u2588\u2588 \u2588\u2588 \u2588\u2588 \u2584\u2588\u2580\u2588\u2588    \u2588\u2588    \u2588\u2588 \u2588\u2588 \u2588\u2588 \u2588\u2588\n\u2588\u2588 \u2588\u2588 \u2580\u2588\u2584\u2588\u2588 \u2588\u2588 \u2588\u2588 \u2588\u2588 \u2580\u2588\u2584\u2588\u2588 \u2588\u2588 \u2580\u2588\u2588\u2588\u2588 \u2588\u2588\u2588\u2588\u2580 \u2588\u2588\u2588\u2588\u2580\n                                    \u2588\u2588    \u2588\u2588\n                                    \u2580\u2580    \u2580\u2580\n\nbuild      : b9591-62061f910\nmodel      : Qwen3.8-27B-IQ1_M.gguf\nmodalities : text\n\navailable commands:\n  /exit or Ctrl+C     stop or exit\n  /regen              regenerate the last response\n  /clear              clear the chat history\n  /read <file>        add a text file\n  /glob <pattern>     add text files using globbing pattern\n\n\n> 1\n\n[Start thinking]\nThe\n\nExiting...\n",
+        "timeout_only_soft": true
+      },
+      "vision_one_token": {
+        "ok": true,
+        "rc": 0,
+        "soft": true,
+        "tail": "warning: no usable GPU found, --gpu-layers option will be ignored\nwarning: one possible reason is that llama.cpp was compiled without GPU support\nwarning: consult docs/build.md for compilation instructions\n",
+        "timeout_only_soft": true
+      }
+    },
+    "label": "IQ1_M-direct-BF16-MTP-Q4_K",
+    "mmproj": {
+      "path": "/mnt/geth-vol1/qwen38_assets/mmproj-Qwen3.8-27B-BF16.gguf",
+      "sha256": "83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53",
+      "size_bytes": 931146432
+    },
+    "ok": true,
+    "path": "/mnt/geth-vol1/qwen38_out/Qwen3.8-27B-IQ1_M.gguf",
+    "sha256": "131cdf5c1c4b547081543382b00434e9ebf3f8eb369ef3714550086074f80bdf",
+    "size_bytes": 7870069760,
+    "ts": 1787011283.288072,
+    "warnings": []
+  }
+}

LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 Alibaba Cloud
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

MMPROJ_GGUF_REPORT.json

@@ -0,0 +1,32 @@
+{
+  "architecture": "clip",
+  "errors": [],
+  "max_tensor_end": 931146432,
+  "ok": true,
+  "path": "/mnt/geth-vol1/qwen38_inputs/unsloth/mmproj-BF16.gguf",
+  "selected_metadata": {
+    "clip.has_vision_encoder": true,
+    "clip.projector_type": "qwen3vl_merger",
+    "general.architecture": "clip",
+    "general.file_type": 32,
+    "general.name": "Qwen3.8-27B",
+    "general.quantization_version": 2,
+    "general.size_label": "461M",
+    "general.type": "mmproj"
+  },
+  "size_bytes": 931146432,
+  "tensor_count": 334,
+  "tensor_type_bytes": {
+    "BF16": 911794176,
+    "F32": 19332032
+  },
+  "tensor_type_count": {
+    "BF16": 110,
+    "F32": 224
+  },
+  "tensor_type_params": {
+    "BF16": 455897088,
+    "F32": 4833008
+  },
+  "total_parameters": 460730096
+}

MTP_Q4K_AUDIT.json

@@ -0,0 +1,27 @@
+{
+  "block64_quantizable_tensors": [
+    "blk.64.attn_k.weight",
+    "blk.64.attn_output.weight",
+    "blk.64.attn_q.weight",
+    "blk.64.attn_v.weight",
+    "blk.64.ffn_down.weight",
+    "blk.64.ffn_gate.weight",
+    "blk.64.ffn_up.weight",
+    "blk.64.nextn.eh_proj.weight"
+  ],
+  "block64_target_type": "Q4_K",
+  "block64_tensor_rows": 15,
+  "dry_run_quant_size_mib": 7495.0,
+  "dry_run_whole_file_bpw": 2.3,
+  "llama_cpp_base_commit": "62061f91088281e65071cc38c5f69ee95c39f14e",
+  "llama_cpp_upstream_fix_merge_commit": "b3ce5cedf4c007b78a45befe839fa3abada03c0b",
+  "llama_cpp_upstream_fix_pr": 24986,
+  "mtp_patch_sha256": "01a53c23afd4ed81b79ffc697b1c4a4b83443253b6095732eee3f8fc58e96b40",
+  "ok": true,
+  "override_rule": "blk[.]64[.].*=q4_k",
+  "override_sha256": "ee2b07a31d8c8811a852f71a748c29b54e3dc599e7cf0d7cdb5e217723e12ec2",
+  "policy": "64 main blocks use imatrix-aware IQ1_M; uncovered auxiliary blk.64 MTP tensors use Q4_K",
+  "quantizer_sha256": "5d3a8456974b28569322dea7ee33941e3c9f09750e100dfce358326110b29187",
+  "toolchain_metadata_path": "/mnt/geth-vol1/qwen38_out/meta/iq1m-quantizer-toolchain.json",
+  "toolchain_metadata_sha256": "d27b0cac0f4253d13df4e6692d8f7267a80c43d216f6ffe5fff6c2d150d283f6"
+}

MTP_n_layer_all.patch

@@ -0,0 +1,13 @@
+diff --git a/src/llama-quant.cpp b/src/llama-quant.cpp
+index 140974dc3..172153cee 100644
+--- a/src/llama-quant.cpp
++++ b/src/llama-quant.cpp
+@@ -849,7 +849,7 @@ static void init_quantize_state_counters(quantize_state_impl & qs, std::vector<t
+             qs.has_tied_embeddings = false;
+         }
+     }
+-    qs.n_ffn_down = qs.n_ffn_gate = qs.n_ffn_up = (int)qs.model.hparams.n_layer();
++    qs.n_ffn_down = qs.n_ffn_gate = qs.n_ffn_up = (int)qs.model.hparams.n_layer_all;
+ }
+ 
+ //

