Add validated Iris-3B weight-only INT8 models and CUDA comparisons

.gitattributes

@@ -33,3 +33,9 @@ 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
+examples/depth-photo.jpg filter=lfs diff=lfs merge=lfs -text
+examples/depth-watercolor.jpg filter=lfs diff=lfs merge=lfs -text
+examples/generate-photo.jpg filter=lfs diff=lfs merge=lfs -text
+examples/generate-watercolor.jpg filter=lfs diff=lfs merge=lfs -text
+examples/upscale-photo.jpg filter=lfs diff=lfs merge=lfs -text
+examples/upscale-watercolor.jpg filter=lfs diff=lfs merge=lfs -text

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 [yyyy] [name of copyright owner]
+
+   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.

NOTICE

@@ -0,0 +1,10 @@
+Iris-3B
+Copyright 2026 Speridlabs
+
+This product is licensed under the Apache License, Version 2.0 (see LICENSE).
+
+Runtime dependencies that are downloaded, not redistributed, keep their own
+licenses, including:
+  - Qwen3-VL-4B-Instruct (text encoder), Alibaba Cloud, Apache License 2.0
+  - DINOv2 (REPA teacher, training only), Meta Platforms, Apache License 2.0
+  - Dion (Muon optimizer implementation), Microsoft, MIT License

