PrismaQuant final: visual NVFP4 (108/110 DP-placed), lm_head BF16 (vLLM ParallelLMHead runtime limit), pos_embed excluded

config.json

@@ -382,6 +382,8 @@
           "re:^language_model[.]model[.]layers[.]9[.]mlp[.]experts[.]gate_up_proj$",
           "re:^mtp[.]layers[.]0[.]mlp[.]experts[.]down_proj$",
           "re:^mtp[.]layers[.]0[.]mlp[.]experts[.]gate_up_proj$",
+          "re:^visual[.]merger[.]linear_fc1$",
+          "re:^visual[.]merger[.]linear_fc2$",
           "re:^language_model[.]model[.]layers[.][0-9]+[.]mlp[.]experts[.][0-9]+[.](gate|up|down)_proj$",
           "re:^mtp[.]layers[.][0-9]+[.]mlp[.]experts[.][0-9]+[.](gate|up|down)_proj$"
         ]
@@ -1002,8 +1004,6 @@
       "visual.blocks.9.attn.qkv",
       "visual.blocks.9.mlp.linear_fc1",
       "visual.blocks.9.mlp.linear_fc2",
-      "visual.merger.linear_fc1",
-      "visual.merger.linear_fc2",
       "visual.pos_embed"
     ],
     "quantization_status": "compressed"

mixed_native_manifest.json

@@ -12,7 +12,7 @@
     "mtp_packed_moe_per_expert/NVFP4": 2,
     "mtp_passthrough/BF16": 9
   },
-  "n_assignment_entries": 590,
+  "n_assignment_entries": 589,
   "ignore": [
     "lm_head"
   ]

model-00001-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9c4fcff082526d1da5b7679802b5d3bf4dd111264c60cb5aa2c6cd2fc86f9a0e
-size 5114226512
+oid sha256:e992111654df517a274cb6d53e39bb938cbe0beec312c1e669132a732d4a25e3
+size 5110883464

model-00002-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a10e9c7164c651ab540d55d0762fd990a8edbf8858783af14438e5970135aaa
-size 5114009072
+oid sha256:c7d55d0029eeb180fba4fc5abf799494c2e70a723143a475603d078845566fe1
+size 5111844976

model-00003-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2f6c8c7222397fd84adf82aa855ac9f52e3eab27d3379be62cb36889f28552a0
-size 5113912496
+oid sha256:24e168889551c4652545d0d77b396d8a7a33ad98cf5500d7a685792d68f986af
+size 5110569432

model-00004-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f368238d4f6b103617baab434c8bf4497ee751d532b99c6f0cf24650511edf85
-size 5106839032
+oid sha256:509c635ee89750f29640f934c259a68ae9b2c84d510a2263e879f7e4b6e32225
+size 5094455024

model-00005-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a060f0705b7d59387fc2250307c8ec4d6d9217d29856961e65e514ebb53a6d1d
-size 5115003872
+oid sha256:4a4aad4d93fda81ef5eb452a194c39b147fc0f049ec17e7140ebb88436ce8bd7
+size 5111654544

model-00006-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b3fdecc210de7448eb05a3d19b0ead3fb40d115b698c8c9ef1b6ab9953ab1477
-size 5114352912
+oid sha256:6a9d1f2004eb9ea451b73b6aef7634abb6e07eee69a2b32acfb1dcef84f4991d
+size 5110615856

model-00007-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8fecaad8e04d4eaaf54daa1e5735ed37de2e625f0dccad00846ab575380d7c1b
-size 5114968200
+oid sha256:98a050de1b4d84b5407b211160577d1818271f9c8fab68a205f3db5de93bc422
+size 5100811840

model-00008-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:493e8e9a8a2ffdda2db8f54dbb266f7267949ceaee0f96f6a501eb06b5d47c86
-size 5114109968
+oid sha256:f620ee070d72ddab944cd6ae4a2813360a3507eb6c61ba183f84fe949a260099
+size 5110565856

model-00009-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:933bdb77f37c90a9177daabbca0c6f0ee406af1fe92c774f0ffd6ddd4997d664
-size 5114086280
+oid sha256:99f6bef674337eee03974f7b563e32482acb66b90396ac782f5662fae55d90b3
+size 5110546200

model-00010-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:695dde7cda4d948f1cb73ce12fc1f9c11c66667200f77c046aefd2b7628553fb
-size 5113661128
+oid sha256:5eeade358a67b520102ce0d0fead204c5ccff73550abccdcf0e455f490201742
+size 5111694072

model-00011-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bf95284b270316d34ade4c012c2a0cbdd3f5473a1cc54b4d911e03dbfd8b6085
-size 5114752408
+oid sha256:25046cb42aa61032d89ce46d5f1a07a87239f922e8f8fc062dd8474ec1b5b4ff
+size 5111409456

model-00012-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:27254df34465a5d06109103fad04f3e7554aed9ca85c8fa9280085ecfa968cec
-size 5115020720
+oid sha256:e9d468cf675eae7e60c67b71959956881a890747bf5ae2d002f18462d070fb6d
+size 5111480504

model-00013-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:35f52f075258f944d906a640d5179e962b9d10414f9c57123e302ee6758ab609
-size 5114947608
+oid sha256:368826a98ceb81b74740d3f00da850c568b867652f615b19c9d6da38c4f61840
+size 5111407656

model-00014-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cb7079e13fca9937723fcdf331a6d9c65c32fcf016e184fc82aa293720c88ace
-size 5113656704
+oid sha256:2ac37e33551c811ca7bad6e6c434c5be53cdfd48c1972f43f7b1837337fb3e9e
+size 5111689640

model-00015-of-00016.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ec8db6fd3477452bea9e784c16f9169b040aecbf84b8cd03fff7914008229872
-size 5029797224
+oid sha256:275fd6d6912dbd11597151564526bb662f5b97419b343fc627d3edf43af18219
+size 5042843944

model.safetensors.index.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:44e7d71867e482556c2c1f4bff9486585871d9666c26ab14fcee802802d282e4
-size 17414423
+oid sha256:452ce7936da6de8aadd1546324cd33eba2eba17c367295291c7168304954f78d
+size 17414985