Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,39 @@ 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
+assets/gallery/000-1440x1440.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/001-1440x1440.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/002-1440x1440.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/003-1248x1664.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/004-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/005-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/006-1248x1664.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/007-1440x1440.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/008-1440x1440.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/009-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/010-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/011-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/012-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/013-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/014-1248x1664.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/015-1248x1664.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/016-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/017-1248x1664.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/018-1248x1664.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/019-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/020-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/021-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/022-1440x1440.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/023-1248x1664.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/024-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/025-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/026-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/027-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/028-1248x1664.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/029-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/030-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/gallery/031-1664x1248.png filter=lfs diff=lfs merge=lfs -text
+assets/teaser.webp filter=lfs diff=lfs merge=lfs -text
+tokenizer/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+assets/teaser.png filter=lfs diff=lfs merge=lfs -text
+assets/comparison/comparison_grid_1to1_q98.webp filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,206 @@
+---
+license: mit
+base_model: microsoft/Lens
+pipeline_tag: text-to-image
+tags:
+- lens
+- text-to-image
+- sdnq
+- uint4
+- static-quantization
+- ablation
+- model-cpu-offload
+---
+
+# Lens SDNQ uint4 static
+
+This is a corrected SDNQ static UINT4 quantized variant of [microsoft/Lens](https://huggingface.co/microsoft/Lens).
+
+The recipe follows the Lens-Turbo ablation result: all-linear UINT4 quantization can introduce periodic grid artifacts and severe text degradation when transformer modulation linears are quantized. This checkpoint keeps `*.img_mod.*` and `*.txt_mod.*` in bfloat16 and quantizes the rest of the denoising transformer with SDNQ UINT4.
+
+## Visual Comparison
+
+**Full-size comparison grid:** the image below is built from native 1440x1440 samples without resampling the image cells and saved as WebP quality 98. Raw file: [assets/comparison/comparison_grid_1to1_q98.webp](https://huggingface.co/WaveCut/Lens-SDNQ-uint4-static/resolve/main/assets/comparison/comparison_grid_1to1_q98.webp).
+
+![Original vs SDNQ comparison grid](assets/comparison/comparison_grid_1to1_q98.webp)
+
+## Quantization Recipe
+
+| Field | Value |
+| --- | --- |
+| Method | SDNQ uint4 static |
+| Source model | `microsoft/Lens` |
+| Quantized component | Denoising transformer |
+| Text encoder | Unchanged upstream GPT-OSS text encoder |
+| VAE | Unchanged upstream VAE |
+| `weights_dtype` | `uint4` |
+| `quantized_matmul_dtype` | `int8` |
+| `use_quantized_matmul` | `true` |
+| `group_size` | `0` |
+| `dequantize_fp32` | `false` |
+| Critical skip rule | `*.img_mod.*`, `*.txt_mod.*` kept in bfloat16 |
+
+## Usage
+
+Run from the cloned [microsoft/Lens](https://github.com/microsoft/Lens) repo root so the custom Lens classes are registered.
+
+```python
+import torch
+from huggingface_hub import snapshot_download
+from lens import LensPipeline, LensTransformer2DModel
+from sdnq import load_sdnq_model
+
+model_dir = snapshot_download("WaveCut/Lens-SDNQ-uint4-static")
+transformer = load_sdnq_model(
+    model_dir + "/transformer",
+    model_cls=LensTransformer2DModel,
+    dtype=torch.bfloat16,
+    device=torch.device("cuda"),
+    dequantize_fp32=False,
+    use_quantized_matmul=True,
+)
+pipe = LensPipeline.from_pretrained(
+    model_dir,
+    transformer=transformer,
+    torch_dtype=torch.bfloat16,
+).to("cuda")
+
+image = pipe(
+    prompt="A cat holding a sign that says hello world",
+    base_resolution=1440,
+    aspect_ratio="1:1",
+    num_inference_steps=20,
+    guidance_scale=5.0,
+    generator=torch.Generator("cuda").manual_seed(0),
+).images[0]
+```
+
+## Benchmark
+
+Hardware: RunPod NVIDIA H100 80GB HBM3 (H100 SXM), PyTorch 2.8.0 CUDA 12.8 container, local container disk only. Benchmark date: 2026-05-24. Generation settings: `base_resolution=1440`, `aspect_ratio="1:1"`, `num_inference_steps=20`, `guidance_scale=5.0`.
+
+| Metric | Original Lens | SDNQ uint4 static |
+| --- | ---: | ---: |
+| Load time, seconds | 17.641 | 14.605 |
+| Load peak allocated VRAM, GB | 22.342 | 18.446 |
+| Load peak reserved VRAM, GB | 22.471 | 18.516 |
+| Transformer tensor storage footprint, GB | 16.417 | 4.301 |
+| Transformer storage reduction vs original | baseline | 73.8% smaller |
+| Average prompt runtime, seconds | 15.357 | 17.937 |
+| Median prompt runtime, seconds | 14.346 | 17.813 |
+| Average generation peak allocated VRAM, GB | 27.462 | 23.533 |
+| Max generation peak allocated VRAM, GB | 27.467 | 23.538 |
+
+Transformer-only footprint is computed from safetensors tensor storage for the denoising transformer parameter tensors only; it excludes allocator overhead and non-transformer components. The original transformer tensors are F32; the corrected SDNQ transformer stores quantized tensors as U8 plus the excluded modulation layers as BF16.
+
+### Model CPU Offload Benchmark
+
+Same 10 prompts, using `pipe.enable_model_cpu_offload()`. The reported load time uses a warm local Hugging Face cache on the container disk, so model download time is excluded. Each model was measured in a fresh Python process. `Cold generation` is P01, the first generation immediately after load/offload setup; `warm generation` aggregates P02-P10.
+
+| Metric | Original Lens | SDNQ uint4 static |
+| --- | ---: | ---: |
+| Offload setup/load time, seconds | 15.510 | 13.573 |
+| Offload setup peak allocated VRAM, GB | 12.582 | 12.582 |
+| Offload setup peak reserved VRAM, GB | 13.881 | 13.881 |
+| Cold generation time, seconds | 26.217 | 21.473 |
+| Cold generation peak allocated VRAM, GB | 19.274 | 15.479 |
+| Cold generation peak reserved VRAM, GB | 19.608 | 19.126 |
+| Warm generation average time, seconds | 18.123 | 17.650 |
+| Warm generation median time, seconds | 17.178 | 17.519 |
+| Warm generation average peak allocated VRAM, GB | 19.271 | 15.480 |
+| Warm generation average peak reserved VRAM, GB | 19.630 | 18.965 |
+| Warm generation max peak allocated VRAM, GB | 19.276 | 15.482 |
+| Warm generation max peak reserved VRAM, GB | 19.803 | 19.210 |
+
+Raw metrics: [`benchmark_metrics.json`](benchmark_metrics.json), [`comparison_matrix.json`](comparison_matrix.json), [`model_cpu_offload_benchmark.json`](model_cpu_offload_benchmark.json), [`sdnq_quantization_summary.json`](sdnq_quantization_summary.json).
+
+## 10-Prompt Matrix
+
+| ID | Scenario | Seed | Original time, s | Quant time, s | Delta | Original peak allocated VRAM, GB | Quant peak allocated VRAM, GB |
+| --- | --- | ---: | ---: | ---: | ---: | ---: | ---: |
+| P01 | Midnight Library Weather Station | 301 | 19.518 | 24.105 | +23.5% | 27.461 | 23.532 |
+| P02 | Desert Observatory Treaty Room | 302 | 16.986 | 18.565 | +9.3% | 27.461 | 23.532 |
+| P03 | Arctic Submarine Greenhouse | 303 | 13.903 | 17.882 | +28.6% | 27.461 | 23.532 |
+| P04 | Long English Museum Labels | 304 | 14.439 | 17.848 | +23.6% | 27.461 | 23.533 |
+| P05 | Tokyo Rooftop Repair Diner | 305 | 14.253 | 17.779 | +24.7% | 27.461 | 23.532 |
+| P06 | Russian Provincial Print Shop | 306 | 16.744 | 18.136 | +8.3% | 27.467 | 23.538 |
+| P07 | Ocean Cartography Bakery | 307 | 13.918 | 14.828 | +6.5% | 27.461 | 23.532 |
+| P08 | Long English Train Notice Wall | 308 | 13.906 | 14.826 | +6.6% | 27.461 | 23.532 |
+| P09 | Orbital Botanical Courtroom | 309 | 16.204 | 17.740 | +9.5% | 27.461 | 23.533 |
+| P10 | Byzantine Data Center Chapel | 310 | 13.703 | 17.665 | +28.9% | 27.461 | 23.532 |
+
+## Full Prompts
+
+<details>
+<summary>P01 - Midnight Library Weather Station</summary>
+
+A vast midnight library converted into a Victorian weather station, brass barometers, hanging cloud chambers, blue lightning outside stained-glass windows, spiral ladders, rainwater collecting in crystal funnels, and readable labels everywhere. Include a large oak sign saying "ARCHIVE OF STORMS - EAST WING", a ledger title saying "BAROMETRIC ANOMALIES 1897-1903", a small drawer label saying "FOG SAMPLES / DO NOT SHAKE", a chalkboard note saying "THUNDER ARRIVES AT 02:17", and a bookmark saying "RETURN TO SHELF C-19". Extremely detailed, cinematic, natural perspective, crisp small typography.
+
+</details>
+
+<details>
+<summary>P02 - Desert Observatory Treaty Room</summary>
+
+An ancient desert observatory at golden hour, now used as a treaty room for astronomers and nomad diplomats, sandstone arches, astrolabes, folded star maps, copper tea service, wind-blown curtains, tiny dust motes, and many readable inscriptions. The central parchment must read "TREATY OF THE SEVEN MOONS". A wall plaque reads "OBSERVATORY OF QASR AL-SUHAIL". A tea label says "CARDAMOM - NO SUGAR". A blue wax seal says "WITNESSED UNDER MARS". A telescope tag says "CALIBRATE BEFORE SUNSET". Hyperreal, warm shadows, intricate surface wear.
+
+</details>
+
+<details>
+<summary>P03 - Arctic Submarine Greenhouse</summary>
+
+A transparent research submarine trapped under Arctic ice, transformed into a warm hydroponic greenhouse with orange grow lights, condensation, polar bears visible above through thick ice, scientists in wool sweaters, algae tanks, and frost patterns on glass. Include readable text on multiple objects: "POLAR BOTANY UNIT 4" on the bulkhead, "EMERGENCY SEED VAULT" on a red locker, "LIGHT CYCLE: 18 HOURS" on a tablet, "DO NOT FEED THE KELP" on a handwritten note, and "RETURN CORE SAMPLES" on a metal tray. Detailed, atmospheric, believable engineering.
+
+</details>
+
+<details>
+<summary>P04 - Long English Museum Labels</summary>
+
+A photorealistic museum exhibit room about impossible machines, with glass cases, velvet ropes, soft spotlights, and several long English placards that must be visible on different parts of the image. Placard one reads: "THE CLOCK THAT REMEMBERED WINTER: assembled from brass, bone, and borrowed tides, circa 1814." Placard two reads: "PLEASE DO NOT TOUCH THE PERPETUAL ENGINE; it becomes anxious when observed too closely." Placard three reads: "CURATOR'S NOTE: every gear was catalogued, polished, numbered, and returned before dawn." Also include ticket stubs, tiny accession numbers, fingerprints on glass, and realistic museum lighting.
+
+</details>
+
+<details>
+<summary>P05 - Tokyo Rooftop Repair Diner</summary>
+
+A rainy Tokyo rooftop diner that doubles as a robot repair shop, neon reflections, steam from ramen bowls, umbrellas, tiny servo motors, handwritten order slips, rain beads on chrome, and a skyline full of antennas. Readable signs: a pink neon sign says "MIDNIGHT RAMEN & REPAIRS", a menu board says "SPECIAL: MISO, BATTERY PACK, GREEN TEA", a repair invoice says "UNIT 7B - LEFT HAND RECALIBRATION", a sticker says "NO DRONES AFTER 2 AM", and a paper lantern says "OPEN WHEN IT RAINS". High detail, shallow depth of field, cinematic realism.
+
+</details>
+
+<details>
+<summary>P06 - Russian Provincial Print Shop</summary>
+
+Старинная провинциальная типография в России, поздний вечер, керосиновые лампы, деревянные кассы со свинцовыми литерами, мокрые афиши на веревках, самовар, иней на окне, реалистичная пыль и следы краски. На большой вывеске долж��о быть написано: "ТИПОГРАФИЯ УЕЗДНЫХ ВЕСТЕЙ". На длинной афише читаемый текст: "Завтра в городском саду: лекция о кометах, духовой оркестр, чай с баранками, начало ровно в семь часов вечера". На ящике: "ЛИТЕРЫ: А-Я, НЕ РОНЯТЬ". На записке: "Срочно отпечатать до рассвета". Очень детально, без мультяшности.
+
+</details>
+
+<details>
+<summary>P07 - Ocean Cartography Bakery</summary>
+
+A cozy bakery inside an old ocean cartography office, with croissants shaped like sea monsters, nautical charts dusted with flour, brass compasses, jars of ink, morning light, and a baker drawing coastlines in powdered sugar. Text elements: "TIDAL BREAD & MAPS" on the front sign, "SOURDOUGH CURRENT - 6:30 AM" on a chalkboard, "UNCHARTED PLUM TARTS" on a pastry label, "DO NOT EAT THE COMPASS" on a note, and "NORTH SEA BATCH 12" stamped on a paper bag. Warm, detailed, whimsical but realistic.
+
+</details>
+
+<details>
+<summary>P08 - Long English Train Notice Wall</summary>
+
+A foggy Edwardian railway platform at dawn with a wall of overlapping long English notices, brass lamps, wet cobblestones, porters, suitcases, pigeons, steam, and reflections. The largest notice must read: "IMPORTANT SERVICE CHANGE: The 6:42 express to Northbridge will depart from Platform Three after the moonlit freight has cleared the signal box." A second poster reads: "LOST PROPERTY: one violin case, two blue gloves, a silver compass, and a letter never posted." A timetable says "WINTER ROUTE - DELAYS EXPECTED NEAR THE MARSH". Ultra detailed, cinematic, legible signs, natural perspective.
+
+</details>
+
+<details>
+<summary>P09 - Orbital Botanical Courtroom</summary>
+
+A surreal but photorealistic courtroom inside an orbital botanical garden, judges in dark robes, enormous ferns, floating pollen, Earth visible through a curved window, holographic evidence screens, and a tiny robot stenographer. Required readable text: "CASE 44-B: THE PEOPLE VS. THE SUNFLOWER" on the main screen, "EVIDENCE: THREE PETALS AND A BROKEN VASE" on a side display, "SILENCE IN THE GREENHOUSE COURT" on a sign, "WITNESS: DR. LYSANDER MOSS" on a nameplate, and "OXYGEN TAX RECEIPT" on a paper slip. Sharp, high-detail, dramatic lighting.
+
+</details>
+
+<details>
+<summary>P10 - Byzantine Data Center Chapel</summary>
+
+A Byzantine chapel converted into a quiet data center, gold mosaics reflecting server LEDs, incense smoke, marble floors, monks maintaining fiber cables, illuminated manuscripts next to diagnostic terminals, and beautiful cable management. Text must appear in multiple places: "SANCTUM SERVER ROOM - AUTHORIZED MONKS ONLY" on a bronze door, "BACKUP PSALMS COMPLETED AT 03:12" on a terminal, "DO NOT UNPLUG THE RELIQUARY" on a warning label, "LATENCY PRAYER REQUESTS" on a clipboard, and "ARCHIVE NODE IX" etched on a server rack. Rich texture, controlled highlights, realistic scale.
+
+</details> 
+
+## Notes
+
+This checkpoint is intended for research and evaluation. It inherits the upstream Lens limitations and responsible AI considerations from the source model. Text rendering remains challenging; the corrected recipe is designed to avoid the obvious grid/printed texture failure seen when transformer modulation linears are quantized.