README.md

@@ -0,0 +1,83 @@
+---
+license: apache-2.0
+language:
+- en
+- ja
+library_name: pytorch
+pipeline_tag: text-to-image
+tags:
+- iris
+- forge-neo
+- int8
+- weight-only
+- image-to-image
+- depth-estimation
+- image-restoration
+base_model: speridlabs/iris-3b
+---
+
+# Iris-3B INT8 for Aikimi Forge Neo
+
+Rowwise symmetric, weight-only INT8 derivatives of Speridlabs' Iris-3B image generation, relative depth, and restoration models. Use the Iris tab in [Aikimi Forge Neo](https://github.com/AiWithYou/aikimi-forge-neo). The integration can also run the original, unquantized weights.
+
+This is a custom checkpoint format (`iris-rowwise-int8-v1`) for Forge Neo's Iris loader. It stores selected Linear weights as INT8 with per-output FP32 scales, then computes Linear operations in BF16 under CUDA autocast. It does not use an integer GEMM kernel. Shared modulation cores and other parameters retain their original format. The Qwen3-VL text encoder is unchanged and downloaded separately for image generation; depth and restoration do not load it.
+
+## Files and use
+
+| Directory | Task | INT8 model bytes | Original model bytes |
+|---|---|---:|---:|
+| Root | Text-to-image | 3,233,164,048 | 11,950,112,680 |
+| `depth/` | Relative depth | 3,233,825,264 | 11,952,738,184 |
+| `upscaler/` | Restoration and 4x enlargement | 3,233,164,048 | 11,950,112,680 |
+
+Each model includes its original configuration and a `manifest.json` recording the source revision, SHA-256, conversion format, quantized Linear count, and output file sizes and SHA-256. Downstream models also include their original empty-prompt embedding. Model weights alone are about 73% smaller. Text-to-image still needs the separate Qwen3-VL-4B-Instruct encoder (about 8.88 GB of weights).
+
+On Windows with an NVIDIA GPU, update Forge Neo and run:
+
+```powershell
+.\aikimi-iris-setup.bat
+```
+
+Or select Iris, a task, and INT8 in the UI, then choose **モデルを準備**. Only the selected task's weights are downloaded. To prepare all three:
+
+```powershell
+.\venv\Scripts\python.exe tools/setup_iris.py --precision int8 --task all
+```
+
+Forge Neo downloads a fixed, validated Hub commit and checks every weight file against size and SHA-256. Users do not need to quantize locally. The dedicated worker environment is separate from the main UI environment. Source inference code is pinned; only its packaging definition moves training dependencies into an optional extra.
+
+The format requires Forge Neo's quantization loader. Loading these files directly with the original upstream loader or a Diffusers pipeline is not supported.
+
+Text-to-image size choices follow the upstream demo's buckets with substantial training mass: 1024x1024 (default), 1344x768, 1280x832, 1152x896, 896x1152, 832x1280, and 768x1344. Additional INT8 UI tests at 256x256 produced collapsed grid-like output and at 512x512 visible grid artifacts, so these sizes are excluded from the integration. The paired comparison images below use 1024x1024.
+
+## Measured comparisons
+
+Tested on 2026-10-09 with one NVIDIA RTX 3090 24 GB, Windows, Python 3.13.14, PyTorch 2.13.0+cu130. Each pair uses identical input and conditions. Text-to-image uses 1024x1024, 100 steps, CFG 3, an empty negative prompt, and fixed seeds. Cold-load smoke tests and initial CPU model construction are excluded from timing; inference wall time includes returning model weights to CPU. Peaks are PyTorch CUDA allocated memory, not total board usage. Each condition was measured once.
+
+Text-to-image's measured peak decreased from approximately 18.9 GiB to 10.7 GiB. Generation times were similar, with INT8 slightly slower in these samples. Both normal and INT8 completed at 1024x1024 on this 24 GB card. No physical 12 GB or 16 GB card was tested.
+
+![Photo comparison](examples/generate-photo.jpg)
+
+![Watercolor comparison](examples/generate-watercolor.jpg)
+
+Both generated pairs preserved the scene and subject in visual inspection, with small changes in fine detail. They are not pixel-identical, and these examples do not establish lossless quantization or a general quality guarantee. Prompts, seeds, raw measurements, and paired differences are in [benchmark.json](benchmark.json).
+
+Depth is relative depth, not metric distance. The two 1024x1024 comparisons have finite depth arrays and normal/INT8 Pearson correlations of 0.999908 and 0.999945; this measures agreement between variants, not ground-truth depth accuracy. PNG visualization and NPY arrays are exported.
+
+![Relative depth comparison](examples/depth-photo.jpg)
+
+Restoration inputs were deliberately downsampled and JPEG-compressed generated references: 256x256 quality 35 to 1024x1024 (one tile), and 384x256 quality 45 to 1536x1024 (two overlapping tiles). Both precisions completed both paths, and floating-point output was checked for finiteness before integer image conversion. Visual inspection found similar structure and detail in these pairs, with small changes. Restoration may reconstruct detail differently from the reference.
+
+![Restoration comparison](examples/upscale-watercolor.jpg)
+
+All twelve paired CUDA cases completed. Depth/restoration peak allocated memory was approximately 15.0–15.1 GiB normal and 6.8–6.9 GiB INT8. These tasks had no pre-run warmup; the first photo includes cold GPU transfer and host memory/disk effects, so their times must not be compared as a steady-state speed benchmark. See the methodology in benchmark.json.
+
+Restoration follows the upstream input budget (short edge up to 512, long edge up to 1024) and enlarges that processed input 4x, using overlapping tiles where required.
+
+## Source and license
+
+- Original model: [speridlabs/iris-3b](https://huggingface.co/speridlabs/iris-3b), revision `7445443349bc9abe3c96f01ff793e2098ca012b3`.
+- Original inference: [speridlabs/iris-3b](https://github.com/speridlabs/iris-3b), commit `a8d15239dea469aba042cfa56ca3bb4e450d5ebc`.
+- Text encoder: [Qwen/Qwen3-VL-4B-Instruct](https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct), revision `ebb281ec70b05090aa6165b016eac8ec08e71b17`.
+
+Apache-2.0. Original LICENSE and NOTICE are retained. Aikimi's modification is the weight-only conversion and Forge Neo integration; this is an independently converted derivative, not an official Speridlabs INT8 release.