PROVENANCE.json

@@ -0,0 +1,67 @@
+{
+  "architecture": "qwen35",
+  "artifacts": {
+    "Qwen3.8-27B-IQ1_M.gguf": {
+      "format": "GGUF",
+      "role": "model",
+      "sha256": "131cdf5c1c4b547081543382b00434e9ebf3f8eb369ef3714550086074f80bdf",
+      "size": 7870069760
+    },
+    "mmproj-Qwen3.8-27B-BF16.gguf": {
+      "format": "GGUF",
+      "role": "multimodal_projector",
+      "sha256": "83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53",
+      "size": 931146432
+    }
+  },
+  "base_license": "Apache-2.0",
+  "base_model": "Qwen/Qwen3.8-27B",
+  "base_revision": "1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0",
+  "bf16_source": {
+    "files": [
+      {
+        "path": "BF16/Qwen3.8-27B-BF16-00001-of-00002.gguf",
+        "sha256": "b9966e82b7a4d87028b5eae061d578ee826305ebf8baea5bfc6e09bad0ba191f",
+        "size": 49986159616
+      },
+      {
+        "path": "BF16/Qwen3.8-27B-BF16-00002-of-00002.gguf",
+        "sha256": "92e3943c4f9bd6292a7bef82369f65fed9bfed088b9df0fb2fa2ce17c9edfa02",
+        "size": 4671576000
+      }
+    ],
+    "repo": "unsloth/Qwen3.8-27B-GGUF",
+    "revision": "f1bfb127c64f7072bdd2cad55f258b9c8b2910fe"
+  },
+  "parameter_count": 27320697856,
+  "quantization": {
+    "direct_from_bf16": true,
+    "mtp_fallback_fraction": 0.015544012903269816,
+    "mtp_fallback_parameters": 424673280,
+    "mtp_fallback_tensors": [
+      "blk.64.attn_k.weight",
+      "blk.64.attn_output.weight",
+      "blk.64.attn_q.weight",
+      "blk.64.attn_v.weight",
+      "blk.64.ffn_down.weight",
+      "blk.64.ffn_gate.weight",
+      "blk.64.ffn_up.weight",
+      "blk.64.nextn.eh_proj.weight"
+    ],
+    "mtp_fallback_type": "Q4_K",
+    "override_rule": "blk[.]64[.].*=q4_k",
+    "override_sha256": "ee2b07a31d8c8811a852f71a748c29b54e3dc599e7cf0d7cdb5e217723e12ec2",
+    "primary_type": "IQ1_M",
+    "whole_file_bpw": 2.3045003612956028
+  },
+  "toolchain": {
+    "llama_cpp_base_commit": "62061f91088281e65071cc38c5f69ee95c39f14e",
+    "metadata_sha256": "d27b0cac0f4253d13df4e6692d8f7267a80c43d216f6ffe5fff6c2d150d283f6",
+    "patch_sha256": "01a53c23afd4ed81b79ffc697b1c4a4b83443253b6095732eee3f8fc58e96b40",
+    "quantizer_sha256": "5d3a8456974b28569322dea7ee33941e3c9f09750e100dfce358326110b29187",
+    "run_script_sha256": "9c44f21e3d6317dad6160b79a31400b8159785cb7b5813c386655736bc321daf",
+    "upstream_fix_merge_commit": "b3ce5cedf4c007b78a45befe839fa3abada03c0b",
+    "upstream_fix_pr": 24986
+  },
+  "validation_report": "Qwen3.8-27B-IQ1_M.gguf.validation.json"
+}

PROVENANCE.md

@@ -0,0 +1,62 @@
+# Reproducible provenance
+
+Generated: `2026-08-18T00:16:18Z`
+
+## Source
+
+- Repository: `unsloth/Qwen3.8-27B-GGUF`
+- Revision: `f1bfb127c64f7072bdd2cad55f258b9c8b2910fe`
+- Conversion source: pinned two-part BF16 GGUF, not an intermediate Q8 or another lossy quantization
+
+- `BF16/Qwen3.8-27B-BF16-00001-of-00002.gguf`: `b9966e82b7a4d87028b5eae061d578ee826305ebf8baea5bfc6e09bad0ba191f` (49986159616 bytes)
+- `BF16/Qwen3.8-27B-BF16-00002-of-00002.gguf`: `92e3943c4f9bd6292a7bef82369f65fed9bfed088b9df0fb2fa2ce17c9edfa02` (4671576000 bytes)
+
+## Quantizer
+
+- llama.cpp base commit: `62061f91088281e65071cc38c5f69ee95c39f14e`
+- Official upstream MTP fix: PR `#24986`, merge commit `b3ce5cedf4c007b78a45befe839fa3abada03c0b`
+- Patched quantizer SHA256: `5d3a8456974b28569322dea7ee33941e3c9f09750e100dfce358326110b29187`
+- MTP patch SHA256: `01a53c23afd4ed81b79ffc697b1c4a4b83443253b6095732eee3f8fc58e96b40`
+- Override file SHA256: `ee2b07a31d8c8811a852f71a748c29b54e3dc599e7cf0d7cdb5e217723e12ec2`
+- Run script SHA256: `9c44f21e3d6317dad6160b79a31400b8159785cb7b5813c386655736bc321daf`
+
+The exact executable and all six loaded llama/ggml shared libraries are recorded in `QUANTIZER_TOOLCHAIN.json`.
+
+## Commands
+
+Dry run:
+
+```text
+/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/llama-quantize --imatrix /mnt/geth-vol1/qwen38_assets/Qwen3.8-27B-agentic.imatrix --tensor-type-file /mnt/geth-vol1/qwen38_assets/iq1m_mtp_q4k_overrides.txt --dry-run /root/qwen38_bf16/BF16/Qwen3.8-27B-BF16-00001-of-00002.gguf /mnt/geth-vol1/qwen38_out/Qwen3.8-27B-IQ1_M.gguf.part IQ1_M 16
+```
+
+Quantization:
+
+```text
+nice -n 5 /mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/llama-quantize --imatrix /mnt/geth-vol1/qwen38_assets/Qwen3.8-27B-agentic.imatrix --tensor-type-file /mnt/geth-vol1/qwen38_assets/iq1m_mtp_q4k_overrides.txt /root/qwen38_bf16/BF16/Qwen3.8-27B-BF16-00001-of-00002.gguf /mnt/geth-vol1/qwen38_out/Qwen3.8-27B-IQ1_M.gguf.part IQ1_M 16
+```
+
+## MTP exception
+
+The pinned importance matrix has no calibration entries for the auxiliary `blk.64` prediction block. The official `n_layer_all` accounting fix makes that extra block visible to the quantizer. Exactly the following eight two-dimensional tensors are kept at Q4_K:
+
+- `blk.64.attn_k.weight`
+- `blk.64.attn_output.weight`
+- `blk.64.attn_q.weight`
+- `blk.64.attn_v.weight`
+- `blk.64.ffn_down.weight`
+- `blk.64.ffn_gate.weight`
+- `blk.64.ffn_up.weight`
+- `blk.64.nextn.eh_proj.weight`
+
+Together they contain 424,673,280 parameters (1.5544% of the model). `MTP_Q4K_AUDIT.json` proves the dry run saw all 15 block-64 rows and assigned Q4_K to exactly these eight BF16 matrices.
+
+## Validation gates
+
+- GGUF magic and `qwen35` architecture: hard gate
+- IQ1_M present and all eight manual block-64 Q4_K override tensors present: hard gate
+- Whole-file BPW range 2.25–2.35: hard gate
+- Projector size and SHA256: hard gate
+- Text runtime probe: completed non-zero exit is a hard failure; timeout is recorded as a soft warning
+- Vision runtime probe with the pinned projector: completed non-zero exit is a hard failure; timeout is recorded as a soft warning
+- Remote Hugging Face LFS size/SHA and Range GGUF magic: hard gates performed by `upload_verify.py`