benchmark_metrics.json

@@ -0,0 +1,349 @@
+{
+  "hardware": "RunPod NVIDIA H100 80GB HBM3 (H100 SXM)",
+  "mode": "full GPU, pipe.to('cuda')",
+  "base_resolution": 1440,
+  "aspect_ratio": "1:1",
+  "num_inference_steps": 20,
+  "guidance_scale": 5.0,
+  "dtype": "torch.bfloat16",
+  "models": {
+    "base": {
+      "load": {
+        "load_time_s": 17.641,
+        "peak_allocated_gb": 22.342,
+        "peak_reserved_gb": 22.471,
+        "end_allocated_gb": 22.342,
+        "end_reserved_gb": 22.471
+      },
+      "summary": {
+        "avg_time_s": 15.357,
+        "median_time_s": 14.346,
+        "avg_peak_allocated_gb": 27.462,
+        "avg_peak_reserved_gb": 31.756,
+        "max_peak_allocated_gb": 27.467,
+        "max_peak_reserved_gb": 31.965
+      },
+      "prompts": [
+        {
+          "id": "P01",
+          "title": "Midnight Library Weather Station",
+          "seed": 301,
+          "time_s": 19.518,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.461,
+          "peak_reserved_gb": 31.74,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.74,
+          "image": "/workspace/lens_sdnq_work/generated/base/p01_base.png"
+        },
+        {
+          "id": "P02",
+          "title": "Desert Observatory Treaty Room",
+          "seed": 302,
+          "time_s": 16.986,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.461,
+          "peak_reserved_gb": 31.74,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.74,
+          "image": "/workspace/lens_sdnq_work/generated/base/p02_base.png"
+        },
+        {
+          "id": "P03",
+          "title": "Arctic Submarine Greenhouse",
+          "seed": 303,
+          "time_s": 13.903,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.461,
+          "peak_reserved_gb": 31.719,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.719,
+          "image": "/workspace/lens_sdnq_work/generated/base/p03_base.png"
+        },
+        {
+          "id": "P04",
+          "title": "Long English Museum Labels",
+          "seed": 304,
+          "time_s": 14.439,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.461,
+          "peak_reserved_gb": 31.719,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.719,
+          "image": "/workspace/lens_sdnq_work/generated/base/p04_base.png"
+        },
+        {
+          "id": "P05",
+          "title": "Tokyo Rooftop Repair Diner",
+          "seed": 305,
+          "time_s": 14.253,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.461,
+          "peak_reserved_gb": 31.74,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.74,
+          "image": "/workspace/lens_sdnq_work/generated/base/p05_base.png"
+        },
+        {
+          "id": "P06",
+          "title": "Russian Provincial Print Shop",
+          "seed": 306,
+          "time_s": 16.744,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.467,
+          "peak_reserved_gb": 31.965,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.965,
+          "image": "/workspace/lens_sdnq_work/generated/base/p06_base.png"
+        },
+        {
+          "id": "P07",
+          "title": "Ocean Cartography Bakery",
+          "seed": 307,
+          "time_s": 13.918,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.461,
+          "peak_reserved_gb": 31.74,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.74,
+          "image": "/workspace/lens_sdnq_work/generated/base/p07_base.png"
+        },
+        {
+          "id": "P08",
+          "title": "Long English Train Notice Wall",
+          "seed": 308,
+          "time_s": 13.906,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.461,
+          "peak_reserved_gb": 31.74,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.74,
+          "image": "/workspace/lens_sdnq_work/generated/base/p08_base.png"
+        },
+        {
+          "id": "P09",
+          "title": "Orbital Botanical Courtroom",
+          "seed": 309,
+          "time_s": 16.204,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.461,
+          "peak_reserved_gb": 31.719,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.719,
+          "image": "/workspace/lens_sdnq_work/generated/base/p09_base.png"
+        },
+        {
+          "id": "P10",
+          "title": "Byzantine Data Center Chapel",
+          "seed": 310,
+          "time_s": 13.703,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 27.461,
+          "peak_reserved_gb": 31.74,
+          "end_allocated_gb": 22.377,
+          "end_reserved_gb": 31.74,
+          "image": "/workspace/lens_sdnq_work/generated/base/p10_base.png"
+        }
+      ]
+    },
+    "quant": {
+      "load": {
+        "load_time_s": 14.605,
+        "peak_allocated_gb": 18.446,
+        "peak_reserved_gb": 18.516,
+        "end_allocated_gb": 18.446,
+        "end_reserved_gb": 18.516
+      },
+      "summary": {
+        "avg_time_s": 17.937,
+        "median_time_s": 17.813,
+        "avg_peak_allocated_gb": 23.533,
+        "avg_peak_reserved_gb": 27.699,
+        "max_peak_allocated_gb": 23.538,
+        "max_peak_reserved_gb": 27.712
+      },
+      "prompts": [
+        {
+          "id": "P01",
+          "title": "Midnight Library Weather Station",
+          "seed": 301,
+          "time_s": 24.105,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.532,
+          "peak_reserved_gb": 27.682,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.682,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p01_quant.png"
+        },
+        {
+          "id": "P02",
+          "title": "Desert Observatory Treaty Room",
+          "seed": 302,
+          "time_s": 18.565,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.532,
+          "peak_reserved_gb": 27.712,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.712,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p02_quant.png"
+        },
+        {
+          "id": "P03",
+          "title": "Arctic Submarine Greenhouse",
+          "seed": 303,
+          "time_s": 17.882,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.532,
+          "peak_reserved_gb": 27.712,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.712,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p03_quant.png"
+        },
+        {
+          "id": "P04",
+          "title": "Long English Museum Labels",
+          "seed": 304,
+          "time_s": 17.848,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.533,
+          "peak_reserved_gb": 27.712,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.712,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p04_quant.png"
+        },
+        {
+          "id": "P05",
+          "title": "Tokyo Rooftop Repair Diner",
+          "seed": 305,
+          "time_s": 17.779,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.532,
+          "peak_reserved_gb": 27.712,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.712,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p05_quant.png"
+        },
+        {
+          "id": "P06",
+          "title": "Russian Provincial Print Shop",
+          "seed": 306,
+          "time_s": 18.136,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.538,
+          "peak_reserved_gb": 27.615,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.615,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p06_quant.png"
+        },
+        {
+          "id": "P07",
+          "title": "Ocean Cartography Bakery",
+          "seed": 307,
+          "time_s": 14.828,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.532,
+          "peak_reserved_gb": 27.712,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.712,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p07_quant.png"
+        },
+        {
+          "id": "P08",
+          "title": "Long English Train Notice Wall",
+          "seed": 308,
+          "time_s": 14.826,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.532,
+          "peak_reserved_gb": 27.712,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.712,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p08_quant.png"
+        },
+        {
+          "id": "P09",
+          "title": "Orbital Botanical Courtroom",
+          "seed": 309,
+          "time_s": 17.74,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.533,
+          "peak_reserved_gb": 27.712,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.712,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p09_quant.png"
+        },
+        {
+          "id": "P10",
+          "title": "Byzantine Data Center Chapel",
+          "seed": 310,
+          "time_s": 17.665,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 23.532,
+          "peak_reserved_gb": 27.712,
+          "end_allocated_gb": 18.448,
+          "end_reserved_gb": 27.712,
+          "image": "/workspace/lens_sdnq_work/generated/quant/p10_quant.png"
+        }
+      ]
+    }
+  }
+}