benchmark.json

@@ -0,0 +1,305 @@
+{
+  "date": "2026-10-09",
+  "environment": {
+    "torch": "2.13.0+cu130",
+    "python": "3.13.14 (main, Jun 11 2026, 04:04:46)",
+    "device": "NVIDIA GeForce RTX 3090",
+    "cuda": "13.0",
+    "torch_threads": 10
+  },
+  "method": "One GPU; one task model at a time; torch inference mode and BF16 autocast. T2I uses Qwen3-VL-4B and 100 steps, 1024x1024, CFG 3 after a 2-step warmup. Other tasks have no warmup; their first photo case includes cold GPU transfer and host memory/disk effects. Times are Runner.run wall time including GPU offload, excluding initial CPU model construction and downloads. Initial 2-step T2I cold-load smoke is excluded. Each condition was measured once, not a statistical speed benchmark. Peak is torch CUDA allocated/reserved memory for the inference. It excludes other processes and is not total board memory. Depth uses the 1024x1024 normal T2I PNG. Restoration inputs are generated references downsampled to 256x256 JPEG quality 35 (one tile) and cropped/downsampled to 384x256 JPEG quality 45 (two overlapping tiles), outputs 1024x1024 and 1536x1024.",
+  "records": [
+    {
+      "task": "generate",
+      "case": "photo",
+      "precision": "normal",
+      "request": {
+        "task": "generate",
+        "precision": "normal",
+        "prompt": "A candid editorial photograph of a young adult woman with short dark hair standing at a rain soaked Tokyo street crossing at dusk, holding a transparent umbrella, reflections of neon signs, realistic skin texture, 50mm lens, natural proportions, cinematic but believable lighting.",
+        "seed": 424242,
+        "width": 1024,
+        "height": 1024,
+        "steps": 100,
+        "cfg": 3.0,
+        "negative_prompt": ""
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 243.98675999999978,
+      "peak_allocated_mib": 19348.044921875,
+      "peak_reserved_mib": 19732.0,
+      "pixel_std": 72.87049865722656
+    },
+    {
+      "task": "generate",
+      "case": "photo",
+      "precision": "int8",
+      "request": {
+        "task": "generate",
+        "precision": "int8",
+        "prompt": "A candid editorial photograph of a young adult woman with short dark hair standing at a rain soaked Tokyo street crossing at dusk, holding a transparent umbrella, reflections of neon signs, realistic skin texture, 50mm lens, natural proportions, cinematic but believable lighting.",
+        "seed": 424242,
+        "width": 1024,
+        "height": 1024,
+        "steps": 100,
+        "cfg": 3.0,
+        "negative_prompt": ""
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 244.0833332000766,
+      "peak_allocated_mib": 10934.3515625,
+      "peak_reserved_mib": 11488.0,
+      "pixel_std": 73.21208953857422
+    },
+    {
+      "task": "generate",
+      "case": "watercolor",
+      "precision": "normal",
+      "request": {
+        "task": "generate",
+        "precision": "normal",
+        "prompt": "A detailed hand painted watercolor illustration of a small wooden reading room above a quiet bookshop. A ginger cat sleeps on a green velvet armchair, warm light from a brass desk lamp, a large window looking onto rainy rooftops, bookshelves with distinct books, delicate line work, paper texture. No text.",
+        "seed": 20261009,
+        "width": 1024,
+        "height": 1024,
+        "steps": 100,
+        "cfg": 3.0,
+        "negative_prompt": ""
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 252.1166963998694,
+      "peak_allocated_mib": 19347.509765625,
+      "peak_reserved_mib": 19752.0,
+      "pixel_std": 66.97412872314453
+    },
+    {
+      "task": "generate",
+      "case": "watercolor",
+      "precision": "int8",
+      "request": {
+        "task": "generate",
+        "precision": "int8",
+        "prompt": "A detailed hand painted watercolor illustration of a small wooden reading room above a quiet bookshop. A ginger cat sleeps on a green velvet armchair, warm light from a brass desk lamp, a large window looking onto rainy rooftops, bookshelves with distinct books, delicate line work, paper texture. No text.",
+        "seed": 20261009,
+        "width": 1024,
+        "height": 1024,
+        "steps": 100,
+        "cfg": 3.0,
+        "negative_prompt": ""
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 254.52246790006757,
+      "peak_allocated_mib": 10933.671875,
+      "peak_reserved_mib": 11528.0,
+      "pixel_std": 67.01991271972656