QUANTIZER_BUILD_PROVENANCE.txt

@@ -0,0 +1,7 @@
+base_commit=62061f91088281e65071cc38c5f69ee95c39f14e
+upstream_fix_commit=b3ce5cedf4c007b78a45befe839fa3abada03c0b
+upstream_pr=https://github.com/ggml-org/llama.cpp/pull/24986
+built_utc=2026-08-17T22:39:38Z
+ src/llama-quant.cpp | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+01a53c23afd4ed81b79ffc697b1c4a4b83443253b6095732eee3f8fc58e96b40  /mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/mtp-n-layer-all.patch

QUANTIZER_TOOLCHAIN.json

@@ -0,0 +1,87 @@
+{
+  "created_utc": "2026-08-17T23:14:40Z",
+  "dry_run_argv": [
+    "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/llama-quantize",
+    "--imatrix",
+    "/mnt/geth-vol1/qwen38_assets/Qwen3.8-27B-agentic.imatrix",
+    "--tensor-type-file",
+    "/mnt/geth-vol1/qwen38_assets/iq1m_mtp_q4k_overrides.txt",
+    "--dry-run",
+    "/root/qwen38_bf16/BF16/Qwen3.8-27B-BF16-00001-of-00002.gguf",
+    "/mnt/geth-vol1/qwen38_out/Qwen3.8-27B-IQ1_M.gguf.part",
+    "IQ1_M",
+    "16"
+  ],
+  "llama_cpp": {
+    "artifacts": {
+      "/mnt/geth-vol1/qwen38_assets/iq1m_mtp_q4k_overrides.txt": {
+        "resolved_path": "/mnt/geth-vol1/qwen38_assets/iq1m_mtp_q4k_overrides.txt",
+        "sha256": "ee2b07a31d8c8811a852f71a748c29b54e3dc599e7cf0d7cdb5e217723e12ec2",
+        "size": 19
+      },
+      "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libggml-base.so.0": {
+        "resolved_path": "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libggml-base.so.0.13.1",
+        "sha256": "733ef2bef04f50a948e7febc21d9acf683fd337b6268366492eb86af7a701f50",
+        "size": 913616
+      },
+      "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libggml-cpu.so.0": {
+        "resolved_path": "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libggml-cpu.so.0.13.1",
+        "sha256": "737da857881f1b790fc7450d681423aa8f0181f010ae8196977f32ae5316634d",
+        "size": 1349064
+      },
+      "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libggml.so.0": {
+        "resolved_path": "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libggml.so.0.13.1",
+        "sha256": "f99b1154804875d08d999e6c49fd79c5927179661b87c76a1420a513e1f61fa0",
+        "size": 56376
+      },
+      "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libllama-common.so.0": {
+        "resolved_path": "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libllama-common.so.0.0.1",
+        "sha256": "edc289279045d1ff3218214a05de564e1970ed9d82b34b2a20376c23a15f0c60",
+        "size": 5826584
+      },
+      "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libllama-quantize-impl.so": {
+        "resolved_path": "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libllama-quantize-impl.so",
+        "sha256": "0073c7ced92f65baa21fcb2d5cfe5430910c682d7fb2c3927acbb2d13b5e95ee",
+        "size": 89744
+      },
+      "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libllama.so.0": {
+        "resolved_path": "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/libllama.so.0.0.1",
+        "sha256": "29489bb2b7975d367e1fb5b4d8fb35acef279d4523f72aee420aa2da6db16abe",
+        "size": 3876024
+      },
+      "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/llama-quantize": {
+        "resolved_path": "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/llama-quantize",
+        "sha256": "5d3a8456974b28569322dea7ee33941e3c9f09750e100dfce358326110b29187",
+        "size": 17920
+      },
+      "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/mtp-n-layer-all.patch": {
+        "resolved_path": "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/mtp-n-layer-all.patch",
+        "sha256": "01a53c23afd4ed81b79ffc697b1c4a4b83443253b6095732eee3f8fc58e96b40",
+        "size": 484
+      }
+    },
+    "base_commit": "62061f91088281e65071cc38c5f69ee95c39f14e",
+    "fix_summary": "quant: use hparams.n_layer_all for FFN quantization counters so the auxiliary MTP layer is counted",
+    "upstream_fix_merge_commit": "b3ce5cedf4c007b78a45befe839fa3abada03c0b",
+    "upstream_fix_pr": 24986
+  },
+  "quantize_argv": [
+    "nice",
+    "-n",
+    "5",
+    "/mnt/geth-vol1/qwen38_tools/llama.cpp-62061f910-mtpfix/build/bin/llama-quantize",
+    "--imatrix",
+    "/mnt/geth-vol1/qwen38_assets/Qwen3.8-27B-agentic.imatrix",
+    "--tensor-type-file",
+    "/mnt/geth-vol1/qwen38_assets/iq1m_mtp_q4k_overrides.txt",
+    "/root/qwen38_bf16/BF16/Qwen3.8-27B-BF16-00001-of-00002.gguf",
+    "/mnt/geth-vol1/qwen38_out/Qwen3.8-27B-IQ1_M.gguf.part",
+    "IQ1_M",
+    "16"
+  ],
+  "run_script": {
+    "path": "/opt/qwen38_iq1m_release/run.sh",
+    "sha256": "9c44f21e3d6317dad6160b79a31400b8159785cb7b5813c386655736bc321daf"
+  },
+  "schema": 1
+}