comparison_matrix.json

@@ -0,0 +1,362 @@
+[
+  {
+    "id": "P01",
+    "title": "Midnight Library Weather Station",
+    "seed": 301,
+    "prompt": "A vast midnight library converted into a Victorian weather station, brass barometers, hanging cloud chambers, blue lightning outside stained-glass windows, spiral ladders, rainwater collecting in crystal funnels, and readable labels everywhere. Include a large oak sign saying \"ARCHIVE OF STORMS - EAST WING\", a ledger title saying \"BAROMETRIC ANOMALIES 1897-1903\", a small drawer label saying \"FOG SAMPLES / DO NOT SHAKE\", a chalkboard note saying \"THUNDER ARRIVES AT 02:17\", and a bookmark saying \"RETURN TO SHELF C-19\". Extremely detailed, cinematic, natural perspective, crisp small typography.",
+    "base": {
+      "id": "P01",
+      "title": "Midnight Library Weather Station",
+      "seed": 301,
+      "time_s": 19.518,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.461,
+      "peak_reserved_gb": 31.74,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.74,
+      "image": "/workspace/lens_sdnq_work/generated/base/p01_base.png"
+    },
+    "quant": {
+      "id": "P01",
+      "title": "Midnight Library Weather Station",
+      "seed": 301,
+      "time_s": 24.105,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.532,
+      "peak_reserved_gb": 27.682,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.682,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p01_quant.png"
+    }
+  },
+  {
+    "id": "P02",
+    "title": "Desert Observatory Treaty Room",
+    "seed": 302,
+    "prompt": "An ancient desert observatory at golden hour, now used as a treaty room for astronomers and nomad diplomats, sandstone arches, astrolabes, folded star maps, copper tea service, wind-blown curtains, tiny dust motes, and many readable inscriptions. The central parchment must read \"TREATY OF THE SEVEN MOONS\". A wall plaque reads \"OBSERVATORY OF QASR AL-SUHAIL\". A tea label says \"CARDAMOM - NO SUGAR\". A blue wax seal says \"WITNESSED UNDER MARS\". A telescope tag says \"CALIBRATE BEFORE SUNSET\". Hyperreal, warm shadows, intricate surface wear.",
+    "base": {
+      "id": "P02",
+      "title": "Desert Observatory Treaty Room",
+      "seed": 302,
+      "time_s": 16.986,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.461,
+      "peak_reserved_gb": 31.74,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.74,
+      "image": "/workspace/lens_sdnq_work/generated/base/p02_base.png"
+    },
+    "quant": {
+      "id": "P02",
+      "title": "Desert Observatory Treaty Room",
+      "seed": 302,
+      "time_s": 18.565,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.532,
+      "peak_reserved_gb": 27.712,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.712,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p02_quant.png"
+    }
+  },
+  {
+    "id": "P03",
+    "title": "Arctic Submarine Greenhouse",
+    "seed": 303,
+    "prompt": "A transparent research submarine trapped under Arctic ice, transformed into a warm hydroponic greenhouse with orange grow lights, condensation, polar bears visible above through thick ice, scientists in wool sweaters, algae tanks, and frost patterns on glass. Include readable text on multiple objects: \"POLAR BOTANY UNIT 4\" on the bulkhead, \"EMERGENCY SEED VAULT\" on a red locker, \"LIGHT CYCLE: 18 HOURS\" on a tablet, \"DO NOT FEED THE KELP\" on a handwritten note, and \"RETURN CORE SAMPLES\" on a metal tray. Detailed, atmospheric, believable engineering.",
+    "base": {
+      "id": "P03",
+      "title": "Arctic Submarine Greenhouse",
+      "seed": 303,
+      "time_s": 13.903,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.461,
+      "peak_reserved_gb": 31.719,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.719,
+      "image": "/workspace/lens_sdnq_work/generated/base/p03_base.png"
+    },
+    "quant": {
+      "id": "P03",
+      "title": "Arctic Submarine Greenhouse",
+      "seed": 303,
+      "time_s": 17.882,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.532,
+      "peak_reserved_gb": 27.712,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.712,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p03_quant.png"
+    }
+  },
+  {
+    "id": "P04",
+    "title": "Long English Museum Labels",
+    "seed": 304,
+    "prompt": "A photorealistic museum exhibit room about impossible machines, with glass cases, velvet ropes, soft spotlights, and several long English placards that must be visible on different parts of the image. Placard one reads: \"THE CLOCK THAT REMEMBERED WINTER: assembled from brass, bone, and borrowed tides, circa 1814.\" Placard two reads: \"PLEASE DO NOT TOUCH THE PERPETUAL ENGINE; it becomes anxious when observed too closely.\" Placard three reads: \"CURATOR'S NOTE: every gear was catalogued, polished, numbered, and returned before dawn.\" Also include ticket stubs, tiny accession numbers, fingerprints on glass, and realistic museum lighting.",
+    "base": {
+      "id": "P04",
+      "title": "Long English Museum Labels",
+      "seed": 304,
+      "time_s": 14.439,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.461,
+      "peak_reserved_gb": 31.719,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.719,
+      "image": "/workspace/lens_sdnq_work/generated/base/p04_base.png"
+    },
+    "quant": {
+      "id": "P04",
+      "title": "Long English Museum Labels",
+      "seed": 304,
+      "time_s": 17.848,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.533,
+      "peak_reserved_gb": 27.712,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.712,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p04_quant.png"
+    }
+  },
+  {
+    "id": "P05",
+    "title": "Tokyo Rooftop Repair Diner",
+    "seed": 305,
+    "prompt": "A rainy Tokyo rooftop diner that doubles as a robot repair shop, neon reflections, steam from ramen bowls, umbrellas, tiny servo motors, handwritten order slips, rain beads on chrome, and a skyline full of antennas. Readable signs: a pink neon sign says \"MIDNIGHT RAMEN & REPAIRS\", a menu board says \"SPECIAL: MISO, BATTERY PACK, GREEN TEA\", a repair invoice says \"UNIT 7B - LEFT HAND RECALIBRATION\", a sticker says \"NO DRONES AFTER 2 AM\", and a paper lantern says \"OPEN WHEN IT RAINS\". High detail, shallow depth of field, cinematic realism.",
+    "base": {
+      "id": "P05",
+      "title": "Tokyo Rooftop Repair Diner",
+      "seed": 305,
+      "time_s": 14.253,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.461,
+      "peak_reserved_gb": 31.74,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.74,
+      "image": "/workspace/lens_sdnq_work/generated/base/p05_base.png"
+    },
+    "quant": {
+      "id": "P05",
+      "title": "Tokyo Rooftop Repair Diner",
+      "seed": 305,
+      "time_s": 17.779,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.532,
+      "peak_reserved_gb": 27.712,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.712,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p05_quant.png"
+    }
+  },
+  {
+    "id": "P06",
+    "title": "Russian Provincial Print Shop",
+    "seed": 306,
+    "prompt": "\u0421\u0442\u0430\u0440\u0438\u043d\u043d\u0430\u044f \u043f\u0440\u043e\u0432\u0438\u043d\u0446\u0438\u0430\u043b\u044c\u043d\u0430\u044f \u0442\u0438\u043f\u043e\u0433\u0440\u0430\u0444\u0438\u044f \u0432 \u0420\u043e\u0441\u0441\u0438\u0438, \u043f\u043e\u0437\u0434\u043d\u0438\u0439 \u0432\u0435\u0447\u0435\u0440, \u043a\u0435\u0440\u043e\u0441\u0438\u043d\u043e\u0432\u044b\u0435 \u043b\u0430\u043c\u043f\u044b, \u0434\u0435\u0440\u0435\u0432\u044f\u043d\u043d\u044b\u0435 \u043a\u0430\u0441\u0441\u044b \u0441\u043e \u0441\u0432\u0438\u043d\u0446\u043e\u0432\u044b\u043c\u0438 \u043b\u0438\u0442\u0435\u0440\u0430\u043c\u0438, \u043c\u043e\u043a\u0440\u044b\u0435 \u0430\u0444\u0438\u0448\u0438 \u043d\u0430 \u0432\u0435\u0440\u0435\u0432\u043a\u0430\u0445, \u0441\u0430\u043c\u043e\u0432\u0430\u0440, \u0438\u043d\u0435\u0439 \u043d\u0430 \u043e\u043a\u043d\u0435, \u0440\u0435\u0430\u043b\u0438\u0441\u0442\u0438\u0447\u043d\u0430\u044f \u043f\u044b\u043b\u044c \u0438 \u0441\u043b\u0435\u0434\u044b \u043a\u0440\u0430\u0441\u043a\u0438. \u041d\u0430 \u0431\u043e\u043b\u044c\u0448\u043e\u0439 \u0432\u044b\u0432\u0435\u0441\u043a\u0435 \u0434\u043e\u043b\u0436\u043d\u043e \u0431\u044b\u0442\u044c \u043d\u0430\u043f\u0438\u0441\u0430\u043d\u043e: \"\u0422\u0418\u041f\u041e\u0413\u0420\u0410\u0424\u0418\u042f \u0423\u0415\u0417\u0414\u041d\u042b\u0425 \u0412\u0415\u0421\u0422\u0415\u0419\". \u041d\u0430 \u0434\u043b\u0438\u043d\u043d\u043e\u0439 \u0430\u0444\u0438\u0448\u0435 \u0447\u0438\u0442\u0430\u0435\u043c\u044b\u0439 \u0442\u0435\u043a\u0441\u0442: \"\u0417\u0430\u0432\u0442\u0440\u0430 \u0432 \u0433\u043e\u0440\u043e\u0434\u0441\u043a\u043e\u043c \u0441\u0430\u0434\u0443: \u043b\u0435\u043a\u0446\u0438\u044f \u043e \u043a\u043e\u043c\u0435\u0442\u0430\u0445, \u0434\u0443\u0445\u043e\u0432\u043e\u0439 \u043e\u0440\u043a\u0435\u0441\u0442\u0440, \u0447\u0430\u0439 \u0441 \u0431\u0430\u0440\u0430\u043d\u043a\u0430\u043c\u0438, \u043d\u0430\u0447\u0430\u043b\u043e \u0440\u043e\u0432\u043d\u043e \u0432 \u0441\u0435\u043c\u044c \u0447\u0430\u0441\u043e\u0432 \u0432\u0435\u0447\u0435\u0440\u0430\". \u041d\u0430 \u044f\u0449\u0438\u043a\u0435: \"\u041b\u0418\u0422\u0415\u0420\u042b: \u0410-\u042f, \u041d\u0415 \u0420\u041e\u041d\u042f\u0422\u042c\". \u041d\u0430 \u0437\u0430\u043f\u0438\u0441\u043a\u0435: \"\u0421\u0440\u043e\u0447\u043d\u043e \u043e\u0442\u043f\u0435\u0447\u0430\u0442\u0430\u0442\u044c \u0434\u043e \u0440\u0430\u0441\u0441\u0432\u0435\u0442\u0430\". \u041e\u0447\u0435\u043d\u044c \u0434\u0435\u0442\u0430\u043b\u044c\u043d\u043e, \u0431\u0435\u0437 \u043c\u0443\u043b\u044c\u0442\u044f\u0448\u043d\u043e\u0441\u0442\u0438.",
+    "base": {
+      "id": "P06",
+      "title": "Russian Provincial Print Shop",
+      "seed": 306,
+      "time_s": 16.744,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.467,
+      "peak_reserved_gb": 31.965,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.965,
+      "image": "/workspace/lens_sdnq_work/generated/base/p06_base.png"
+    },
+    "quant": {
+      "id": "P06",
+      "title": "Russian Provincial Print Shop",
+      "seed": 306,
+      "time_s": 18.136,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.538,
+      "peak_reserved_gb": 27.615,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.615,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p06_quant.png"
+    }
+  },
+  {
+    "id": "P07",
+    "title": "Ocean Cartography Bakery",
+    "seed": 307,
+    "prompt": "A cozy bakery inside an old ocean cartography office, with croissants shaped like sea monsters, nautical charts dusted with flour, brass compasses, jars of ink, morning light, and a baker drawing coastlines in powdered sugar. Text elements: \"TIDAL BREAD & MAPS\" on the front sign, \"SOURDOUGH CURRENT - 6:30 AM\" on a chalkboard, \"UNCHARTED PLUM TARTS\" on a pastry label, \"DO NOT EAT THE COMPASS\" on a note, and \"NORTH SEA BATCH 12\" stamped on a paper bag. Warm, detailed, whimsical but realistic.",
+    "base": {
+      "id": "P07",
+      "title": "Ocean Cartography Bakery",
+      "seed": 307,
+      "time_s": 13.918,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.461,
+      "peak_reserved_gb": 31.74,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.74,
+      "image": "/workspace/lens_sdnq_work/generated/base/p07_base.png"
+    },
+    "quant": {
+      "id": "P07",
+      "title": "Ocean Cartography Bakery",
+      "seed": 307,
+      "time_s": 14.828,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.532,
+      "peak_reserved_gb": 27.712,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.712,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p07_quant.png"
+    }
+  },
+  {
+    "id": "P08",
+    "title": "Long English Train Notice Wall",
+    "seed": 308,
+    "prompt": "A foggy Edwardian railway platform at dawn with a wall of overlapping long English notices, brass lamps, wet cobblestones, porters, suitcases, pigeons, steam, and reflections. The largest notice must read: \"IMPORTANT SERVICE CHANGE: The 6:42 express to Northbridge will depart from Platform Three after the moonlit freight has cleared the signal box.\" A second poster reads: \"LOST PROPERTY: one violin case, two blue gloves, a silver compass, and a letter never posted.\" A timetable says \"WINTER ROUTE - DELAYS EXPECTED NEAR THE MARSH\". Ultra detailed, cinematic, legible signs, natural perspective.",
+    "base": {
+      "id": "P08",
+      "title": "Long English Train Notice Wall",
+      "seed": 308,
+      "time_s": 13.906,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.461,
+      "peak_reserved_gb": 31.74,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.74,
+      "image": "/workspace/lens_sdnq_work/generated/base/p08_base.png"
+    },
+    "quant": {
+      "id": "P08",
+      "title": "Long English Train Notice Wall",
+      "seed": 308,
+      "time_s": 14.826,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.532,
+      "peak_reserved_gb": 27.712,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.712,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p08_quant.png"
+    }
+  },
+  {
+    "id": "P09",
+    "title": "Orbital Botanical Courtroom",
+    "seed": 309,
+    "prompt": "A surreal but photorealistic courtroom inside an orbital botanical garden, judges in dark robes, enormous ferns, floating pollen, Earth visible through a curved window, holographic evidence screens, and a tiny robot stenographer. Required readable text: \"CASE 44-B: THE PEOPLE VS. THE SUNFLOWER\" on the main screen, \"EVIDENCE: THREE PETALS AND A BROKEN VASE\" on a side display, \"SILENCE IN THE GREENHOUSE COURT\" on a sign, \"WITNESS: DR. LYSANDER MOSS\" on a nameplate, and \"OXYGEN TAX RECEIPT\" on a paper slip. Sharp, high-detail, dramatic lighting.",
+    "base": {
+      "id": "P09",
+      "title": "Orbital Botanical Courtroom",
+      "seed": 309,
+      "time_s": 16.204,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.461,
+      "peak_reserved_gb": 31.719,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.719,
+      "image": "/workspace/lens_sdnq_work/generated/base/p09_base.png"
+    },
+    "quant": {
+      "id": "P09",
+      "title": "Orbital Botanical Courtroom",
+      "seed": 309,
+      "time_s": 17.74,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.533,
+      "peak_reserved_gb": 27.712,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.712,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p09_quant.png"
+    }
+  },
+  {
+    "id": "P10",
+    "title": "Byzantine Data Center Chapel",
+    "seed": 310,
+    "prompt": "A Byzantine chapel converted into a quiet data center, gold mosaics reflecting server LEDs, incense smoke, marble floors, monks maintaining fiber cables, illuminated manuscripts next to diagnostic terminals, and beautiful cable management. Text must appear in multiple places: \"SANCTUM SERVER ROOM - AUTHORIZED MONKS ONLY\" on a bronze door, \"BACKUP PSALMS COMPLETED AT 03:12\" on a terminal, \"DO NOT UNPLUG THE RELIQUARY\" on a warning label, \"LATENCY PRAYER REQUESTS\" on a clipboard, and \"ARCHIVE NODE IX\" etched on a server rack. Rich texture, controlled highlights, realistic scale.",
+    "base": {
+      "id": "P10",
+      "title": "Byzantine Data Center Chapel",
+      "seed": 310,
+      "time_s": 13.703,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 27.461,
+      "peak_reserved_gb": 31.74,
+      "end_allocated_gb": 22.377,
+      "end_reserved_gb": 31.74,
+      "image": "/workspace/lens_sdnq_work/generated/base/p10_base.png"
+    },
+    "quant": {
+      "id": "P10",
+      "title": "Byzantine Data Center Chapel",
+      "seed": 310,
+      "time_s": 17.665,
+      "image_size": [
+        1440,
+        1440
+      ],
+      "peak_allocated_gb": 23.532,
+      "peak_reserved_gb": 27.712,
+      "end_allocated_gb": 18.448,
+      "end_reserved_gb": 27.712,
+      "image": "/workspace/lens_sdnq_work/generated/quant/p10_quant.png"
+    }
+  }
+]

