WaveCut commited on
Commit
cfeedc5
·
verified ·
1 Parent(s): 8115025

Repair OrbitQuant artifact metadata

Browse files
README.md CHANGED
@@ -16,6 +16,8 @@ This is an OrbitQuant artifact generated from the source model listed above.
16
  - Bits: `W2A3`
17
  - Runtime mode: `dequant_bf16`
18
  - Activation kernel backend: `auto`
 
 
19
  - Target policy: `z_image`
20
  - Rotation: `rpbh`
21
  - Rotation seed: `0`
 
16
  - Bits: `W2A3`
17
  - Runtime mode: `dequant_bf16`
18
  - Activation kernel backend: `auto`
19
+ - Quantization device: `cuda`
20
+ - Weight quantization backend: `triton_cuda`
21
  - Target policy: `z_image`
22
  - Rotation: `rpbh`
23
  - Rotation seed: `0`
SHA256SUMS CHANGED
@@ -34,7 +34,7 @@ e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 .cache/hugging
34
  dd4a8142f4d9d6adde53062c1b3839d57851cbd280f035df06ba98319e6d9d88 .cache/huggingface/download/quantization_config.json.metadata
35
  29e503a9b27827c2388071dd72725bd656c057a62c6292d4e82daffb89472443 .cache/huggingface/trees/36c530692b9a57b77f042d2ebe4ab0193393f3ea.json
36
  11ad7efa24975ee4b0c3c3a38ed18737f0658a5f75a0a96787b576a78a023361 .gitattributes
37
- 0a1c959c5eead1157ea45d9b93698c7025f399a7075478621d95437471a22a94 README.md
38
  e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 assets/.gitkeep
39
  427ec5087e732f9ef392535df5dc1c8ccd8efeb0b2e7e958b6bd291f86aa4aa7 assets/original_vs_orbitquant_z-image-native_seed0_W2A3_simple-object.webp
40
  5618d77395474d9a60a7b628f43ccfe892e51cbd55931fd311256622f164473f assets/z-image-native_seed0_W2A3_simple-object.png
@@ -45,11 +45,11 @@ a13d7c5e403dadf518f22586bef4ccaab1713cafe2904f638f88ed79d2dcbd4e assets/z-image
45
  8699ac9eb6d973a4900f8678046c3aa53e96e800d39d16d6058652c334f02175 benchmark/orbitquant.metrics.jsonl
46
  af4941b2147bc8c44345cb807ce27e433e8e174fca4700ee04ad313744a10264 benchmark/original.metrics.csv
47
  bdc756d15f8aeddfa1440becea374f09e31eefb46a26743295a98db23279cb35 benchmark/original.metrics.jsonl
48
- 45213c814886feb6c14f7e8fded16dfc2b6c03f53ebff68a9e3f9ea530b8b085 benchmark/summary.json
49
  0a635f376c86c2dd6722a46dc267244cf8bc3353c1202d1a2816d62dd82fee79 model.safetensors
50
- 25f3914c0e52ca2053ec4209b93fc02683901a5c65b19eabd4e06f4e4ab3c8b5 model_index.json
51
  04da03ff11a91f322d72c88fe5686e7ebf66972de6ed5ab606e5282e682a332f orbitquant_codebooks.safetensors
52
- 22f193e778f1beb54ae689137c82630bf9123e214b3fa2f75c17bf313ca94b54 orbitquant_manifest.json
53
  e1724c9394bf83ea05da6b457b8072d2bcb3d5a533dafce916da8deb517d3825 orbitquant_rotations.safetensors
54
  65ec8afff39a6cf80feee2eac2f06fa66e51cd471a8382f4355a04fd74eda0af prompts.json
55
  027e947dcac727df2af2d04916df320c6933566400e959e711a3a32c2d31411c quantization_config.json
 
34
  dd4a8142f4d9d6adde53062c1b3839d57851cbd280f035df06ba98319e6d9d88 .cache/huggingface/download/quantization_config.json.metadata
35
  29e503a9b27827c2388071dd72725bd656c057a62c6292d4e82daffb89472443 .cache/huggingface/trees/36c530692b9a57b77f042d2ebe4ab0193393f3ea.json
36
  11ad7efa24975ee4b0c3c3a38ed18737f0658a5f75a0a96787b576a78a023361 .gitattributes
37
+ b3240bd02706cc1ba3d39d618cea03b68e6cafcf9f865645d02c72c2d503e907 README.md
38
  e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 assets/.gitkeep
39
  427ec5087e732f9ef392535df5dc1c8ccd8efeb0b2e7e958b6bd291f86aa4aa7 assets/original_vs_orbitquant_z-image-native_seed0_W2A3_simple-object.webp
40
  5618d77395474d9a60a7b628f43ccfe892e51cbd55931fd311256622f164473f assets/z-image-native_seed0_W2A3_simple-object.png
 