Qwen3.8-27B-IQ1_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:131cdf5c1c4b547081543382b00434e9ebf3f8eb369ef3714550086074f80bdf
+size 7870069760

Qwen3.8-27B-IQ1_M.gguf.validation.json

@@ -0,0 +1,65 @@
+{
+  "checks": {
+    "architecture": {
+      "ok": true,
+      "value": "qwen35"
+    },
+    "gguf_hash": {
+      "ok": true,
+      "rc": 0,
+      "tail": "sha256    03e23858f3daf5794e968a4cb48ffed5d1b1bc6f4b4565ccf4eaa76d889fe0ba  /mnt/geth-vol1/qwen38_out/Qwen3.8-27B-IQ1_M.gguf\n"
+    },
+    "gguf_structure": {
+      "ok": true,
+      "rc": 0,
+      "tail": "ms = 2, ne = (5120, 1024, 1, 1), name = blk.63.attn_v.weight, data = 0x73d35fb2b9b0\nblk.63.attn_v.weight data[:10] : 0.000000 -61206667715034480640.000000 -53298240.000000 -336439519620563392423160593572360617984.000000 -0.000000 340143099973783539667810026250043392.000000 260816666624.000000 -0.000000 -0.000000 31278922549570343796736.000000 \n\ngguf_ex_read_1: reading tensor 847 data\ngguf_ex_read_1: tensor[847]: n_dims = 2, ne = (17408, 5120, 1, 1), name = blk.63.ffn_down.weight, data = 0x73d35fdfb9b0\nblk.63.ffn_down.weight data[:10] : 1.579176 0.000000 907486167040.000000 0.000228 0.000000 174.888931 -30085307675661518343635894462840832.000000 0.000000 -3220002832384.000000 -0.000000 \n\ngguf_ex_read_1: reading tensor 848 data\ngguf_ex_read_1: tensor[848]: n_dims = 2, ne = (5120, 17408, 1, 1), name = blk.63.ffn_gate.weight, data = 0x73d3610939b0\nblk.63.ffn_gate.weight data[:10] : -139671229077918688194068480.000000 0.017403 34851.042969 -1222996110550339282620383232.000000 -0.000000 -57473286144.000000 -0.000000 0.000277 -0.000000 -192248972461473792.000000 \n\ngguf_ex_read_1: reading tensor 849 data\ngguf_ex_read_1: tensor[849]: n_dims = 2, ne = (5120, 17408, 1, 1), name = blk.63.ffn_up.weight, data = 0x73d36232b9b0\nblk.63.ffn_up.weight data[:10] : 0.000000 -178719968003268542464.000000 0.000000 3308479178082405362696192.000000 0.000000 -0.000069 -4658421132768507204730880.000000 -37492529948254162386944.000000 8719009579008.000000 -0.000000 \n\ngguf_ex_read_1: reading tensor 850 data\ngguf_ex_read_1: tensor[850]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.63.post_attention_norm.weight, data = 0x73d3635c39b0\nblk.63.post_attention_norm.weight data[:10] : 1.245117 1.146484 1.209961 1.213867 1.102051 1.255859 1.275391 1.232422 1.215820 1.231445 \n\ngguf_ex_read_1: reading tensor 851 data\ngguf_ex_read_1: tensor[851]: n_dims = 2, ne = (5120, 1024, 1, 1), name = blk.64.attn_k.weight, data = 0x73d3635c89b0\nblk.64.attn_k.weight data[:10] : 0.000000 -0.000000 -0.000000 77783911728075348258548875264.000000 -0.000000 -779807621120.000000 -116419366345506816.000000 -0.000000 -3565253426151424.000000 -0.341793 \n\ngguf_ex_read_1: reading tensor 852 data\ngguf_ex_read_1: tensor[852]: n_dims = 1, ne = (256, 1, 1, 1), name = blk.64.attn_k_norm.weight, data = 0x73d3638989b0\nblk.64.attn_k_norm.weight data[:10] : 1.308594 0.885742 1.699219 1.263672 1.605469 1.941406 1.550781 1.714844 1.310547 1.710938 \n\ngguf_ex_read_1: reading tensor 853 data\ngguf_ex_read_1: tensor[853]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.attn_norm.weight, data = 0x73d363898db0\nblk.64.attn_norm.weight data[:10] : 1.011536 1.086914 1.051514 1.046631 1.306641 1.049316 1.024048 1.145508 1.234375 1.024536 \n\ngguf_ex_read_1: reading tensor 854 data\ngguf_ex_read_1: tensor[854]: n_dims = 2, ne = (6144, 5120, 1, 1), name = blk.64.attn_output.weight, data = 0x73d36389ddb0\nblk.64.attn_output.weight data[:10] : 0.000000 -nan -0.000231 -588164655477598002671369217638400.000000 0.000000 -0.000000 -70893914414901329129066281568305152.000000 -0.000000 -0.000000 427431294489308781141138740346880.000000 \n\ngguf_ex_read_1: reading tensor 855 data\ngguf_ex_read_1: tensor[855]: n_dims = 2, ne = (5120, 12288, 1, 1), name = blk.64.attn_q.weight, data = 0x73d36497ddb0\nblk.64.attn_q.weight data[:10] : 0.000000 -0.000479 544778361270436276207616.000000 -42490689775796476996741959936580255744.000000 -0.000000 0.000209 0.000000 -1730.768677 -0.000000 -7098020793417728.000000 \n\ngguf_ex_read_1: reading tensor 856 data\ngguf_ex_read_1: tensor[856]: n_dims = 1, ne = (256, 1, 1, 1), name = blk.64.attn_q_norm.weight, data = 0x73d366b3ddb0\nblk.64.attn_q_norm.weight data[:10] : 1.244141 1.072754 1.816406 1.253906 1.605469 1.941406 1.494141 1.863281 1.298828 1.550781 \n\ngguf_ex_read_1: reading tensor 857 data\ngguf_ex_read_1: tensor[857]: n_dims = 2, ne = (5120, 1024, 1, 1), name = blk.64.attn_v.weight, data = 0x73d366b3e1b0\nblk.64.attn_v.weight data[:10] : 0.000000 -0.000000 -0.000000 -0.000000 5447.940430 -14293357591483383808.000000 0.000000 5937133812701839162740289447133184.000000 -0.000000 75039989832789001467068416.000000 \n\ngguf_ex_read_1: reading tensor 858 data\ngguf_ex_read_1: tensor[858]: n_dims = 2, ne = (17408, 5120, 1, 1), name = blk.64.ffn_down.weight, data = 0x73d366e0e1b0\nblk.64.ffn_down.weight data[:10] : 0.000000 -10061968781293743379290446102528.000000 -0.000000 -0.000000 0.000000 225828.859375 0.000000 -1020286891632021366355436645646336.000000 0.000000 51912808.000000 \n\ngguf_ex_read_1: reading tensor 859 data\ngguf_ex_read_1: tensor[859]: n_dims = 2, ne = (5120, 17408, 1, 1), name = blk.64.ffn_gate.weight, data = 0x73d369dde1b0\nblk.64.ffn_gate.weight data[:10] : 0.000000 -39567656087807327807204908597248.000000 -8563729381619400704.000000 -100886693044704566552415174727269613568.000000 1195184917292318819663597264699392.000000 -0.000016 -207488335099243366305018937344.000000 67749950533943957651456.000000 -0.000000 457139.312500 \n\ngguf_ex_read_1: reading tensor 860 data\ngguf_ex_read_1: tensor[860]: n_dims = 2, ne = (5120, 17408, 1, 1), name = blk.64.ffn_up.weight, data = 0x73d36cdae1b0\nblk.64.ffn_up.weight data[:10] : 0.000000 -0.000000 2250904178786304.000000 -13267505927444837330583552.000000 -0.000000 -0.000240 -0.000000 270797587349504.000000 18843456574607247956150831967895552.000000 944346309187725626185875456.000000 \n\ngguf_ex_read_1: reading tensor 861 data\ngguf_ex_read_1: tensor[861]: n_dims = 2, ne = (10240, 5120, 1, 1), name = blk.64.nextn.eh_proj.weight, data = 0x73d36fd7e1b0\nblk.64.nextn.eh_proj.weight data[:10] : 0.000000 -0.000000 -144442774580488843072372736.000000 8119155712.000000 76078228243781219482940852409089392640.000000 17060836274248755462719215239168.000000 -0.000000 0.197302 -0.000000 256557.281250 \n\ngguf_ex_read_1: reading tensor 862 data\ngguf_ex_read_1: tensor[862]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.nextn.enorm.weight, data = 0x73d37199e1b0\nblk.64.nextn.enorm.weight data[:10] : 0.410156 0.511719 0.728516 0.554688 0.621094 0.697266 0.390625 0.658203 0.509766 0.691406 \n\ngguf_ex_read_1: reading tensor 863 data\ngguf_ex_read_1: tensor[863]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.nextn.hnorm.weight, data = 0x73d3719a31b0\nblk.64.nextn.hnorm.weight data[:10] : 0.762695 0.875000 0.784180 0.835938 0.935059 0.822266 0.773438 0.822266 0.830078 0.865234 \n\ngguf_ex_read_1: reading tensor 864 data\ngguf_ex_read_1: tensor[864]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.nextn.shared_head_norm.weight, data = 0x73d3719a81b0\nblk.64.nextn.shared_head_norm.weight data[:10] : 2.312500 1.972656 2.117188 2.093750 1.593750 2.257812 2.296875 2.085938 2.203125 2.015625 \n\ngguf_ex_read_1: reading tensor 865 data\ngguf_ex_read_1: tensor[865]: n_dims = 1, ne = (5120, 1, 1, 1), name = blk.64.post_attention_norm.weight, data = 0x73d3719ad1b0\nblk.64.post_attention_norm.weight data[:10] : 1.273438 1.210938 1.263672 1.249023 1.425781 1.261719 1.289062 1.310547 1.371094 1.204102 \n\ngguf_ex_read_1: ctx_data size: 7859392080\n"
+    },
+    "magic": {
+      "ok": true,
+      "value": "GGUF"
+    },
+    "mmproj_hash": {
+      "ok": true,
+      "rc": 0,
+      "tail": "sha256    c03a8b045fd65613ca56acc139c6112eef41e74f7d7290e9b29657c087098230  /mnt/geth-vol1/qwen38_assets/mmproj-Qwen3.8-27B-BF16.gguf\n"
+    },
+    "mmproj_structure": {
+      "ok": true,
+      "rc": 0,
+      "tail": "sor 309 data\ngguf_ex_read_1: tensor[309]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.8.ln1.weight, data = 0x7877fde59430\nv.blk.8.ln1.weight data[:10] : 1.000000 0.796875 0.843750 0.648438 1.007812 0.847656 0.957031 0.839844 1.171875 0.941406 \n\ngguf_ex_read_1: reading tensor 310 data\ngguf_ex_read_1: tensor[310]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.8.ln2.bias, data = 0x7877fde5a630\nv.blk.8.ln2.bias data[:10] : 0.210938 -0.180664 -0.176758 0.008301 0.048584 0.043945 0.041992 0.069336 -0.021484 0.056396 \n\ngguf_ex_read_1: reading tensor 311 data\ngguf_ex_read_1: tensor[311]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.8.ln2.weight, data = 0x7877fde5b830\nv.blk.8.ln2.weight data[:10] : 1.132812 1.078125 0.949219 0.832031 1.101562 1.078125 1.062500 1.179688 1.304688 1.062500 \n\ngguf_ex_read_1: reading tensor 312 data\ngguf_ex_read_1: tensor[312]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.attn_out.bias, data = 