model_cpu_offload_benchmark.json

@@ -0,0 +1,363 @@
+{
+  "benchmark": "model_cpu_offload_cold_warm",
+  "hardware": "RunPod NVIDIA H100 80GB HBM3 (H100 SXM)",
+  "mode": "Diffusers enable_model_cpu_offload()",
+  "cache_state": "warm local HF cache; no download time included",
+  "process_isolation": "single model per fresh Python process",
+  "base_resolution": 1440,
+  "aspect_ratio": "1:1",
+  "num_inference_steps": 20,
+  "guidance_scale": 5.0,
+  "dtype": "torch.bfloat16",
+  "definitions": {
+    "load_time_s": "Pipeline load plus enable_model_cpu_offload setup from warm local HF cache; download time excluded.",
+    "cold_generation": "P01, first generation immediately after fresh process load/offload setup.",
+    "warm_generation": "P02-P10 after the cold P01 generation."
+  },
+  "models": {
+    "base": {
+      "hardware": "RunPod NVIDIA H100 80GB HBM3 (H100 SXM)",
+      "mode": "Diffusers enable_model_cpu_offload()",
+      "cache_state": "warm local HF cache; no download time included",
+      "process_isolation": "single model per fresh Python process",
+      "base_resolution": 1440,
+      "aspect_ratio": "1:1",
+      "num_inference_steps": 20,
+      "guidance_scale": 5.0,
+      "dtype": "torch.bfloat16",
+      "kind": "base",
+      "load": {
+        "load_time_s": 15.51,
+        "peak_allocated_gb": 12.582,
+        "peak_reserved_gb": 13.881,
+        "end_allocated_gb": 10.185,
+        "end_reserved_gb": 10.679
+      },
+      "summary": {
+        "cold_time_s": 26.217,
+        "cold_peak_allocated_gb": 19.274,
+        "cold_peak_reserved_gb": 19.608,
+        "warm_avg_time_s": 18.123,
+        "warm_median_time_s": 17.178,
+        "warm_avg_peak_allocated_gb": 19.271,
+        "warm_avg_peak_reserved_gb": 19.63,
+        "warm_max_peak_allocated_gb": 19.276,
+        "warm_max_peak_reserved_gb": 19.803
+      },
+      "prompts": [
+        {
+          "id": "P01",
+          "title": "Midnight Library Weather Station",
+          "seed": 301,
+          "time_s": 26.217,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.274,
+          "peak_reserved_gb": 19.608,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.884
+        },
+        {
+          "id": "P02",
+          "title": "Desert Observatory Treaty Room",
+          "seed": 302,
+          "time_s": 21.285,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.274,
+          "peak_reserved_gb": 19.608,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.884
+        },
+        {
+          "id": "P03",
+          "title": "Arctic Submarine Greenhouse",
+          "seed": 303,
+          "time_s": 17.271,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.275,
+          "peak_reserved_gb": 19.608,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.809
+        },
+        {
+          "id": "P04",
+          "title": "Long English Museum Labels",
+          "seed": 304,
+          "time_s": 17.178,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.275,
+          "peak_reserved_gb": 19.608,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.884
+        },
+        {
+          "id": "P05",
+          "title": "Tokyo Rooftop Repair Diner",
+          "seed": 305,
+          "time_s": 17.058,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.275,
+          "peak_reserved_gb": 19.608,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.813
+        },
+        {
+          "id": "P06",
+          "title": "Russian Provincial Print Shop",
+          "seed": 306,
+          "time_s": 20.163,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.242,
+          "peak_reserved_gb": 19.803,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 13.16
+        },
+        {
+          "id": "P07",
+          "title": "Ocean Cartography Bakery",
+          "seed": 307,
+          "time_s": 16.903,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.274,
+          "peak_reserved_gb": 19.608,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.884
+        },
+        {
+          "id": "P08",
+          "title": "Long English Train Notice Wall",
+          "seed": 308,
+          "time_s": 16.857,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.274,
+          "peak_reserved_gb": 19.608,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.884
+        },
+        {
+          "id": "P09",
+          "title": "Orbital Botanical Courtroom",
+          "seed": 309,
+          "time_s": 19.664,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.276,
+          "peak_reserved_gb": 19.61,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.909
+        },
+        {
+          "id": "P10",
+          "title": "Byzantine Data Center Chapel",
+          "seed": 310,
+          "time_s": 16.731,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 19.276,
+          "peak_reserved_gb": 19.608,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.813
+        }
+      ]
+    },
+    "quant": {
+      "hardware": "RunPod NVIDIA H100 80GB HBM3 (H100 SXM)",
+      "mode": "Diffusers enable_model_cpu_offload()",
+      "cache_state": "warm local HF cache; no download time included",
+      "process_isolation": "single model per fresh Python process",
+      "base_resolution": 1440,
+      "aspect_ratio": "1:1",
+      "num_inference_steps": 20,
+      "guidance_scale": 5.0,
+      "dtype": "torch.bfloat16",
+      "kind": "quant",
+      "load": {
+        "load_time_s": 13.573,
+        "peak_allocated_gb": 12.582,
+        "peak_reserved_gb": 13.881,
+        "end_allocated_gb": 10.185,
+        "end_reserved_gb": 10.679
+      },
+      "summary": {
+        "cold_time_s": 21.473,
+        "cold_peak_allocated_gb": 15.479,
+        "cold_peak_reserved_gb": 19.126,
+        "warm_avg_time_s": 17.65,
+        "warm_median_time_s": 17.519,
+        "warm_avg_peak_allocated_gb": 15.48,
+        "warm_avg_peak_reserved_gb": 18.965,
+        "warm_max_peak_allocated_gb": 15.482,
+        "warm_max_peak_reserved_gb": 19.21
+      },
+      "prompts": [
+        {
+          "id": "P01",
+          "title": "Midnight Library Weather Station",
+          "seed": 301,
+          "time_s": 21.473,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.479,
+          "peak_reserved_gb": 19.126,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.865
+        },
+        {
+          "id": "P02",
+          "title": "Desert Observatory Treaty Room",
+          "seed": 302,
+          "time_s": 19.047,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.479,
+          "peak_reserved_gb": 19.126,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.865
+        },
+        {
+          "id": "P03",
+          "title": "Arctic Submarine Greenhouse",
+          "seed": 303,
+          "time_s": 17.133,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.479,
+          "peak_reserved_gb": 19.185,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.792
+        },
+        {
+          "id": "P04",
+          "title": "Long English Museum Labels",
+          "seed": 304,
+          "time_s": 17.73,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.479,
+          "peak_reserved_gb": 19.126,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.863
+        },
+        {
+          "id": "P05",
+          "title": "Tokyo Rooftop Repair Diner",
+          "seed": 305,
+          "time_s": 17.332,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.479,
+          "peak_reserved_gb": 19.21,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.815
+        },
+        {
+          "id": "P06",
+          "title": "Russian Provincial Print Shop",
+          "seed": 306,
+          "time_s": 17.519,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.482,
+          "peak_reserved_gb": 17.404,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 13.141
+        },
+        {
+          "id": "P07",
+          "title": "Ocean Cartography Bakery",
+          "seed": 307,
+          "time_s": 17.614,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.479,
+          "peak_reserved_gb": 19.126,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.865
+        },
+        {
+          "id": "P08",
+          "title": "Long English Train Notice Wall",
+          "seed": 308,
+          "time_s": 17.498,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.479,
+          "peak_reserved_gb": 19.126,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.865
+        },
+        {
+          "id": "P09",
+          "title": "Orbital Botanical Courtroom",
+          "seed": 309,
+          "time_s": 17.546,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.481,
+          "peak_reserved_gb": 19.172,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.888
+        },
+        {
+          "id": "P10",
+          "title": "Byzantine Data Center Chapel",
+          "seed": 310,
+          "time_s": 17.434,
+          "image_size": [
+            1440,
+            1440
+          ],
+          "peak_allocated_gb": 15.479,
+          "peak_reserved_gb": 19.21,
+          "end_allocated_gb": 10.221,
+          "end_reserved_gb": 10.813
+        }
+      ]
+    }
+  }
+}

model_index.json

@@ -0,0 +1,24 @@
+{
+  "_class_name": "LensPipeline",
+  "_diffusers_version": "0.37.1",
+  "scheduler": [
+    "diffusers",
+    "FlowMatchEulerDiscreteScheduler"
+  ],
+  "vae": [
+    "diffusers",
+    "AutoencoderKLFlux2"
+  ],
+  "text_encoder": [
+    "transformers",
+    "LensGptOssEncoder"
+  ],
+  "tokenizer": [
+    "transformers",
+    "PreTrainedTokenizerFast"
+  ],
+  "transformer": [
+    "diffusers",
+    "LensTransformer2DModel"
+  ]
+}

scheduler/scheduler_config.json

@@ -0,0 +1,18 @@
+{
+  "_class_name": "FlowMatchEulerDiscreteScheduler",
+  "_diffusers_version": "0.37.1",
+  "base_image_seq_len": 256,
+  "base_shift": 0.5,
+  "invert_sigmas": false,
+  "max_image_seq_len": 4096,
+  "max_shift": 1.15,
+  "num_train_timesteps": 1000,
+  "shift": 3.0,
+  "shift_terminal": null,
+  "stochastic_sampling": false,
+  "time_shift_type": "exponential",
+  "use_beta_sigmas": false,
+  "use_dynamic_shifting": true,
+  "use_exponential_sigmas": false,
+  "use_karras_sigmas": false
+}

sdnq_quantization_summary.json

@@ -0,0 +1,62 @@
+{
+  "source_model": "microsoft/Lens",
+  "target_model": "WaveCut/Lens-SDNQ-uint4-static",
+  "method": "SDNQ uint4 static",
+  "corrected_recipe": true,
+  "weights_dtype": "uint4",
+  "quantized_matmul_dtype": "int8",
+  "group_size": 0,
+  "use_quantized_matmul": true,
+  "dequantize_fp32": false,
+  "modules_to_not_convert_user": [
+    ".final_layer",
+    "pos_embed",
+    ".norm_out",
+    ".y_embedder",
+    ".context_embedder",
+    ".condition_embedder",
+    ".x_embedder",
+    ".vid_out",
+    ".emb_out",
+    ".img_in",
+    "patch_embed",
+    ".time_embed",
+    ".t_embedder",
+    "multi_modal_projector",
+    "patch_emb",
+    "norm",
+    ".img_out",
+    "patch_embedding",
+    "lm_head",
+    ".proj_out",
+    ".vid_in",
+    ".txt_in",
+    "wte",
+    "time_text_embed",
+    ".txt_out",
+    ".emb_in",
+    "*.img_mod.*",
+    "*.txt_mod.*"
+  ],
+  "root_cause_from_turbo_ablation": "Do not quantize transformer modulation linears (*.img_mod.* and *.txt_mod.*); all-linear UINT4 caused periodic grid artifacts and text degradation on Lens-Turbo.",
+  "transformer_load_time_s": 3.677,
+  "transformer_load_peak_allocated_gb": 8.359,
+  "transformer_load_peak_reserved_gb": 8.424,
+  "quantization_time_s": 0.313,
+  "quantization_peak_allocated_gb": 8.425,
+  "quantization_peak_reserved_gb": 8.485,
+  "base_transformer_tensor_storage_gb": 16.417,
+  "quant_transformer_tensor_storage_gb": 4.301,
+  "transformer_storage_reduction_percent": 73.8,
+  "base_transformer_repo_files_gb": 16.417,
+  "quant_transformer_repo_files_gb": 4.302,
+  "base_transformer_dtypes": {
+    "FLOAT32": 16416900608
+  },
+  "quant_transformer_dtypes": {
+    "BFLOAT16": 2942501632,
+    "UINT8": 1358954496
+  },
+  "base_transformer_tensors": 1264,
+  "quant_transformer_tensors": 2224
+}

text_encoder/config.json

@@ -0,0 +1,78 @@
+{
+  "architectures": [
+    "GptOssForCausalLM"
+  ],
+  "attention_bias": true,
+  "attention_dropout": 0.0,
+  "bos_token_id": null,
+  "dtype": "bfloat16",
+  "eos_token_id": 200002,
+  "experts_per_token": 4,
+  "head_dim": 64,
+  "hidden_act": "silu",
+  "hidden_size": 2880,
+  "initial_context_length": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 2880,
+  "layer_types": [
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 131072,
+  "model_type": "gpt_oss",
+  "num_attention_heads": 64,
+  "num_experts_per_tok": 4,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 8,
+  "num_local_experts": 32,
+  "output_router_logits": false,
+  "pad_token_id": 199999,
+  "quantization_config": {
+    "modules_to_not_convert": [
+      "model.layers.*.self_attn",
+      "model.layers.*.mlp.router",
+      "model.embed_tokens",
+      "lm_head"
+    ],
+    "quant_method": "mxfp4"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_parameters": {
+    "beta_fast": 32.0,
+    "beta_slow": 1.0,
+    "factor": 32.0,
+    "original_max_position_embeddings": 4096,
+    "rope_theta": 150000,
+    "rope_type": "yarn",
+    "truncate": false
+  },
+  "router_aux_loss_coef": 0.9,
+  "sliding_window": 128,
+  "swiglu_limit": 7.0,
+  "tie_word_embeddings": false,
+  "transformers_version": "5.8.0",
+  "use_cache": true,
+  "vocab_size": 201088
+}

text_encoder/generation_config.json

@@ -0,0 +1,11 @@
+{
+  "bos_token_id": 199998,
+  "do_sample": true,
+  "eos_token_id": [
+    200002,
+    199999,
+    200012
+  ],
+  "pad_token_id": 199999,
+  "transformers_version": "5.8.0"
+}

text_encoder/model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec5a256b3cdeea38db22f0988f9f156554fc9753f95857f5ecaea56048f43bd7
+size 4845744456

text_encoder/model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:914d5c38cd798a3393e348942e29c38497851b38c09df529ee162f4bd1f13b64
+size 4774186632

text_encoder/model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:931a994c03fc5554d6e343d95a991ae162502459eae6035cd9cc5ea773b598e3
+size 4154656824

text_encoder/model.safetensors.index.json

@@ -0,0 +1,467 @@
+{
+  "metadata": {
+    "total_parameters": 1804459584,
+    "total_size": 13774535808
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00001-of-00003.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.experts.down_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.experts.down_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.experts.down_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.experts.gate_up_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.experts.gate_up_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.experts.gate_up_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.router.bias": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.router.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.o_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.sinks": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.experts.down_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.experts.down_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.experts.down_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.experts.gate_up_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.experts.gate_up_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.experts.gate_up_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.router.bias": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.router.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.o_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.sinks": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.experts.gate_up_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.15.mlp.experts.gate_up_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.15.mlp.router.bias": "model-00003-of-00003.safetensors",
+    "model.layers.15.mlp.router.weight": "model-00003-of-00003.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.15.self_attn.o_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.15.self_attn.sinks": "model-00003-of-00003.safetensors",
+    "model.layers.15.self_attn.v_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.16.mlp.experts.down_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.16.mlp.experts.down_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.16.mlp.experts.down_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.16.mlp.experts.gate_up_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.16.mlp.experts.gate_up_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.16.mlp.experts.gate_up_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.16.mlp.router.bias": "model-00003-of-00003.safetensors",
+    "model.layers.16.mlp.router.weight": "model-00003-of-00003.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.16.self_attn.k_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.16.self_attn.o_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.16.self_attn.q_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.16.self_attn.sinks": "model-00003-of-00003.safetensors",
+    "model.layers.16.self_attn.v_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.17.mlp.experts.down_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.17.mlp.experts.down_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.17.mlp.experts.down_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.17.mlp.experts.gate_up_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.17.mlp.experts.gate_up_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.17.mlp.experts.gate_up_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.17.mlp.router.bias": "model-00003-of-00003.safetensors",
+    "model.layers.17.mlp.router.weight": "model-00003-of-00003.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.17.self_attn.o_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.17.self_attn.sinks": "model-00003-of-00003.safetensors",
+    "model.layers.17.self_attn.v_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.18.mlp.experts.down_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.18.mlp.experts.down_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.18.mlp.experts.down_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.18.mlp.experts.gate_up_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.18.mlp.experts.gate_up_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.18.mlp.experts.gate_up_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.18.mlp.router.bias": "model-00003-of-00003.safetensors",
+    "model.layers.18.mlp.router.weight": "model-00003-of-00003.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.18.self_attn.o_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.18.self_attn.sinks": "model-00003-of-00003.safetensors",
+    "model.layers.18.self_attn.v_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.19.mlp.experts.down_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.19.mlp.experts.down_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.19.mlp.experts.down_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.19.mlp.experts.gate_up_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.19.mlp.experts.gate_up_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.19.mlp.experts.gate_up_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.19.mlp.router.bias": "model-00003-of-00003.safetensors",
+    "model.layers.19.mlp.router.weight": "model-00003-of-00003.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.19.self_attn.o_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.19.self_attn.sinks": "model-00003-of-00003.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.experts.down_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.experts.down_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.experts.down_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.experts.gate_up_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.experts.gate_up_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.experts.gate_up_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.router.bias": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.router.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.o_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.sinks": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.20.mlp.experts.down_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.20.mlp.experts.down_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.20.mlp.experts.down_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.20.mlp.experts.gate_up_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.20.mlp.experts.gate_up_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.20.mlp.experts.gate_up_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.20.mlp.router.bias": "model-00003-of-00003.safetensors",
+    "model.layers.20.mlp.router.weight": "model-00003-of-00003.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.20.self_attn.o_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.20.self_attn.sinks": "model-00003-of-00003.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.21.mlp.experts.down_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.21.mlp.experts.down_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.21.mlp.experts.down_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.21.mlp.experts.gate_up_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.21.mlp.experts.gate_up_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.21.mlp.experts.gate_up_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.21.mlp.router.bias": "model-00003-of-00003.safetensors",
+    "model.layers.21.mlp.router.weight": "model-00003-of-00003.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.21.self_attn.k_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.21.self_attn.o_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.21.self_attn.sinks": "model-00003-of-00003.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.experts.down_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.experts.down_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.experts.down_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.experts.gate_up_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.experts.gate_up_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.experts.gate_up_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.router.bias": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.router.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.o_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.sinks": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.experts.down_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.experts.down_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.experts.down_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.experts.gate_up_proj_bias": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.experts.gate_up_proj_blocks": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.experts.gate_up_proj_scales": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.router.bias": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.router.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.o_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.sinks": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.experts.down_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.experts.down_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.experts.down_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.experts.gate_up_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.experts.gate_up_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.experts.gate_up_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.router.bias": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.router.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.o_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.sinks": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.experts.down_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.experts.down_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.experts.down_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.experts.gate_up_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.experts.gate_up_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.experts.gate_up_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.router.bias": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.router.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.o_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.sinks": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.experts.down_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.experts.down_proj_blocks": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.experts.down_proj_scales": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.experts.gate_up_proj_bias": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.5.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.5.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.5.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.5.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.5.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.5.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.5.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.5.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.6.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.6.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.6.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.6.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.6.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.6.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.6.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.6.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.6.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.6.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.6.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.7.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.7.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.7.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.7.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.7.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.7.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.7.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.7.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.7.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.7.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.7.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.8.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.8.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.8.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.8.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.8.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.8.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.8.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.8.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.8.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.8.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.8.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.9.mlp.experts.down_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.9.mlp.experts.down_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.9.mlp.experts.down_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.9.mlp.experts.gate_up_proj_bias": "model-00002-of-00003.safetensors",
+    "model.layers.9.mlp.experts.gate_up_proj_blocks": "model-00002-of-00003.safetensors",
+    "model.layers.9.mlp.experts.gate_up_proj_scales": "model-00002-of-00003.safetensors",
+    "model.layers.9.mlp.router.bias": "model-00002-of-00003.safetensors",
+    "model.layers.9.mlp.router.weight": "model-00002-of-00003.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.9.self_attn.k_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.9.self_attn.o_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.9.self_attn.q_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.9.self_attn.sinks": "model-00002-of-00003.safetensors",
+    "model.layers.9.self_attn.v_proj.bias": "model-00002-of-00003.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.norm.weight": "model-00003-of-00003.safetensors"
+  }
+}