+    },
+    {
+      "task": "depth",
+      "case": "photo",
+      "precision": "normal",
+      "request": {
+        "task": "depth",
+        "precision": "normal"
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 283.81315309996717,
+      "peak_allocated_mib": 15374.9541015625,
+      "peak_reserved_mib": 16048.0,
+      "pixel_std": 85.49915313720703
+    },
+    {
+      "task": "depth",
+      "case": "photo",
+      "precision": "int8",
+      "request": {
+        "task": "depth",
+        "precision": "int8"
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 87.89453630009666,
+      "peak_allocated_mib": 6960.0576171875,
+      "peak_reserved_mib": 7674.0,
+      "pixel_std": 85.4561996459961
+    },
+    {
+      "task": "depth",
+      "case": "watercolor",
+      "precision": "normal",
+      "request": {
+        "task": "depth",
+        "precision": "normal"
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 8.675420299870893,
+      "peak_allocated_mib": 15406.9013671875,
+      "peak_reserved_mib": 16058.0,
+      "pixel_std": 80.67788696289062
+    },
+    {
+      "task": "depth",
+      "case": "watercolor",
+      "precision": "int8",
+      "request": {
+        "task": "depth",
+        "precision": "int8"
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 3.3528326998930424,
+      "peak_allocated_mib": 6989.7626953125,
+      "peak_reserved_mib": 7708.0,
+      "pixel_std": 80.69136047363281
+    },
+    {
+      "task": "upscale",
+      "case": "photo",
+      "precision": "normal",
+      "request": {
+        "task": "upscale",
+        "precision": "normal"
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 300.6608500001021,
+      "peak_allocated_mib": 15366.005859375,
+      "peak_reserved_mib": 16018.0,
+      "pixel_std": 74.8482437133789
+    },
+    {
+      "task": "upscale",
+      "case": "photo",
+      "precision": "int8",
+      "request": {
+        "task": "upscale",
+        "precision": "int8"
+      },
+      "size": [
+        1024,
+        1024
+      ],
+      "wall_seconds": 91.08279359992594,
+      "peak_allocated_mib": 6953.5810546875,
+      "peak_reserved_mib": 7654.0,
+      "pixel_std": 74.83624267578125
+    },
+    {
+      "task": "upscale",
+      "case": "watercolor",
+      "precision": "normal",
+      "request": {
+        "task": "upscale",
+        "precision": "normal"
+      },
+      "size": [
+        1536,
+        1024
+      ],
+      "wall_seconds": 9.50692670000717,
+      "peak_allocated_mib": 15452.650390625,
+      "peak_reserved_mib": 16100.0,
+      "pixel_std": 72.71382141113281
+    },
+    {
+      "task": "upscale",
+      "case": "watercolor",
+      "precision": "int8",
+      "request": {
+        "task": "upscale",
+        "precision": "int8"
+      },
+      "size": [
+        1536,
+        1024
+      ],
+      "wall_seconds": 4.555513399885967,
+      "peak_allocated_mib": 7039.1875,
+      "peak_reserved_mib": 7754.0,
+      "pixel_std": 72.731201171875
+    }
+  ],
+  "comparisons": [
+    {
+      "task": "generate",
+      "case": "photo",
+      "pixel_mae_0_255": 2.3095102310180664,
+      "identical_pixels_fraction": 0.023015975952148438
+    },
+    {
+      "task": "generate",
+      "case": "watercolor",
+      "pixel_mae_0_255": 4.086854457855225,
+      "identical_pixels_fraction": 0.0079193115234375
+    },
+    {
+      "task": "depth",
+      "case": "photo",
+      "pixel_mae_0_255": 0.5848824381828308,
+      "identical_pixels_fraction": 0.5262269973754883,
+      "depth_pearson": 0.9999082843818213,
+      "depth_shapes": [
+        [
+          1024,
+          1024
+        ],
+        [
+          1024,
+          1024
+        ]
+      ]
+    },
+    {
+      "task": "depth",
+      "case": "watercolor",
+      "pixel_mae_0_255": 0.6243413090705872,
+      "identical_pixels_fraction": 0.5460033416748047,
+      "depth_pearson": 0.9999447206409359,
+      "depth_shapes": [
+        [
+          1024,
+          1024
+        ],
+        [
+          1024,
+          1024
+        ]
+      ]
+    },
+    {
+      "task": "upscale",
+      "case": "photo",
+      "pixel_mae_0_255": 0.6896772384643555,
+      "identical_pixels_fraction": 0.09531784057617188
+    },
+    {
+      "task": "upscale",
+      "case": "watercolor",
+      "pixel_mae_0_255": 1.0330185890197754,
+      "identical_pixels_fraction": 0.07715797424316406
+    }
+  ]
+}