0x7877fde5ca30\nv.blk.9.attn_out.bias data[:10] : 0.028198 -0.052002 0.011719 0.073730 0.086426 0.068359 -0.188477 0.204102 -0.037842 0.077148 \n\ngguf_ex_read_1: reading tensor 313 data\ngguf_ex_read_1: tensor[313]: n_dims = 2, ne = (1152, 1152, 1, 1), name = v.blk.9.attn_out.weight, data = 0x7877fde5dc30\nv.blk.9.attn_out.weight data[:10] : 0.001108 0.012679 0.003055 0.004982 -0.010421 -0.014663 0.019498 0.000313 0.018583 0.010787 \n\ngguf_ex_read_1: reading tensor 314 data\ngguf_ex_read_1: tensor[314]: n_dims = 1, ne = (3456, 1, 1, 1), name = v.blk.9.attn_qkv.bias, data = 0x7877fe0e5c30\nv.blk.9.attn_qkv.bias data[:10] : 0.098145 -0.263672 -0.648438 -0.132812 -0.222656 -0.182617 -5.500000 -0.042969 -0.056885 0.024170 \n\ngguf_ex_read_1: reading tensor 315 data\ngguf_ex_read_1: tensor[315]: n_dims = 2, ne = (1152, 3456, 1, 1), name = v.blk.9.attn_qkv.weight, data = 0x7877fe0e9230\nv.blk.9.attn_qkv.weight data[:10] : 0.029020 -0.014846 -0.002666 0.000901 0.001493 0.042171 -0.016447 0.013137 -0.018095 -0.021391 \n\ngguf_ex_read_1: reading tensor 316 data\ngguf_ex_read_1: tensor[316]: n_dims = 1, ne = (4304, 1, 1, 1), name = v.blk.9.ffn_up.bias, data = 0x7877fe881230\nv.blk.9.ffn_up.bias data[:10] : -1.101562 -1.773438 -1.843750 -2.281250 -2.421875 -0.218750 -1.265625 -1.953125 -0.832031 -1.523438 \n\ngguf_ex_read_1: reading tensor 317 data\ngguf_ex_read_1: tensor[317]: n_dims = 2, ne = (1152, 4304, 1, 1), name = v.blk.9.ffn_up.weight, data = 0x7877fe885570\nv.blk.9.ffn_up.weight data[:10] : -0.002498 -0.005683 -0.001531 0.002743 0.003376 -0.011672 0.004371 0.027128 -0.011061 0.023100 \n\ngguf_ex_read_1: reading tensor 318 data\ngguf_ex_read_1: tensor[318]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ffn_down.bias, data = 0x7877ff1fa570\nv.blk.9.ffn_down.bias data[:10] : -0.194336 -0.386719 -0.068359 -0.149414 0.188477 -0.074707 0.335938 0.082520 0.063965 -0.017578 \n\ngguf_ex_read_1: reading tensor 319 data\ngguf_ex_read_1: tensor[319]: n_dims = 2, ne = (4304, 1152, 1, 1), name = v.blk.9.ffn_down.weight, data = 0x7877ff1fb770\nv.blk.9.ffn_down.weight data[:10] : 0.005439 0.005607 -0.010055 -0.009353 0.004051 0.005958 0.008925 0.006217 0.002064 0.001108 \n\ngguf_ex_read_1: reading tensor 320 data\ngguf_ex_read_1: tensor[320]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ln1.bias, data = 0x7877ffb70770\nv.blk.9.ln1.bias data[:10] : -0.038086 -0.055420 -0.062500 -0.006104 0.062988 0.094727 0.161133 -0.009949 -0.194336 -0.058594 \n\ngguf_ex_read_1: reading tensor 321 data\ngguf_ex_read_1: tensor[321]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ln1.weight, data = 0x7877ffb71970\nv.blk.9.ln1.weight data[:10] : 1.078125 0.968750 0.890625 0.726562 1.164062 0.972656 1.093750 1.007812 1.421875 1.039062 \n\ngguf_ex_read_1: reading tensor 322 data\ngguf_ex_read_1: tensor[322]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ln2.bias, data = 0x7877ffb72b70\nv.blk.9.ln2.bias data[:10] : -0.223633 -0.011475 -0.066406 -0.148438 -0.049316 0.095215 0.494141 -0.287109 -0.175781 -0.194336 \n\ngguf_ex_read_1: reading tensor 323 data\ngguf_ex_read_1: tensor[323]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.blk.9.ln2.weight, data = 0x7877ffb73d70\nv.blk.9.ln2.weight data[:10] : 1.171875 1.195312 1.000000 0.914062 1.203125 1.234375 1.039062 1.117188 1.421875 1.203125 \n\ngguf_ex_read_1: reading tensor 324 data\ngguf_ex_read_1: tensor[324]: n_dims = 1, ne = (4608, 1, 1, 1), name = mm.0.bias, data = 0x7877ffb74f70\nmm.0.bias data[:10] : 0.024536 0.007996 0.020142 0.022705 0.024170 0.017456 0.020386 0.025879 0.035156 0.018311 \n\ngguf_ex_read_1: reading tensor 325 data\ngguf_ex_read_1: tensor[325]: n_dims = 2, ne = (4608, 4608, 1, 1), name = mm.0.weight, data = 0x7877ffb79770\nmm.0.weight data[:10] : -0.020231 -0.012984 -0.011489 0.003628 -0.013350 -0.005119 0.003124 0.018949 -0.001943 -0.014998 \n\ngguf_ex_read_1: reading tensor 326 data\ngguf_ex_read_1: tensor[326]: n_dims = 1, ne = (5120, 1, 1, 1), name = mm.2.bias, data = 0x7878023f9770\nmm.2.bias data[:10] : 0.009338 0.000195 -0.007538 0.005615 -0.010071 -0.008667 0.028687 -0.018799 0.015076 0.000706 \n\ngguf_ex_read_1: reading tensor 327 data\ngguf_ex_read_1: tensor[327]: n_dims = 2, ne = (4608, 5120, 1, 1), name = mm.2.weight, data = 0x7878023fe770\nmm.2.weight data[:10] : 0.006339 0.027616 -0.002262 0.002880 -0.004585 -0.004310 0.003974 0.006446 0.011275 -0.022978 \n\ngguf_ex_read_1: reading tensor 328 data\ngguf_ex_read_1: tensor[328]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.post_ln.bias, data = 0x7878050fe770\nv.post_ln.bias data[:10] : 0.010437 0.021973 -0.063477 0.167969 -0.085449 -0.233398 -0.154297 -0.080078 -0.121582 -0.065430 \n\ngguf_ex_read_1: reading tensor 329 data\ngguf_ex_read_1: tensor[329]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.post_ln.weight, data = 0x7878050ff970\nv.post_ln.weight data[:10] : 1.335938 1.273438 1.257812 1.343750 1.437500 1.210938 1.320312 1.351562 1.203125 1.390625 \n\ngguf_ex_read_1: reading tensor 330 data\ngguf_ex_read_1: tensor[330]: n_dims = 1, ne = (1152, 1, 1, 1), name = v.patch_embd.bias, data = 0x787805100b70\nv.patch_embd.bias data[:10] : 0.107910 -0.341797 0.000954 -0.163086 0.081055 -0.076172 0.043213 -0.318359 0.045166 -0.392578 \n\ngguf_ex_read_1: reading tensor 331 data\ngguf_ex_read_1: tensor[331]: n_dims = 4, ne = (16, 16, 3, 1152), name = v.patch_embd.weight, data = 0x787805101d70\nv.patch_embd.weight data[:10] : 0.008179 -0.007172 -0.004852 -0.011902 -0.004822 0.010376 -0.008911 0.000607 -0.001587 0.024658 \n\ngguf_ex_read_1: reading tensor 332 data\ngguf_ex_read_1: tensor[332]: n_dims = 4, ne = (16, 16, 3, 1152), name = v.patch_embd.weight.1, data = 0x787805461d70\nv.patch_embd.weight.1 data[:10] : 0.009277 -0.006104 -0.003723 -0.010681 -0.003418 0.012024 -0.007751 0.001305 -0.000683 0.025879 \n\ngguf_ex_read_1: reading tensor 333 data\ngguf_ex_read_1: tensor[333]: n_dims = 2, ne = (1152, 2304, 1, 1), name = v.position_embd.weight, data = 0x7878057c1d70\nv.position_embd.weight data[:10] : -0.010620 -0.004089 -0.074707 0.092285 -0.141602 0.002914 -0.005249 0.100586 -0.010498 0.298828 \n\ngguf_ex_read_1: ctx_data size: 931249488\n"
+    },
+    "projector_identity": {
+      "actual_sha256": "83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53",
+      "actual_size_bytes": 931146432,
+      "expected_sha256": "83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53",
+      "expected_size_bytes": 931146432,
+      "ok": true
+    },
+    "text_one_token": {
+      "ok": true,
+      "rc": 0,
+      "soft": true,
+      "tail": "warning: no usable GPU found, --gpu-layers option will be ignored\nwarning: one possible reason is that llama.cpp was compiled without GPU support\nwarning: consult docs/build.md for compilation instructions\n--no-conversation is not supported by llama-cli\nplease use llama-completion instead\n\nLoading model... \n\n\n\u2584\u2584 \u2584\u2584\n\u2588\u2588 \u2588\u2588\n\u2588\u2588 \u2588\u2588  \u2580\u2580\u2588\u2584 \u2588\u2588\u2588\u2584\u2588\u2588\u2588\u2584  \u2580\u2580\u2588\u2584    \u2584\u2588\u2588\u2588\u2588 \u2588\u2588\u2588\u2588\u2584 \u2588\u2588\u2588\u2588\u2584\n\u2588\u2588 \u2588\u2588 \u2584\u2588\u2580\u2588\u2588 \u2588\u2588 \u2588\u2588 \u2588\u2588 \u2584\u2588\u2580\u2588\u2588    \u2588\u2588    \u2588\u2588 \u2588\u2588 \u2588\u2588 \u2588\u2588\n\u2588\u2588 \u2588\u2588 \u2580\u2588\u2584\u2588\u2588 \u2588\u2588 \u2588\u2588 \u2588\u2588 \u2580\u2588\u2584\u2588\u2588 \u2588\u2588 \u2580\u2588\u2588\u2588\u2588 \u2588\u2588\u2588\u2588\u2580 \u2588\u2588\u2588\u2588\u2580\n                                    \u2588\u2588    \u2588\u2588\n                                    \u2580\u2580    \u2580\u2580\n\nbuild      : b9591-62061f910\nmodel      : Qwen3.8-27B-IQ1_M.gguf\nmodalities : text\n\navailable commands:\n  /exit or Ctrl+C     stop or exit\n  /regen              regenerate the last response\n  /clear              clear the chat history\n  /read <file>        add a text file\n  /glob <pattern>     add text files using globbing pattern\n\n\n> 1\n\n[Start thinking]\nThe\n\nExiting...\n",
+      "timeout_only_soft": true
+    },
+    "vision_one_token": {
+      "ok": true,
+      "rc": 0,
+      "soft": true,
+      "tail": "warning: no usable GPU found, --gpu-layers option will be ignored\nwarning: one possible reason is that llama.cpp was compiled without GPU support\nwarning: consult docs/build.md for compilation instructions\n",
+      "timeout_only_soft": true
+    }
+  },
+  "label": "IQ1_M-direct-BF16-MTP-Q4_K",
+  "mmproj": {
+    "path": "/mnt/geth-vol1/qwen38_assets/mmproj-Qwen3.8-27B-BF16.gguf",
+    "sha256": "83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53",
+    "size_bytes": 931146432
+  },
+  "ok": true,
+  "path": "/mnt/geth-vol1/qwen38_out/Qwen3.8-27B-IQ1_M.gguf",
+  "sha256": "131cdf5c1c4b547081543382b00434e9ebf3f8eb369ef3714550086074f80bdf",
+  "size_bytes": 7870069760,
+  "ts": 1787011283.288072,
+  "warnings": []
+}