tokenizer/chat_template.jinja

@@ -0,0 +1,331 @@
+{#-
+  In addition to the normal inputs of `messages` and `tools`, this template also accepts the
+  following kwargs:
+  - "builtin_tools": A list, can contain "browser" and/or "python".
+  - "model_identity": A string that optionally describes the model identity.
+  - "reasoning_effort": A string that describes the reasoning effort, defaults to "medium".
+ #}
+
+{#- Tool Definition Rendering ============================================== #}
+{%- macro render_typescript_type(param_spec, required_params, is_nullable=false) -%}
+    {%- if param_spec.type == "array" -%}
+        {%- if param_spec['items'] -%}
+            {%- if param_spec['items']['type'] == "string" -%}
+                {{- "string[]" }}
+            {%- elif param_spec['items']['type'] == "number" -%}
+                {{- "number[]" }}
+            {%- elif param_spec['items']['type'] == "integer" -%}
+                {{- "number[]" }}
+            {%- elif param_spec['items']['type'] == "boolean" -%}
+                {{- "boolean[]" }}
+            {%- else -%}
+                {%- set inner_type = render_typescript_type(param_spec['items'], required_params) -%}
+                {%- if inner_type == "object | object" or inner_type|length > 50 -%}
+                    {{- "any[]" }}
+                {%- else -%}
+                    {{- inner_type + "[]" }}
+                {%- endif -%}
+            {%- endif -%}
+            {%- if param_spec.nullable -%}
+                {{- " | null" }}
+            {%- endif -%}
+        {%- else -%}
+            {{- "any[]" }}
+            {%- if param_spec.nullable -%}
+                {{- " | null" }}
+            {%- endif -%}
+        {%- endif -%}
+    {%- elif param_spec.type is defined and param_spec.type is iterable and param_spec.type is not string and param_spec.type is not mapping and param_spec.type[0] is defined -%}
+        {#- Handle array of types like ["object", "object"] from Union[dict, list] #}
+        {%- if param_spec.type | length > 1 -%}
+            {{- param_spec.type | join(" | ") }}
+        {%- else -%}
+            {{- param_spec.type[0] }}
+        {%- endif -%}
+    {%- elif param_spec.oneOf -%}
+        {#- Handle oneOf schemas - check for complex unions and fallback to any #}
+        {%- set has_object_variants = false -%}
+        {%- for variant in param_spec.oneOf -%}
+            {%- if variant.type == "object" -%}
+                {%- set has_object_variants = true -%}
+            {%- endif -%}
+        {%- endfor -%}
+        {%- if has_object_variants and param_spec.oneOf|length > 1 -%}
+            {{- "any" }}
+        {%- else -%}
+            {%- for variant in param_spec.oneOf -%}
+                {{- render_typescript_type(variant, required_params) -}}
+                {%- if variant.description %}
+                    {{- "// " + variant.description }}
+                {%- endif -%}
+                {%- if variant.default is defined %}
+                    {{ "// default: " + variant.default|tojson }}
+                {%- endif -%}
+                {%- if not loop.last %}
+                    {{- " | " }}
+                {% endif -%}
+            {%- endfor -%}
+        {%- endif -%}
+    {%- elif param_spec.type == "string" -%}
+        {%- if param_spec.enum -%}
+            {{- '"' + param_spec.enum|join('" | "') + '"' -}}
+        {%- else -%}
+            {{- "string" }}
+            {%- if param_spec.nullable %}
+                {{- " | null" }}
+            {%- endif -%}
+        {%- endif -%}
+    {%- elif param_spec.type == "number" -%}
+        {{- "number" }}
+    {%- elif param_spec.type == "integer" -%}
+        {{- "number" }}
+    {%- elif param_spec.type == "boolean" -%}
+        {{- "boolean" }}
+
+    {%- elif param_spec.type == "object" -%}
+        {%- if param_spec.properties -%}
+            {{- "{\n" }}
+            {%- for prop_name, prop_spec in param_spec.properties.items() -%}
+                {{- prop_name -}}
+                {%- if prop_name not in (param_spec.required or []) -%}
+                    {{- "?" }}
+                {%- endif -%}
+                {{- ": " }}
+                {{ render_typescript_type(prop_spec, param_spec.required or []) }}
+                {%- if not loop.last -%}
+                    {{-", " }}
+                {%- endif -%}
+            {%- endfor -%}
+            {{- "}" }}
+        {%- else -%}
+            {{- "object" }}
+        {%- endif -%}
+    {%- else -%}
+        {{- "any" }}
+    {%- endif -%}
+{%- endmacro -%}
+
+{%- macro render_tool_namespace(namespace_name, tools) -%}
+    {{- "## " + namespace_name + "\n\n" }}
+    {{- "namespace " + namespace_name + " {\n\n" }}
+    {%- for tool in tools %}
+        {%- set tool = tool.function %}
+        {{- "// " + tool.description + "\n" }}
+        {{- "type "+ tool.name + " = " }}
+        {%- if tool.parameters and tool.parameters.properties %}
+            {{- "(_: {\n" }}
+            {%- for param_name, param_spec in tool.parameters.properties.items() %}
+                {%- if param_spec.description %}
+                    {{- "// " + param_spec.description + "\n" }}
+                {%- endif %}
+                {{- param_name }}
+                {%- if param_name not in (tool.parameters.required or []) -%}
+                    {{- "?" }}
+                {%- endif -%}
+                {{- ": " }}
+                {{- render_typescript_type(param_spec, tool.parameters.required or []) }}
+                {%- if param_spec.default is defined -%}
+                    {%- if param_spec.enum %}
+                        {{- ", // default: " + param_spec.default }}
+                    {%- elif param_spec.oneOf %}
+                        {{- "// default: " + param_spec.default }}
+                    {%- else %}
+                        {{- ", // default: " + param_spec.default|tojson }}
+                    {%- endif -%}
+                {%- endif -%}
+                {%- if not loop.last %}
+                    {{- ",\n" }}
+                {%- else %}
+                    {{- ",\n" }}
+                {%- endif -%}
+            {%- endfor %}
+            {{- "}) => any;\n\n" }}
+        {%- else -%}
+            {{- "() => any;\n\n" }}
+        {%- endif -%}
+    {%- endfor %}
+    {{- "} // namespace " + namespace_name }}
+{%- endmacro -%}
+
+{%- macro render_builtin_tools(browser_tool, python_tool) -%}
+    {%- if browser_tool %}
+        {{- "## browser\n\n" }}
+        {{- "// Tool for browsing.\n" }}
+        {{- "// The `cursor` appears in brackets before each browsing display: `[{cursor}]`.\n" }}
+        {{- "// Cite information from the tool using the following format:\n" }}
+        {{- "// `【{cursor}†L{line_start}(-L{line_end})?】`, for example: `【6†L9-L11】` or `【8†L3】`.\n" }}
+        {{- "// Do not quote more than 10 words directly from the tool output.\n" }}
+        {{- "// sources=web (default: web)\n" }}
+        {{- "namespace browser {\n\n" }}
+        {{- "// Searches for information related to `query` and displays `topn` results.\n" }}
+        {{- "type search = (_: {\n" }}
+        {{- "query: string,\n" }}
+        {{- "topn?: number, // default: 10\n" }}
+        {{- "source?: string,\n" }}
+        {{- "}) => any;\n\n" }}
+        {{- "// Opens the link `id` from the page indicated by `cursor` starting at line number `loc`, showing `num_lines` lines.\n" }}
+        {{- "// Valid link ids are displayed with the formatting: `【{id}†.*】`.\n" }}
+        {{- "// If `cursor` is not provided, the most recent page is implied.\n" }}
+        {{- "// If `id` is a string, it is treated as a fully qualified URL associated with `source`.\n" }}
+        {{- "// If `loc` is not provided, the viewport will be positioned at the beginning of the document or centered on the most relevant passage, if available.\n" }}
+        {{- "// Use this function without `id` to scroll to a new location of an opened page.\n" }}
+        {{- "type open = (_: {\n" }}
+        {{- "id?: number | string, // default: -1\n" }}
+        {{- "cursor?: number, // default: -1\n" }}
+        {{- "loc?: number, // default: -1\n" }}
+        {{- "num_lines?: number, // default: -1\n" }}
+        {{- "view_source?: boolean, // default: false\n" }}
+        {{- "source?: string,\n" }}
+        {{- "}) => any;\n\n" }}
+        {{- "// Finds exact matches of `pattern` in the current page, or the page given by `cursor`.\n" }}
+        {{- "type find = (_: {\n" }}
+        {{- "pattern: string,\n" }}
+        {{- "cursor?: number, // default: -1\n" }}
+        {{- "}) => any;\n\n" }}
+        {{- "} // namespace browser\n\n" }}
+    {%- endif -%}
+
+    {%- if python_tool %}
+        {{- "## python\n\n" }}
+        {{- "Use this tool to execute Python code in your chain of thought. The code will not be shown to the user. This tool should be used for internal reasoning, but not for code that is intended to be visible to the user (e.g. when creating plots, tables, or files).\n\n" }}
+        {{- "When you send a message containing Python code to python, it will be executed in a stateful Jupyter notebook environment. python will respond with the output of the execution or time out after 120.0 seconds. The drive at '/mnt/data' can be used to save and persist user files. Internet access for this session is UNKNOWN. Depends on the cluster.\n\n" }}
+    {%- endif -%}
+{%- endmacro -%}
+
+{#- System Message Construction ============================================ #}
+{%- macro build_system_message() -%}
+    {%- if model_identity is not defined %}
+        {%- set model_identity = "You are ChatGPT, a large language model trained by OpenAI." %}
+    {%- endif %}
+    {{- model_identity + "\n" }}
+    {{- "Knowledge cutoff: 2024-06\n" }}
+    {{- "Current date: " + strftime_now("%Y-%m-%d") + "\n\n" }}
+    {%- if reasoning_effort is not defined %}
+        {%- set reasoning_effort = "medium" %}
+    {%- endif %}
+    {{- "Reasoning: " + reasoning_effort + "\n\n" }}
+    {%- if builtin_tools %}
+        {{- "# Tools\n\n" }}
+        {%- set available_builtin_tools = namespace(browser=false, python=false) %}
+        {%- for tool in builtin_tools %}
+            {%- if tool == "browser" %}
+                {%- set available_builtin_tools.browser = true %}
+            {%- elif tool == "python" %}
+                {%- set available_builtin_tools.python = true %}
+            {%- endif %}
+        {%- endfor %}
+        {{- render_builtin_tools(available_builtin_tools.browser, available_builtin_tools.python) }}
+    {%- endif -%}
+    {{- "# Valid channels: analysis, commentary, final. Channel must be included for every message." }}
+    {%- if tools -%}
+        {{- "\nCalls to these tools must go to the commentary channel: 'functions'." }}
+    {%- endif -%}
+{%- endmacro -%}
+
+{#- Main Template Logic ================================================= #}
+{#- Set defaults #}
+
+{#- Render system message #}
+{{- "<|start|>system<|message|>" }}
+{{- build_system_message() }}
+{{- "<|end|>" }}
+
+{#- Extract developer message #}
+{%- if messages[0].role == "developer" or messages[0].role == "system" %}
+    {%- set developer_message = messages[0].content %}
+    {%- set loop_messages = messages[1:] %}
+{%- else %}
+    {%- set developer_message = "" %}
+    {%- set loop_messages = messages %}
+{%- endif %}
+
+{#- Render developer message #}
+{%- if developer_message or tools %}
+    {{- "<|start|>developer<|message|>" }}
+    {%- if developer_message %}
+        {{- "# Instructions\n\n" }}
+        {{- developer_message }}
+        {{- "\n\n" }}
+    {%- endif %}
+    {%- if tools -%}
+        {{- "# Tools\n\n" }}
+        {{- render_tool_namespace("functions", tools) }}
+    {%- endif -%}
+    {{- "<|end|>" }}
+{%- endif %}
+
+{#- Render messages #}
+{%- set last_tool_call = namespace(name=none) %}
+{%- for message in loop_messages -%}
+    {#- At this point only assistant/user/tool messages should remain #}
+    {%- if message.role == 'assistant' -%}
+        {#- Checks to ensure the messages are being passed in the format we expect #}
+        {%- if "content" in message %}
+            {%- if "<|channel|>analysis<|message|>" in message.content or "<|channel|>final<|message|>" in message.content %}
+                {{- raise_exception("You have passed a message containing <|channel|> tags in the content field. Instead of doing this, you should pass analysis messages (the string between '<|message|>' and '<|end|>') in the 'thinking' field, and final messages (the string between '<|message|>' and '<|end|>') in the 'content' field.") }}
+            {%- endif %}
+        {%- endif %}
+        {%- if "thinking" in message %}
+            {%- if "<|channel|>analysis<|message|>" in message.thinking or "<|channel|>final<|message|>" in message.thinking %}
+                {{- raise_exception("You have passed a message containing <|channel|> tags in the thinking field. Instead of doing this, you should pass analysis messages (the string between '<|message|>' and '<|end|>') in the 'thinking' field, and final messages (the string between '<|message|>' and '<|end|>') in the 'content' field.") }}
+            {%- endif %}
+        {%- endif %}
+        {%- if "tool_calls" in message %}
+            {#- We need very careful handling here - we want to drop the tool call analysis message if the model #}
+            {#- has output a later <|final|> message, but otherwise we want to retain it. This is the only case #}
+            {#- when we render CoT/analysis messages in inference. #}
+            {%- set future_final_message = namespace(found=false) %}
+            {%- for future_message in loop_messages[loop.index:] %}
+                {%- if future_message.role == 'assistant' and "tool_calls" not in future_message %}
+                    {%- set future_final_message.found = true %}
+                {%- endif %}
+            {%- endfor %}
+            {#- We assume max 1 tool call per message, and so we infer the tool call name #}
+            {#- in "tool" messages from the most recent assistant tool call name #}
+            {%- set tool_call = message.tool_calls[0] %}
+            {%- if tool_call.function %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {%- if message.content and message.thinking %}
+                {{- raise_exception("Cannot pass both content and thinking in an assistant message with tool calls! Put the analysis message in one or the other, but not both.") }}
+            {%- elif message.content and not future_final_message.found %}
+                {{- "<|start|>assistant<|channel|>analysis<|message|>" + message.content + "<|end|>" }}
+            {%- elif message.thinking and not future_final_message.found %}
+                {{- "<|start|>assistant<|channel|>analysis<|message|>" + message.thinking + "<|end|>" }}
+            {%- endif %}
+            {{- "<|start|>assistant to=" }}
+            {{- "functions." + tool_call.name + "<|channel|>commentary " }}
+            {{- (tool_call.content_type if tool_call.content_type is defined else "json") + "<|message|>" }}
+            {{- tool_call.arguments|tojson }}
+            {{- "<|call|>" }}
+            {%- set last_tool_call.name = tool_call.name %}
+        {%- elif loop.last and not add_generation_prompt %}
+            {#- Only render the CoT if the final turn is an assistant turn and add_generation_prompt is false #}
+            {#- This is a situation that should only occur in training, never in inference. #}
+            {%- if "thinking" in message %}
+                {{- "<|start|>assistant<|channel|>analysis<|message|>" + message.thinking + "<|end|>" }}
+            {%- endif %}
+            {#- <|return|> indicates the end of generation, but <|end|> does not #}
+            {#- <|return|> should never be an input to the model, but we include it as the final token #}
+            {#- when training, so the model learns to emit it. #}
+            {{- "<|start|>assistant<|channel|>final<|message|>" + message.content + "<|return|>" }}
+        {%- else %}
+            {#- CoT is dropped during all previous turns, so we never render it for inference #}
+            {{- "<|start|>assistant<|channel|>final<|message|>" + message.content + "<|end|>" }}
+            {%- set last_tool_call.name = none %}
+        {%- endif %}
+    {%- elif message.role == 'tool' -%}
+        {%- if last_tool_call.name is none %}
+            {{- raise_exception("Message has tool role, but there was no previous assistant message with a tool call!") }}
+        {%- endif %}
+        {{- "<|start|>functions." + last_tool_call.name }}
+        {{- " to=assistant<|channel|>commentary<|message|>" + message.content|tojson + "<|end|>" }}
+    {%- elif message.role == 'user' -%}
+        {{- "<|start|>user<|message|>" + message.content + "<|end|>" }}
+    {%- endif -%}
+{%- endfor -%}
+
+{#- Generation prompt #}
+{%- if add_generation_prompt -%}
+<|start|>assistant
+{%- endif -%}