45
  8699ac9eb6d973a4900f8678046c3aa53e96e800d39d16d6058652c334f02175 benchmark/orbitquant.metrics.jsonl
46
  af4941b2147bc8c44345cb807ce27e433e8e174fca4700ee04ad313744a10264 benchmark/original.metrics.csv
47
  bdc756d15f8aeddfa1440becea374f09e31eefb46a26743295a98db23279cb35 benchmark/original.metrics.jsonl
48
+ db4a5d11492c9c5026c7b8a9a517c208f57f41aa7657203eb90cb5d4b0bc4a67 benchmark/summary.json
49
  0a635f376c86c2dd6722a46dc267244cf8bc3353c1202d1a2816d62dd82fee79 model.safetensors
50
+ 54542cf15ffc82072e9e90bdf688f254d809eceea1a5a181e520848910f0ef5c model_index.json
51
  04da03ff11a91f322d72c88fe5686e7ebf66972de6ed5ab606e5282e682a332f orbitquant_codebooks.safetensors
52
+ af8cd292a2dde7fe855225958c90e313beac12fcf38d5943f6f3deed340a3318 orbitquant_manifest.json
53
  e1724c9394bf83ea05da6b457b8072d2bcb3d5a533dafce916da8deb517d3825 orbitquant_rotations.safetensors
54
  65ec8afff39a6cf80feee2eac2f06fa66e51cd471a8382f4355a04fd74eda0af prompts.json
55
  027e947dcac727df2af2d04916df320c6933566400e959e711a3a32c2d31411c quantization_config.json
benchmark/summary.json CHANGED
@@ -142,5 +142,7 @@
142
  }
143
  }
144
  }
145
- }
 
 
146
  }
 
142
  }
143
  }
144
  }
145
+ },
146
+ "quantization_device": "cuda",
147
+ "weight_quantization_backend": "triton_cuda"
148
  }
model_index.json CHANGED
@@ -15,5 +15,7 @@
15
  "activation_bits": 3,
16
  "target_policy": "z_image",
17
  "runtime_mode": "dequant_bf16",
18
- "activation_kernel_backend": "auto"
 
 
19
  }
 
15
  "activation_bits": 3,
16
  "target_policy": "z_image",
17
  "runtime_mode": "dequant_bf16",
18
+ "activation_kernel_backend": "auto",
19
+ "quantization_device": "cuda",
20
+ "weight_quantization_backend": "triton_cuda"
21
  }
orbitquant_manifest.json CHANGED
@@ -16,6 +16,8 @@
16
  "row_norm_dtype": "bfloat16",
17
  "runtime_mode": "dequant_bf16",
18
  "activation_kernel_backend": "auto",
 
 
19
  "target_policy": "z_image",
20
  "adaln_policy": "int4_rtn_group64_bf16_activation",
21
  "quantized_modules": [
@@ -2723,9 +2725,9 @@
2723
  "orbitquant_codebooks.safetensors": "04da03ff11a91f322d72c88fe5686e7ebf66972de6ed5ab606e5282e682a332f",
2724
  "orbitquant_rotations.safetensors": "e1724c9394bf83ea05da6b457b8072d2bcb3d5a533dafce916da8deb517d3825",
2725
  "quantization_config.json": "027e947dcac727df2af2d04916df320c6933566400e959e711a3a32c2d31411c",
2726
- "model_index.json": "25f3914c0e52ca2053ec4209b93fc02683901a5c65b19eabd4e06f4e4ab3c8b5",
2727
  "prompts.json": "65ec8afff39a6cf80feee2eac2f06fa66e51cd471a8382f4355a04fd74eda0af",
2728
- "benchmark/summary.json": "45213c814886feb6c14f7e8fded16dfc2b6c03f53ebff68a9e3f9ea530b8b085",
2729
  "benchmark/original.metrics.jsonl": "bdc756d15f8aeddfa1440becea374f09e31eefb46a26743295a98db23279cb35",
2730
  "benchmark/orbitquant.metrics.jsonl": "8699ac9eb6d973a4900f8678046c3aa53e96e800d39d16d6058652c334f02175",
2731
  "benchmark/original.metrics.csv": "af4941b2147bc8c44345cb807ce27e433e8e174fca4700ee04ad313744a10264",
 
16
  "row_norm_dtype": "bfloat16",
17
  "runtime_mode": "dequant_bf16",
18
  "activation_kernel_backend": "auto",
19
+ "quantization_device": "cuda",
20
+ "weight_quantization_backend": "triton_cuda",
21
  "target_policy": "z_image",
22
  "adaln_policy": "int4_rtn_group64_bf16_activation",
23
  "quantized_modules": [
 
2725
  "orbitquant_codebooks.safetensors": "04da03ff11a91f322d72c88fe5686e7ebf66972de6ed5ab606e5282e682a332f",
2726
  "orbitquant_rotations.safetensors": "e1724c9394bf83ea05da6b457b8072d2bcb3d5a533dafce916da8deb517d3825",
2727
  "quantization_config.json": "027e947dcac727df2af2d04916df320c6933566400e959e711a3a32c2d31411c",
2728
+ "model_index.json": "54542cf15ffc82072e9e90bdf688f254d809eceea1a5a181e520848910f0ef5c",
2729
  "prompts.json": "65ec8afff39a6cf80feee2eac2f06fa66e51cd471a8382f4355a04fd74eda0af",
2730
+ "benchmark/summary.json": "db4a5d11492c9c5026c7b8a9a517c208f57f41aa7657203eb90cb5d4b0bc4a67",
2731
  "benchmark/original.metrics.jsonl": "bdc756d15f8aeddfa1440becea374f09e31eefb46a26743295a98db23279cb35",
2732
  "benchmark/orbitquant.metrics.jsonl": "8699ac9eb6d973a4900f8678046c3aa53e96e800d39d16d6058652c334f02175",
2733
  "benchmark/original.metrics.csv": "af4941b2147bc8c44345cb807ce27e433e8e174fca4700ee04ad313744a10264",