README.md

@@ -0,0 +1,65 @@
+---
+base_model: Qwen/Qwen3.8-27B
+license: apache-2.0
+library_name: gguf
+pipeline_tag: image-text-to-text
+tags:
+- qwen3.8
+- qwen35
+- gguf
+- llama.cpp
+- multimodal
+- iq1_m
+- q4_k
+---
+
+# Qwen3.8-27B IQ1_M GGUF
+
+A runnable multimodal GGUF quantization of `Qwen/Qwen3.8-27B`, produced directly from the pinned two-part BF16 GGUF source at revision `f1bfb127c64f7072bdd2cad55f258b9c8b2910fe`.
+
+## Quantization policy
+
+- The 64 calibrated main transformer blocks use imatrix-aware **IQ1_M**.
+- The auxiliary MTP block (`blk.64`) has no entries in the pinned importance matrix, so exactly eight two-dimensional weight tensors are **manually overridden to Q4_K** instead of being forced into an uncalibrated extreme-low-bit type. Other supporting tensors may independently use Q4_K or their normal GGUF types under the IQ1_M policy.
+- Those eight matrices contain 424,673,280 parameters, 1.5544% of the 27,320,697,856-parameter model.
+- Norm and other non-quantized tensors retain their normal GGUF types.
+- The resulting complete model is 2.3045 whole-file bits per weight. This is different from the nominal IQ1_M tensor rate because metadata and the protected Q4_K/F32 tensors are included.
+
+The eight manual block-64 Q4_K override tensors are:
+
+- `blk.64.attn_k.weight`
+- `blk.64.attn_output.weight`
+- `blk.64.attn_q.weight`
+- `blk.64.attn_v.weight`
+- `blk.64.ffn_down.weight`
+- `blk.64.ffn_gate.weight`
+- `blk.64.ffn_up.weight`
+- `blk.64.nextn.eh_proj.weight`
+
+## Toolchain
+
+The quantizer is built from llama.cpp `62061f91088281e65071cc38c5f69ee95c39f14e` plus the official MTP accounting fix from PR #24986, merge commit `b3ce5cedf4c007b78a45befe839fa3abada03c0b`. The executable SHA256 is `5d3a8456974b28569322dea7ee33941e3c9f09750e100dfce358326110b29187` and the applied patch SHA256 is `01a53c23afd4ed81b79ffc697b1c4a4b83443253b6095732eee3f8fc58e96b40`. Every loaded llama/ggml shared library is pinned in `QUANTIZER_TOOLCHAIN.json`.
+
+## Files
+
+- `Qwen3.8-27B-IQ1_M.gguf`: model GGUF, 7.33 GiB, SHA256 `131cdf5c1c4b547081543382b00434e9ebf3f8eb369ef3714550086074f80bdf`
+- `mmproj-Qwen3.8-27B-BF16.gguf`: vision projector, 888.01 MiB, SHA256 `83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53`
+- `Qwen3.8-27B-IQ1_M.gguf.validation.json`: structural, text-runtime and vision-runtime acceptance results
+- `GGUF_REPORT.json`: tensor-type histogram, architecture and whole-file BPW
+- `MTP_Q4K_AUDIT.json`: exact dry-run audit of the eight protected MTP tensors
+- `QUANTIZER_TOOLCHAIN.json`: executable, library, patch, source and argv hashes
+- `PROVENANCE.md`: source revisions, commands and verification details
+- `CHECKSUMS.sha256`: hashes for every release file
+- `PROVENANCE.json`: machine-readable model/projector identity contract consumed by the uploader
+
+## llama.cpp example
+
+```bash
+llama-mtmd-cli \
+  -m Qwen3.8-27B-IQ1_M.gguf \
+  --mmproj mmproj-Qwen3.8-27B-BF16.gguf \
+  --image image.jpg \
+  -p "Describe this image."
+```
+
+Extreme low-bit quantization trades quality for a much smaller artifact. Use a higher-bit quant when accuracy matters more than footprint.

SOURCE_DOWNLOAD_MANIFEST.json

@@ -0,0 +1,14 @@
+{
+  "repo": "unsloth/Qwen3.8-27B-GGUF",
+  "revision": "f1bfb127c64f7072bdd2cad55f258b9c8b2910fe",
+  "files": {
+    "BF16/Qwen3.8-27B-BF16-00001-of-00002.gguf": {
+      "size": 49986159616,
+      "sha256": "b9966e82b7a4d87028b5eae061d578ee826305ebf8baea5bfc6e09bad0ba191f"
+    },
+    "BF16/Qwen3.8-27B-BF16-00002-of-00002.gguf": {
+      "size": 4671576000,
+      "sha256": "92e3943c4f9bd6292a7bef82369f65fed9bfed088b9df0fb2fa2ce17c9edfa02"
+    }
+  }
+}

TENSOR_TYPE_OVERRIDES.txt

@@ -0,0 +1 @@
+blk[.]64[.].*=q4_k

mmproj-Qwen3.8-27B-BF16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83ee4f4f205fa514161778c41df1ea14144faa0f713510893b63c2395f5c2d53
+size 931146432