tokenizer/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0614fe83cadab421296e664e1f48f4261fa8fef6e03e63bb75c20f38e37d07d3
+size 27868174

tokenizer/tokenizer_config.json

@@ -0,0 +1,15 @@
+{
+  "backend": "tokenizers",
+  "bos_token": "<|startoftext|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|return|>",
+  "is_local": false,
+  "local_files_only": false,
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "TokenizersBackend"
+}

transformer/config.json

@@ -0,0 +1,476 @@
+{
+  "_class_name": "LensTransformer2DModel",
+  "_diffusers_version": "0.38.0",
+  "_name_or_path": "microsoft/Lens",
+  "attention_head_dim": 64,
+  "axes_dims_rope": [
+    8,
+    28,
+    28
+  ],
+  "enc_hidden_dim": 2880,
+  "gate_mlp": true,
+  "in_channels": 128,
+  "inner_dim": 1536,
+  "multi_layer_encoder_feature": true,
+  "num_attention_heads": 24,
+  "num_layers": 48,
+  "out_channels": 32,
+  "patch_size": 2,
+  "quantization_config": {
+    "add_skip_keys": false,
+    "dequantize_fp32": false,
+    "dynamic_loss_threshold": null,
+    "group_size": 0,
+    "hadamard_group_size": 128,
+    "is_integer": true,
+    "is_training": false,
+    "modules_dtype_dict": {},
+    "modules_quant_config": {},
+    "modules_to_not_convert": [
+      ".emb_in",
+      ".emb_out",
+      "norm",
+      ".y_embedder",
+      "patch_emb",
+      "lm_head",
+      ".condition_embedder",
+      ".norm_out",
+      ".t_embedder",
+      "patch_embedding",
+      ".vid_in",
+      ".txt_in",
+      ".txt_out",
+      ".time_embed",
+      "*.img_mod.*",
+      ".proj_out",
+      ".img_out",
+      "*.txt_mod.*",
+      ".context_embedder",
+      ".vid_out",
+      "time_text_embed",
+      "patch_embed",
+      "wte",
+      ".img_in",
+      ".final_layer",
+      ".x_embedder",
+      "multi_modal_projector",
+      "pos_embed",
+      "txt_norm.0.weight",
+      "txt_norm.1.weight",
+      "txt_norm.2.weight",
+      "txt_norm.3.weight",
+      "transformer_blocks.0.attn.norm_q.weight",
+      "transformer_blocks.0.attn.norm_k.weight",
+      "transformer_blocks.0.attn.norm_added_q.weight",
+      "transformer_blocks.0.attn.norm_added_k.weight",
+      "transformer_blocks.0.img_norm1.weight",
+      "transformer_blocks.0.img_norm2.weight",
+      "transformer_blocks.0.txt_norm1.weight",
+      "transformer_blocks.0.txt_norm2.weight",
+      "transformer_blocks.1.attn.norm_q.weight",
+      "transformer_blocks.1.attn.norm_k.weight",
+      "transformer_blocks.1.attn.norm_added_q.weight",
+      "transformer_blocks.1.attn.norm_added_k.weight",
+      "transformer_blocks.1.img_norm1.weight",
+      "transformer_blocks.1.img_norm2.weight",
+      "transformer_blocks.1.txt_norm1.weight",
+      "transformer_blocks.1.txt_norm2.weight",
+      "transformer_blocks.2.attn.norm_q.weight",
+      "transformer_blocks.2.attn.norm_k.weight",
+      "transformer_blocks.2.attn.norm_added_q.weight",
+      "transformer_blocks.2.attn.norm_added_k.weight",
+      "transformer_blocks.2.img_norm1.weight",
+      "transformer_blocks.2.img_norm2.weight",
+      "transformer_blocks.2.txt_norm1.weight",
+      "transformer_blocks.2.txt_norm2.weight",
+      "transformer_blocks.3.attn.norm_q.weight",
+      "transformer_blocks.3.attn.norm_k.weight",
+      "transformer_blocks.3.attn.norm_added_q.weight",
+      "transformer_blocks.3.attn.norm_added_k.weight",
+      "transformer_blocks.3.img_norm1.weight",
+      "transformer_blocks.3.img_norm2.weight",
+      "transformer_blocks.3.txt_norm1.weight",
+      "transformer_blocks.3.txt_norm2.weight",
+      "transformer_blocks.4.attn.norm_q.weight",
+      "transformer_blocks.4.attn.norm_k.weight",
+      "transformer_blocks.4.attn.norm_added_q.weight",
+      "transformer_blocks.4.attn.norm_added_k.weight",
+      "transformer_blocks.4.img_norm1.weight",
+      "transformer_blocks.4.img_norm2.weight",
+      "transformer_blocks.4.txt_norm1.weight",
+      "transformer_blocks.4.txt_norm2.weight",
+      "transformer_blocks.5.attn.norm_q.weight",
+      "transformer_blocks.5.attn.norm_k.weight",
+      "transformer_blocks.5.attn.norm_added_q.weight",
+      "transformer_blocks.5.attn.norm_added_k.weight",
+      "transformer_blocks.5.img_norm1.weight",
+      "transformer_blocks.5.img_norm2.weight",
+      "transformer_blocks.5.txt_norm1.weight",
+      "transformer_blocks.5.txt_norm2.weight",
+      "transformer_blocks.6.attn.norm_q.weight",
+      "transformer_blocks.6.attn.norm_k.weight",
+      "transformer_blocks.6.attn.norm_added_q.weight",
+      "transformer_blocks.6.attn.norm_added_k.weight",
+      "transformer_blocks.6.img_norm1.weight",
+      "transformer_blocks.6.img_norm2.weight",
+      "transformer_blocks.6.txt_norm1.weight",
+      "transformer_blocks.6.txt_norm2.weight",
+      "transformer_blocks.7.attn.norm_q.weight",
+      "transformer_blocks.7.attn.norm_k.weight",
+      "transformer_blocks.7.attn.norm_added_q.weight",
+      "transformer_blocks.7.attn.norm_added_k.weight",
+      "transformer_blocks.7.img_norm1.weight",
+      "transformer_blocks.7.img_norm2.weight",
+      "transformer_blocks.7.txt_norm1.weight",
+      "transformer_blocks.7.txt_norm2.weight",
+      "transformer_blocks.8.attn.norm_q.weight",
+      "transformer_blocks.8.attn.norm_k.weight",
+      "transformer_blocks.8.attn.norm_added_q.weight",
+      "transformer_blocks.8.attn.norm_added_k.weight",
+      "transformer_blocks.8.img_norm1.weight",
+      "transformer_blocks.8.img_norm2.weight",
+      "transformer_blocks.8.txt_norm1.weight",
+      "transformer_blocks.8.txt_norm2.weight",
+      "transformer_blocks.9.attn.norm_q.weight",
+      "transformer_blocks.9.attn.norm_k.weight",
+      "transformer_blocks.9.attn.norm_added_q.weight",
+      "transformer_blocks.9.attn.norm_added_k.weight",
+      "transformer_blocks.9.img_norm1.weight",
+      "transformer_blocks.9.img_norm2.weight",
+      "transformer_blocks.9.txt_norm1.weight",
+      "transformer_blocks.9.txt_norm2.weight",
+      "transformer_blocks.10.attn.norm_q.weight",
+      "transformer_blocks.10.attn.norm_k.weight",
+      "transformer_blocks.10.attn.norm_added_q.weight",
+      "transformer_blocks.10.attn.norm_added_k.weight",
+      "transformer_blocks.10.img_norm1.weight",
+      "transformer_blocks.10.img_norm2.weight",
+      "transformer_blocks.10.txt_norm1.weight",
+      "transformer_blocks.10.txt_norm2.weight",
+      "transformer_blocks.11.attn.norm_q.weight",
+      "transformer_blocks.11.attn.norm_k.weight",
+      "transformer_blocks.11.attn.norm_added_q.weight",
+      "transformer_blocks.11.attn.norm_added_k.weight",
+      "transformer_blocks.11.img_norm1.weight",
+      "transformer_blocks.11.img_norm2.weight",
+      "transformer_blocks.11.txt_norm1.weight",
+      "transformer_blocks.11.txt_norm2.weight",
+      "transformer_blocks.12.attn.norm_q.weight",
+      "transformer_blocks.12.attn.norm_k.weight",
+      "transformer_blocks.12.attn.norm_added_q.weight",
+      "transformer_blocks.12.attn.norm_added_k.weight",
+      "transformer_blocks.12.img_norm1.weight",
+      "transformer_blocks.12.img_norm2.weight",
+      "transformer_blocks.12.txt_norm1.weight",
+      "transformer_blocks.12.txt_norm2.weight",
+      "transformer_blocks.13.attn.norm_q.weight",
+      "transformer_blocks.13.attn.norm_k.weight",
+      "transformer_blocks.13.attn.norm_added_q.weight",
+      "transformer_blocks.13.attn.norm_added_k.weight",
+      "transformer_blocks.13.img_norm1.weight",
+      "transformer_blocks.13.img_norm2.weight",
+      "transformer_blocks.13.txt_norm1.weight",
+      "transformer_blocks.13.txt_norm2.weight",
+      "transformer_blocks.14.attn.norm_q.weight",
+      "transformer_blocks.14.attn.norm_k.weight",
+      "transformer_blocks.14.attn.norm_added_q.weight",
+      "transformer_blocks.14.attn.norm_added_k.weight",
+      "transformer_blocks.14.img_norm1.weight",
+      "transformer_blocks.14.img_norm2.weight",
+      "transformer_blocks.14.txt_norm1.weight",
+      "transformer_blocks.14.txt_norm2.weight",
+      "transformer_blocks.15.attn.norm_q.weight",
+      "transformer_blocks.15.attn.norm_k.weight",
+      "transformer_blocks.15.attn.norm_added_q.weight",
+      "transformer_blocks.15.attn.norm_added_k.weight",
+      "transformer_blocks.15.img_norm1.weight",
+      "transformer_blocks.15.img_norm2.weight",
+      "transformer_blocks.15.txt_norm1.weight",
+      "transformer_blocks.15.txt_norm2.weight",
+      "transformer_blocks.16.attn.norm_q.weight",
+      "transformer_blocks.16.attn.norm_k.weight",
+      "transformer_blocks.16.attn.norm_added_q.weight",
+      "transformer_blocks.16.attn.norm_added_k.weight",
+      "transformer_blocks.16.img_norm1.weight",
+      "transformer_blocks.16.img_norm2.weight",
+      "transformer_blocks.16.txt_norm1.weight",
+      "transformer_blocks.16.txt_norm2.weight",
+      "transformer_blocks.17.attn.norm_q.weight",
+      "transformer_blocks.17.attn.norm_k.weight",
+      "transformer_blocks.17.attn.norm_added_q.weight",
+      "transformer_blocks.17.attn.norm_added_k.weight",
+      "transformer_blocks.17.img_norm1.weight",
+      "transformer_blocks.17.img_norm2.weight",
+      "transformer_blocks.17.txt_norm1.weight",
+      "transformer_blocks.17.txt_norm2.weight",
+      "transformer_blocks.18.attn.norm_q.weight",
+      "transformer_blocks.18.attn.norm_k.weight",
+      "transformer_blocks.18.attn.norm_added_q.weight",
+      "transformer_blocks.18.attn.norm_added_k.weight",
+      "transformer_blocks.18.img_norm1.weight",
+      "transformer_blocks.18.img_norm2.weight",
+      "transformer_blocks.18.txt_norm1.weight",
+      "transformer_blocks.18.txt_norm2.weight",
+      "transformer_blocks.19.attn.norm_q.weight",
+      "transformer_blocks.19.attn.norm_k.weight",
+      "transformer_blocks.19.attn.norm_added_q.weight",
+      "transformer_blocks.19.attn.norm_added_k.weight",
+      "transformer_blocks.19.img_norm1.weight",
+      "transformer_blocks.19.img_norm2.weight",
+      "transformer_blocks.19.txt_norm1.weight",
+      "transformer_blocks.19.txt_norm2.weight",
+      "transformer_blocks.20.attn.norm_q.weight",
+      "transformer_blocks.20.attn.norm_k.weight",
+      "transformer_blocks.20.attn.norm_added_q.weight",
+      "transformer_blocks.20.attn.norm_added_k.weight",
+      "transformer_blocks.20.img_norm1.weight",
+      "transformer_blocks.20.img_norm2.weight",
+      "transformer_blocks.20.txt_norm1.weight",
+      "transformer_blocks.20.txt_norm2.weight",
+      "transformer_blocks.21.attn.norm_q.weight",
+      "transformer_blocks.21.attn.norm_k.weight",
+      "transformer_blocks.21.attn.norm_added_q.weight",
+      "transformer_blocks.21.attn.norm_added_k.weight",
+      "transformer_blocks.21.img_norm1.weight",
+      "transformer_blocks.21.img_norm2.weight",
+      "transformer_blocks.21.txt_norm1.weight",
+      "transformer_blocks.21.txt_norm2.weight",
+      "transformer_blocks.22.attn.norm_q.weight",
+      "transformer_blocks.22.attn.norm_k.weight",
+      "transformer_blocks.22.attn.norm_added_q.weight",
+      "transformer_blocks.22.attn.norm_added_k.weight",
+      "transformer_blocks.22.img_norm1.weight",