config.yaml

@@ -0,0 +1,77 @@
+model:
+  block: single_stream
+  dual_depth: 8
+  final_block_text: keep
+  hidden_size: 2560
+  depth: 24
+  num_heads: 20
+  num_kv_heads: 5
+  gated_attention: true
+  sandwich_norm: true
+  patch_size: 16
+  in_channels: 3
+  mlp_ratio: 4.0
+  qkv_bias: false
+  qk_norm: true
+  norm_eps: 1.0e-06
+  attn_backend: sdpa
+  modulation: shared_bias
+  modulation_rank: 64
+  timestep_max_period: 10.0
+  adaln_zero_init: true
+  rope_theta: 10000.0
+  rope_scale: 16.0
+  rope_aspect: isotropic
+  rope_frame_pairs: 0
+  rope_frame_theta: 10.0
+  text_rope: true
+  text_rope_theta: 10000.0
+  text_abs_pos_embed: true
+  text_dim: 2560
+  text_len: 300
+  text_adapter: lap_blocks2
+  text_lap_num_layers: 12
+  text_lap_num_heads: 32
+  text_lap_mlp_ratio: 1.3
+  repa_layer: 10
+  pixel:
+    enabled: true
+    depth: 4
+    hidden_size: 16
+    attn_hidden_size: 1280
+    num_heads: 10
+    mlp_ratio: 4.0
+    modulation: post
+    abs_pos_embed: true
+text_encoder:
+  name: qwen3_vl
+  pretrained: Qwen/Qwen3-VL-4B-Instruct
+  dim: 2560
+  max_length: 300
+  dtype: bfloat16
+  attn_implementation: sdpa
+  hidden_layers:
+  - 2
+  - 5
+  - 8
+  - 11
+  - 14
+  - 17
+  - 20
+  - 23
+  - 26
+  - 29
+  - 32
+  - 35
+  compile: false
+  on_caption_overflow: warn
+flow:
+  num_train_timesteps: 1000
+  shift: 4.0
+  timestep_sampler: logit_normal
+  logit_mean: 0.0
+  logit_std: 1.0
+  prediction: v
+  x_pred_sigma_min: 0.05
+  shift_law: none
+  shift_base_tokens: 256