+      "transformer_blocks.22.img_norm2.weight",
+      "transformer_blocks.22.txt_norm1.weight",
+      "transformer_blocks.22.txt_norm2.weight",
+      "transformer_blocks.23.attn.norm_q.weight",
+      "transformer_blocks.23.attn.norm_k.weight",
+      "transformer_blocks.23.attn.norm_added_q.weight",
+      "transformer_blocks.23.attn.norm_added_k.weight",
+      "transformer_blocks.23.img_norm1.weight",
+      "transformer_blocks.23.img_norm2.weight",
+      "transformer_blocks.23.txt_norm1.weight",
+      "transformer_blocks.23.txt_norm2.weight",
+      "transformer_blocks.24.attn.norm_q.weight",
+      "transformer_blocks.24.attn.norm_k.weight",
+      "transformer_blocks.24.attn.norm_added_q.weight",
+      "transformer_blocks.24.attn.norm_added_k.weight",
+      "transformer_blocks.24.img_norm1.weight",
+      "transformer_blocks.24.img_norm2.weight",
+      "transformer_blocks.24.txt_norm1.weight",
+      "transformer_blocks.24.txt_norm2.weight",
+      "transformer_blocks.25.attn.norm_q.weight",
+      "transformer_blocks.25.attn.norm_k.weight",
+      "transformer_blocks.25.attn.norm_added_q.weight",
+      "transformer_blocks.25.attn.norm_added_k.weight",
+      "transformer_blocks.25.img_norm1.weight",
+      "transformer_blocks.25.img_norm2.weight",
+      "transformer_blocks.25.txt_norm1.weight",
+      "transformer_blocks.25.txt_norm2.weight",
+      "transformer_blocks.26.attn.norm_q.weight",
+      "transformer_blocks.26.attn.norm_k.weight",
+      "transformer_blocks.26.attn.norm_added_q.weight",
+      "transformer_blocks.26.attn.norm_added_k.weight",
+      "transformer_blocks.26.img_norm1.weight",
+      "transformer_blocks.26.img_norm2.weight",
+      "transformer_blocks.26.txt_norm1.weight",
+      "transformer_blocks.26.txt_norm2.weight",
+      "transformer_blocks.27.attn.norm_q.weight",
+      "transformer_blocks.27.attn.norm_k.weight",
+      "transformer_blocks.27.attn.norm_added_q.weight",
+      "transformer_blocks.27.attn.norm_added_k.weight",
+      "transformer_blocks.27.img_norm1.weight",
+      "transformer_blocks.27.img_norm2.weight",
+      "transformer_blocks.27.txt_norm1.weight",
+      "transformer_blocks.27.txt_norm2.weight",
+      "transformer_blocks.28.attn.norm_q.weight",
+      "transformer_blocks.28.attn.norm_k.weight",
+      "transformer_blocks.28.attn.norm_added_q.weight",
+      "transformer_blocks.28.attn.norm_added_k.weight",
+      "transformer_blocks.28.img_norm1.weight",
+      "transformer_blocks.28.img_norm2.weight",
+      "transformer_blocks.28.txt_norm1.weight",
+      "transformer_blocks.28.txt_norm2.weight",
+      "transformer_blocks.29.attn.norm_q.weight",
+      "transformer_blocks.29.attn.norm_k.weight",
+      "transformer_blocks.29.attn.norm_added_q.weight",
+      "transformer_blocks.29.attn.norm_added_k.weight",
+      "transformer_blocks.29.img_norm1.weight",
+      "transformer_blocks.29.img_norm2.weight",
+      "transformer_blocks.29.txt_norm1.weight",
+      "transformer_blocks.29.txt_norm2.weight",
+      "transformer_blocks.30.attn.norm_q.weight",
+      "transformer_blocks.30.attn.norm_k.weight",
+      "transformer_blocks.30.attn.norm_added_q.weight",
+      "transformer_blocks.30.attn.norm_added_k.weight",
+      "transformer_blocks.30.img_norm1.weight",
+      "transformer_blocks.30.img_norm2.weight",
+      "transformer_blocks.30.txt_norm1.weight",
+      "transformer_blocks.30.txt_norm2.weight",
+      "transformer_blocks.31.attn.norm_q.weight",
+      "transformer_blocks.31.attn.norm_k.weight",
+      "transformer_blocks.31.attn.norm_added_q.weight",
+      "transformer_blocks.31.attn.norm_added_k.weight",
+      "transformer_blocks.31.img_norm1.weight",
+      "transformer_blocks.31.img_norm2.weight",
+      "transformer_blocks.31.txt_norm1.weight",
+      "transformer_blocks.31.txt_norm2.weight",
+      "transformer_blocks.32.attn.norm_q.weight",
+      "transformer_blocks.32.attn.norm_k.weight",
+      "transformer_blocks.32.attn.norm_added_q.weight",
+      "transformer_blocks.32.attn.norm_added_k.weight",
+      "transformer_blocks.32.img_norm1.weight",
+      "transformer_blocks.32.img_norm2.weight",
+      "transformer_blocks.32.txt_norm1.weight",
+      "transformer_blocks.32.txt_norm2.weight",
+      "transformer_blocks.33.attn.norm_q.weight",
+      "transformer_blocks.33.attn.norm_k.weight",
+      "transformer_blocks.33.attn.norm_added_q.weight",
+      "transformer_blocks.33.attn.norm_added_k.weight",
+      "transformer_blocks.33.img_norm1.weight",
+      "transformer_blocks.33.img_norm2.weight",
+      "transformer_blocks.33.txt_norm1.weight",
+      "transformer_blocks.33.txt_norm2.weight",
+      "transformer_blocks.34.attn.norm_q.weight",
+      "transformer_blocks.34.attn.norm_k.weight",
+      "transformer_blocks.34.attn.norm_added_q.weight",
+      "transformer_blocks.34.attn.norm_added_k.weight",
+      "transformer_blocks.34.img_norm1.weight",
+      "transformer_blocks.34.img_norm2.weight",
+      "transformer_blocks.34.txt_norm1.weight",
+      "transformer_blocks.34.txt_norm2.weight",
+      "transformer_blocks.35.attn.norm_q.weight",
+      "transformer_blocks.35.attn.norm_k.weight",
+      "transformer_blocks.35.attn.norm_added_q.weight",
+      "transformer_blocks.35.attn.norm_added_k.weight",
+      "transformer_blocks.35.img_norm1.weight",
+      "transformer_blocks.35.img_norm2.weight",
+      "transformer_blocks.35.txt_norm1.weight",
+      "transformer_blocks.35.txt_norm2.weight",
+      "transformer_blocks.36.attn.norm_q.weight",
+      "transformer_blocks.36.attn.norm_k.weight",
+      "transformer_blocks.36.attn.norm_added_q.weight",
+      "transformer_blocks.36.attn.norm_added_k.weight",
+      "transformer_blocks.36.img_norm1.weight",
+      "transformer_blocks.36.img_norm2.weight",
+      "transformer_blocks.36.txt_norm1.weight",
+      "transformer_blocks.36.txt_norm2.weight",
+      "transformer_blocks.37.attn.norm_q.weight",
+      "transformer_blocks.37.attn.norm_k.weight",
+      "transformer_blocks.37.attn.norm_added_q.weight",
+      "transformer_blocks.37.attn.norm_added_k.weight",
+      "transformer_blocks.37.img_norm1.weight",
+      "transformer_blocks.37.img_norm2.weight",
+      "transformer_blocks.37.txt_norm1.weight",
+      "transformer_blocks.37.txt_norm2.weight",
+      "transformer_blocks.38.attn.norm_q.weight",
+      "transformer_blocks.38.attn.norm_k.weight",
+      "transformer_blocks.38.attn.norm_added_q.weight",
+      "transformer_blocks.38.attn.norm_added_k.weight",
+      "transformer_blocks.38.img_norm1.weight",
+      "transformer_blocks.38.img_norm2.weight",
+      "transformer_blocks.38.txt_norm1.weight",
+      "transformer_blocks.38.txt_norm2.weight",
+      "transformer_blocks.39.attn.norm_q.weight",
+      "transformer_blocks.39.attn.norm_k.weight",
+      "transformer_blocks.39.attn.norm_added_q.weight",
+      "transformer_blocks.39.attn.norm_added_k.weight",
+      "transformer_blocks.39.img_norm1.weight",
+      "transformer_blocks.39.img_norm2.weight",
+      "transformer_blocks.39.txt_norm1.weight",
+      "transformer_blocks.39.txt_norm2.weight",
+      "transformer_blocks.40.attn.norm_q.weight",
+      "transformer_blocks.40.attn.norm_k.weight",
+      "transformer_blocks.40.attn.norm_added_q.weight",
+      "transformer_blocks.40.attn.norm_added_k.weight",
+      "transformer_blocks.40.img_norm1.weight",
+      "transformer_blocks.40.img_norm2.weight",
+      "transformer_blocks.40.txt_norm1.weight",
+      "transformer_blocks.40.txt_norm2.weight",
+      "transformer_blocks.41.attn.norm_q.weight",
+      "transformer_blocks.41.attn.norm_k.weight",
+      "transformer_blocks.41.attn.norm_added_q.weight",
+      "transformer_blocks.41.attn.norm_added_k.weight",
+      "transformer_blocks.41.img_norm1.weight",
+      "transformer_blocks.41.img_norm2.weight",
+      "transformer_blocks.41.txt_norm1.weight",
+      "transformer_blocks.41.txt_norm2.weight",
+      "transformer_blocks.42.attn.norm_q.weight",
+      "transformer_blocks.42.attn.norm_k.weight",
+      "transformer_blocks.42.attn.norm_added_q.weight",
+      "transformer_blocks.42.attn.norm_added_k.weight",
+      "transformer_blocks.42.img_norm1.weight",
+      "transformer_blocks.42.img_norm2.weight",
+      "transformer_blocks.42.txt_norm1.weight",
+      "transformer_blocks.42.txt_norm2.weight",
+      "transformer_blocks.43.attn.norm_q.weight",
+      "transformer_blocks.43.attn.norm_k.weight",
+      "transformer_blocks.43.attn.norm_added_q.weight",
+      "transformer_blocks.43.attn.norm_added_k.weight",
+      "transformer_blocks.43.img_norm1.weight",
+      "transformer_blocks.43.img_norm2.weight",
+      "transformer_blocks.43.txt_norm1.weight",
+      "transformer_blocks.43.txt_norm2.weight",
+      "transformer_blocks.44.attn.norm_q.weight",
+      "transformer_blocks.44.attn.norm_k.weight",
+      "transformer_blocks.44.attn.norm_added_q.weight",
+      "transformer_blocks.44.attn.norm_added_k.weight",
+      "transformer_blocks.44.img_norm1.weight",
+      "transformer_blocks.44.img_norm2.weight",
+      "transformer_blocks.44.txt_norm1.weight",
+      "transformer_blocks.44.txt_norm2.weight",
+      "transformer_blocks.45.attn.norm_q.weight",
+      "transformer_blocks.45.attn.norm_k.weight",
+      "transformer_blocks.45.attn.norm_added_q.weight",
+      "transformer_blocks.45.attn.norm_added_k.weight",
+      "transformer_blocks.45.img_norm1.weight",
+      "transformer_blocks.45.img_norm2.weight",
+      "transformer_blocks.45.txt_norm1.weight",
+      "transformer_blocks.45.txt_norm2.weight",
+      "transformer_blocks.46.attn.norm_q.weight",
+      "transformer_blocks.46.attn.norm_k.weight",
+      "transformer_blocks.46.attn.norm_added_q.weight",
+      "transformer_blocks.46.attn.norm_added_k.weight",
+      "transformer_blocks.46.img_norm1.weight",
+      "transformer_blocks.46.img_norm2.weight",
+      "transformer_blocks.46.txt_norm1.weight",
+      "transformer_blocks.46.txt_norm2.weight",
+      "transformer_blocks.47.attn.norm_q.weight",
+      "transformer_blocks.47.attn.norm_k.weight",
+      "transformer_blocks.47.attn.norm_added_q.weight",
+      "transformer_blocks.47.attn.norm_added_k.weight",
+      "transformer_blocks.47.img_norm1.weight",
+      "transformer_blocks.47.img_norm2.weight",
+      "transformer_blocks.47.txt_norm1.weight",
+      "transformer_blocks.47.txt_norm2.weight"
+    ],
+    "modules_to_not_use_matmul": [],
+    "non_blocking": false,
+    "quant_conv": false,
+    "quant_embedding": false,
+    "quant_method": "sdnq",
+    "quantization_device": null,
+    "quantized_matmul_dtype": "int8",
+    "return_device": null,
+    "sdnq_version": "0.2.0",
+    "svd_rank": 32,
+    "svd_steps": 8,
+    "use_dynamic_quantization": false,
+    "use_grad_ckpt": true,
+    "use_hadamard": false,
+    "use_quantized_matmul": true,
+    "use_quantized_matmul_conv": false,
+    "use_static_quantization": true,
+    "use_stochastic_rounding": false,
+    "use_svd": false,
+    "weights_dtype": "uint4"
+  },
+  "rms_norm": true,
+  "selected_layer_index": [
+    5,
+    11,
+    17,
+    23
+  ]
+}