depth/config.yaml

@@ -0,0 +1,79 @@
+model:
+  block: single_stream
+  dual_depth: 8
+  final_block_text: keep
+  hidden_size: 2560
+  depth: 24
+  num_heads: 20
+  num_kv_heads: 5
+  gated_attention: true
+  sandwich_norm: true
+  patch_size: 16
+  in_channels: 3
+  mlp_ratio: 4.0
+  qkv_bias: false
+  qk_norm: true
+  norm_eps: 1.0e-06
+  attn_backend: sdpa
+  modulation: shared_bias
+  modulation_rank: 64
+  timestep_max_period: 10.0
+  adaln_zero_init: true
+  rope_theta: 10000.0
+  rope_scale: 16.0
+  rope_aspect: isotropic
+  rope_frame_pairs: 0
+  rope_frame_theta: 10.0
+  text_rope: true
+  text_rope_theta: 10000.0
+  text_abs_pos_embed: true
+  text_dim: 2560
+  text_len: 300
+  text_adapter: lap_blocks2
+  text_lap_num_layers: 12
+  text_lap_num_heads: 32
+  text_lap_mlp_ratio: 1.3
+  repa_layer: 10
+  pixel:
+    enabled: true
+    depth: 4
+    hidden_size: 16
+    attn_hidden_size: 1280
+    num_heads: 10
+    mlp_ratio: 4.0
+    modulation: post
+    abs_pos_embed: true
+text_encoder:
+  name: qwen3_vl
+  pretrained: Qwen/Qwen3-VL-4B-Instruct
+  dim: 2560
+  max_length: 300
+  dtype: bfloat16
+  attn_implementation: sdpa
+  hidden_layers:
+  - 2
+  - 5
+  - 8
+  - 11
+  - 14
+  - 17
+  - 20
+  - 23
+  - 26
+  - 29
+  - 32
+  - 35
+  compile: false
+  on_caption_overflow: warn
+flow:
+  num_train_timesteps: 1000
+  shift: 4.0
+  timestep_sampler: logit_normal
+  logit_mean: 0.0
+  logit_std: 1.0
+  prediction: v
+  x_pred_sigma_min: 0.05
+  shift_law: none
+  shift_base_tokens: 256
+task:
+  name: depth

depth/empty_prompt.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea77e942f2ac7698bf4e5a2657d1820839b7e47212762ae31fbe5234beef33e6
+size 36864468

depth/manifest.json

@@ -0,0 +1,27 @@
+{
+  "format": "iris-rowwise-int8-v1",
+  "linear_count": 298,
+  "bytes": 3233728048,
+  "task": "depth",
+  "source_repo": "speridlabs/iris-3b",
+  "source_revision": "7445443349bc9abe3c96f01ff793e2098ca012b3",
+  "source_sha256": "e7aabbd5992637914407bfb5541f67c727ae30516d63eb32a85c541f8e5d8952",
+  "code_revision": "a8d15239dea469aba042cfa56ca3bb4e450d5ebc",
+  "files": [
+    {
+      "path": "model.safetensors",
+      "size": 3233825264,
+      "sha256": "ddb618148d1011bba0431aa24b9a5f27b11d3083f3628203f425d45c162a75a5"
+    },
+    {
+      "path": "config.yaml",
+      "size": 1427,
+      "sha256": "f2a9fbdccabc1a27ba5475ed0ad4bfc35dc854d1008be6a167f8b3de8bfd4dcf"
+    },
+    {
+      "path": "empty_prompt.safetensors",
+      "size": 36864468,
+      "sha256": "ea77e942f2ac7698bf4e5a2657d1820839b7e47212762ae31fbe5234beef33e6"
+    }
+  ]
+}

depth/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddb618148d1011bba0431aa24b9a5f27b11d3083f3628203f425d45c162a75a5
+size 3233825264

manifest.json

@@ -0,0 +1,22 @@
+{
+  "format": "iris-rowwise-int8-v1",
+  "linear_count": 298,
+  "bytes": 3233072608,
+  "task": "generate",
+  "source_repo": "speridlabs/iris-3b",
+  "source_revision": "7445443349bc9abe3c96f01ff793e2098ca012b3",
+  "source_sha256": "a90e1f4c197bac8491f82078def68884b6a3527ea7fc97a693e3e367372aa80f",
+  "code_revision": "a8d15239dea469aba042cfa56ca3bb4e450d5ebc",
+  "files": [
+    {
+      "path": "model.safetensors",
+      "size": 3233164048,
+      "sha256": "0c71d95482b0a3db64245a9754618e5082929d76e1b5fecc33dda6ba1fd948e5"
+    },
+    {
+      "path": "config.yaml",
+      "size": 1407,
+      "sha256": "077ce19156d9425a7e665a0059be02b3c5a1fcd7aea7607047b4f7486efa6cd5"
+    }
+  ]
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c71d95482b0a3db64245a9754618e5082929d76e1b5fecc33dda6ba1fd948e5
+size 3233164048

upscaler/config.yaml

@@ -0,0 +1,81 @@
+model:
+  block: single_stream
+  dual_depth: 8
+  final_block_text: keep
+  hidden_size: 2560
+  depth: 24
+  num_heads: 20
+  num_kv_heads: 5
+  gated_attention: true
+  sandwich_norm: true
+  patch_size: 16
+  in_channels: 3
+  mlp_ratio: 4.0
+  qkv_bias: false
+  qk_norm: true
+  norm_eps: 1.0e-06
+  attn_backend: sdpa
+  modulation: shared_bias
+  modulation_rank: 64
+  timestep_max_period: 10.0
+  adaln_zero_init: true
+  rope_theta: 10000.0
+  rope_scale: 16.0
+  rope_aspect: isotropic
+  rope_frame_pairs: 0
+  rope_frame_theta: 10.0
+  text_rope: true
+  text_rope_theta: 10000.0
+  text_abs_pos_embed: true
+  text_dim: 2560
+  text_len: 300
+  text_adapter: lap_blocks2
+  text_lap_num_layers: 12
+  text_lap_num_heads: 32
+  text_lap_mlp_ratio: 1.3
+  repa_layer: 10
+  pixel:
+    enabled: true
+    depth: 4
+    hidden_size: 16
+    attn_hidden_size: 1280
+    num_heads: 10
+    mlp_ratio: 4.0
+    modulation: post
+    abs_pos_embed: true
+text_encoder:
+  name: qwen3_vl
+  pretrained: Qwen/Qwen3-VL-4B-Instruct
+  dim: 2560
+  max_length: 300
+  dtype: bfloat16
+  attn_implementation: sdpa
+  hidden_layers:
+  - 2
+  - 5
+  - 8
+  - 11
+  - 14
+  - 17
+  - 20
+  - 23
+  - 26
+  - 29
+  - 32
+  - 35
+  compile: false
+  on_caption_overflow: warn
+flow:
+  num_train_timesteps: 1000
+  shift: 4.0
+  timestep_sampler: logit_normal
+  logit_mean: 0.0
+  logit_std: 1.0
+  prediction: v
+  x_pred_sigma_min: 0.05
+  shift_law: none
+  shift_base_tokens: 256
+task:
+  name: restoration
+  sigma: 0.5
+  tile: 1024

upscaler/empty_prompt.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3cef1162f07bb81f7fb37fc7c997a9305a000f4de2563e61982d5d373a61d225
+size 36864468

upscaler/manifest.json

@@ -0,0 +1,27 @@
+{
+  "format": "iris-rowwise-int8-v1",
+  "linear_count": 298,
+  "bytes": 3233072608,
+  "task": "upscale",
+  "source_repo": "speridlabs/iris-3b",
+  "source_revision": "7445443349bc9abe3c96f01ff793e2098ca012b3",
+  "source_sha256": "b4235e5f60bd1c0cb761f13b9549e31aefd25e3814aded4349f1f156ce0abeaa",
+  "code_revision": "a8d15239dea469aba042cfa56ca3bb4e450d5ebc",
+  "files": [
+    {
+      "path": "model.safetensors",
+      "size": 3233164048,
+      "sha256": "d7806539bb65ea6e1035b988707482e1cca6bc805801871a1bba49495a16e49e"
+    },
+    {
+      "path": "config.yaml",
+      "size": 1459,
+      "sha256": "da09cc9a84c6ed6014f9c788640ee24bca0b79c1f9fa55b947a52357ca32a51d"
+    },
+    {
+      "path": "empty_prompt.safetensors",
+      "size": 36864468,
+      "sha256": "3cef1162f07bb81f7fb37fc7c997a9305a000f4de2563e61982d5d373a61d225"
+    }
+  ]
+}

upscaler/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7806539bb65ea6e1035b988707482e1cca6bc805801871a1bba49495a16e49e
+size 3233164048