transformer/diffusion_pytorch_model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abd3791d4c163f82fb8e95f7f53bd7af38043e24b19bc5bcab600e2c2633ec87
+size 3987262976

transformer/diffusion_pytorch_model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1065a7a189f5d7dbe23820538667563f615c92685e566e2c8b7d8777ea350c5e
+size 314450920

transformer/quantization_config.json

@@ -0,0 +1,449 @@
+{
+  "add_skip_keys": false,
+  "dequantize_fp32": false,
+  "dynamic_loss_threshold": null,
+  "group_size": 0,
+  "hadamard_group_size": 128,
+  "is_integer": true,
+  "is_training": false,
+  "modules_dtype_dict": {},
+  "modules_quant_config": {},
+  "modules_to_not_convert": [
+    ".emb_in",
+    ".emb_out",
+    "norm",
+    ".y_embedder",
+    "patch_emb",
+    "lm_head",
+    ".condition_embedder",
+    ".norm_out",
+    ".t_embedder",
+    "patch_embedding",
+    ".vid_in",
+    ".txt_in",
+    ".txt_out",
+    ".time_embed",
+    "*.img_mod.*",
+    ".proj_out",
+    ".img_out",
+    "*.txt_mod.*",
+    ".context_embedder",
+    ".vid_out",
+    "time_text_embed",
+    "patch_embed",
+    "wte",
+    ".img_in",
+    ".final_layer",
+    ".x_embedder",
+    "multi_modal_projector",
+    "pos_embed",
+    "txt_norm.0.weight",
+    "txt_norm.1.weight",
+    "txt_norm.2.weight",
+    "txt_norm.3.weight",
+    "transformer_blocks.0.attn.norm_q.weight",
+    "transformer_blocks.0.attn.norm_k.weight",
+    "transformer_blocks.0.attn.norm_added_q.weight",
+    "transformer_blocks.0.attn.norm_added_k.weight",
+    "transformer_blocks.0.img_norm1.weight",
+    "transformer_blocks.0.img_norm2.weight",
+    "transformer_blocks.0.txt_norm1.weight",
+    "transformer_blocks.0.txt_norm2.weight",
+    "transformer_blocks.1.attn.norm_q.weight",
+    "transformer_blocks.1.attn.norm_k.weight",
+    "transformer_blocks.1.attn.norm_added_q.weight",
+    "transformer_blocks.1.attn.norm_added_k.weight",
+    "transformer_blocks.1.img_norm1.weight",
+    "transformer_blocks.1.img_norm2.weight",
+    "transformer_blocks.1.txt_norm1.weight",
+    "transformer_blocks.1.txt_norm2.weight",
+    "transformer_blocks.2.attn.norm_q.weight",
+    "transformer_blocks.2.attn.norm_k.weight",
+    "transformer_blocks.2.attn.norm_added_q.weight",
+    "transformer_blocks.2.attn.norm_added_k.weight",
+    "transformer_blocks.2.img_norm1.weight",
+    "transformer_blocks.2.img_norm2.weight",
+    "transformer_blocks.2.txt_norm1.weight",
+    "transformer_blocks.2.txt_norm2.weight",
+    "transformer_blocks.3.attn.norm_q.weight",
+    "transformer_blocks.3.attn.norm_k.weight",
+    "transformer_blocks.3.attn.norm_added_q.weight",
+    "transformer_blocks.3.attn.norm_added_k.weight",
+    "transformer_blocks.3.img_norm1.weight",
+    "transformer_blocks.3.img_norm2.weight",
+    "transformer_blocks.3.txt_norm1.weight",
+    "transformer_blocks.3.txt_norm2.weight",
+    "transformer_blocks.4.attn.norm_q.weight",
+    "transformer_blocks.4.attn.norm_k.weight",
+    "transformer_blocks.4.attn.norm_added_q.weight",
+    "transformer_blocks.4.attn.norm_added_k.weight",
+    "transformer_blocks.4.img_norm1.weight",
+    "transformer_blocks.4.img_norm2.weight",
+    "transformer_blocks.4.txt_norm1.weight",
+    "transformer_blocks.4.txt_norm2.weight",
+    "transformer_blocks.5.attn.norm_q.weight",
+    "transformer_blocks.5.attn.norm_k.weight",
+    "transformer_blocks.5.attn.norm_added_q.weight",
+    "transformer_blocks.5.attn.norm_added_k.weight",
+    "transformer_blocks.5.img_norm1.weight",
+    "transformer_blocks.5.img_norm2.weight",
+    "transformer_blocks.5.txt_norm1.weight",
+    "transformer_blocks.5.txt_norm2.weight",
+    "transformer_blocks.6.attn.norm_q.weight",
+    "transformer_blocks.6.attn.norm_k.weight",
+    "transformer_blocks.6.attn.norm_added_q.weight",
+    "transformer_blocks.6.attn.norm_added_k.weight",
+    "transformer_blocks.6.img_norm1.weight",
+    "transformer_blocks.6.img_norm2.weight",
+    "transformer_blocks.6.txt_norm1.weight",
+    "transformer_blocks.6.txt_norm2.weight",
+    "transformer_blocks.7.attn.norm_q.weight",
+    "transformer_blocks.7.attn.norm_k.weight",
+    "transformer_blocks.7.attn.norm_added_q.weight",
+    "transformer_blocks.7.attn.norm_added_k.weight",
+    "transformer_blocks.7.img_norm1.weight",
+    "transformer_blocks.7.img_norm2.weight",
+    "transformer_blocks.7.txt_norm1.weight",
+    "transformer_blocks.7.txt_norm2.weight",
+    "transformer_blocks.8.attn.norm_q.weight",
+    "transformer_blocks.8.attn.norm_k.weight",
+    "transformer_blocks.8.attn.norm_added_q.weight",
+    "transformer_blocks.8.attn.norm_added_k.weight",
+    "transformer_blocks.8.img_norm1.weight",
+    "transformer_blocks.8.img_norm2.weight",
+    "transformer_blocks.8.txt_norm1.weight",
+    "transformer_blocks.8.txt_norm2.weight",
+    "transformer_blocks.9.attn.norm_q.weight",
+    "transformer_blocks.9.attn.norm_k.weight",
+    "transformer_blocks.9.attn.norm_added_q.weight",
+    "transformer_blocks.9.attn.norm_added_k.weight",
+    "transformer_blocks.9.img_norm1.weight",
+    "transformer_blocks.9.img_norm2.weight",
+    "transformer_blocks.9.txt_norm1.weight",
+    "transformer_blocks.9.txt_norm2.weight",
+    "transformer_blocks.10.attn.norm_q.weight",
+    "transformer_blocks.10.attn.norm_k.weight",
+    "transformer_blocks.10.attn.norm_added_q.weight",
+    "transformer_blocks.10.attn.norm_added_k.weight",
+    "transformer_blocks.10.img_norm1.weight",
+    "transformer_blocks.10.img_norm2.weight",
+    "transformer_blocks.10.txt_norm1.weight",
+    "transformer_blocks.10.txt_norm2.weight",
+    "transformer_blocks.11.attn.norm_q.weight",
+    "transformer_blocks.11.attn.norm_k.weight",
+    "transformer_blocks.11.attn.norm_added_q.weight",
+    "transformer_blocks.11.attn.norm_added_k.weight",
+    "transformer_blocks.11.img_norm1.weight",
+    "transformer_blocks.11.img_norm2.weight",
+    "transformer_blocks.11.txt_norm1.weight",
+    "transformer_blocks.11.txt_norm2.weight",
+    "transformer_blocks.12.attn.norm_q.weight",
+    "transformer_blocks.12.attn.norm_k.weight",
+    "transformer_blocks.12.attn.norm_added_q.weight",
+    "transformer_blocks.12.attn.norm_added_k.weight",
+    "transformer_blocks.12.img_norm1.weight",
+    "transformer_blocks.12.img_norm2.weight",
+    "transformer_blocks.12.txt_norm1.weight",
+    "transformer_blocks.12.txt_norm2.weight",
+    "transformer_blocks.13.attn.norm_q.weight",
+    "transformer_blocks.13.attn.norm_k.weight",
+    "transformer_blocks.13.attn.norm_added_q.weight",
+    "transformer_blocks.13.attn.norm_added_k.weight",
+    "transformer_blocks.13.img_norm1.weight",
+    "transformer_blocks.13.img_norm2.weight",
+    "transformer_blocks.13.txt_norm1.weight",
+    "transformer_blocks.13.txt_norm2.weight",
+    "transformer_blocks.14.attn.norm_q.weight",
+    "transformer_blocks.14.attn.norm_k.weight",
+    "transformer_blocks.14.attn.norm_added_q.weight",
+    "transformer_blocks.14.attn.norm_added_k.weight",
+    "transformer_blocks.14.img_norm1.weight",
+    "transformer_blocks.14.img_norm2.weight",
+    "transformer_blocks.14.txt_norm1.weight",
+    "transformer_blocks.14.txt_norm2.weight",
+    "transformer_blocks.15.attn.norm_q.weight",
+    "transformer_blocks.15.attn.norm_k.weight",
+    "transformer_blocks.15.attn.norm_added_q.weight",
+    "transformer_blocks.15.attn.norm_added_k.weight",
+    "transformer_blocks.15.img_norm1.weight",
+    "transformer_blocks.15.img_norm2.weight",
+    "transformer_blocks.15.txt_norm1.weight",
+    "transformer_blocks.15.txt_norm2.weight",
+    "transformer_blocks.16.attn.norm_q.weight",
+    "transformer_blocks.16.attn.norm_k.weight",
+    "transformer_blocks.16.attn.norm_added_q.weight",
+    "transformer_blocks.16.attn.norm_added_k.weight",
+    "transformer_blocks.16.img_norm1.weight",
+    "transformer_blocks.16.img_norm2.weight",
+    "transformer_blocks.16.txt_norm1.weight",
+    "transformer_blocks.16.txt_norm2.weight",
+    "transformer_blocks.17.attn.norm_q.weight",
+    "transformer_blocks.17.attn.norm_k.weight",
+    "transformer_blocks.17.attn.norm_added_q.weight",
+    "transformer_blocks.17.attn.norm_added_k.weight",
+    "transformer_blocks.17.img_norm1.weight",
+    "transformer_blocks.17.img_norm2.weight",
+    "transformer_blocks.17.txt_norm1.weight",
+    "transformer_blocks.17.txt_norm2.weight",
+    "transformer_blocks.18.attn.norm_q.weight",
+    "transformer_blocks.18.attn.norm_k.weight",
+    "transformer_blocks.18.attn.norm_added_q.weight",
+    "transformer_blocks.18.attn.norm_added_k.weight",
+    "transformer_blocks.18.img_norm1.weight",
+    "transformer_blocks.18.img_norm2.weight",
+    "transformer_blocks.18.txt_norm1.weight",
+    "transformer_blocks.18.txt_norm2.weight",
+    "transformer_blocks.19.attn.norm_q.weight",
+    "transformer_blocks.19.attn.norm_k.weight",
+    "transformer_blocks.19.attn.norm_added_q.weight",
+    "transformer_blocks.19.attn.norm_added_k.weight",
+    "transformer_blocks.19.img_norm1.weight",
+    "transformer_blocks.19.img_norm2.weight",
+    "transformer_blocks.19.txt_norm1.weight",
+    "transformer_blocks.19.txt_norm2.weight",
+    "transformer_blocks.20.attn.norm_q.weight",
+    "transformer_blocks.20.attn.norm_k.weight",
+    "transformer_blocks.20.attn.norm_added_q.weight",
+    "transformer_blocks.20.attn.norm_added_k.weight",
+    "transformer_blocks.20.img_norm1.weight",
+    "transformer_blocks.20.img_norm2.weight",
+    "transformer_blocks.20.txt_norm1.weight",
+    "transformer_blocks.20.txt_norm2.weight",
+    "transformer_blocks.21.attn.norm_q.weight",
+    "transformer_blocks.21.attn.norm_k.weight",
+    "transformer_blocks.21.attn.norm_added_q.weight",
+    "transformer_blocks.21.attn.norm_added_k.weight",
+    "transformer_blocks.21.img_norm1.weight",
+    "transformer_blocks.21.img_norm2.weight",
+    "transformer_blocks.21.txt_norm1.weight",
+    "transformer_blocks.21.txt_norm2.weight",
+    "transformer_blocks.22.attn.norm_q.weight",
+    "transformer_blocks.22.attn.norm_k.weight",
+    "transformer_blocks.22.attn.norm_added_q.weight",
+    "transformer_blocks.22.attn.norm_added_k.weight",
+    "transformer_blocks.22.img_norm1.weight",
+    "transformer_blocks.22.img_norm2.weight",
+    "transformer_blocks.22.txt_norm1.weight",
+    "transformer_blocks.22.txt_norm2.weight",
+    "transformer_blocks.23.attn.norm_q.weight",
+    "transformer_blocks.23.attn.norm_k.weight",
+    "transformer_blocks.23.attn.norm_added_q.weight",
+    "transformer_blocks.23.attn.norm_added_k.weight",
+    "transformer_blocks.23.img_norm1.weight",
+    "transformer_blocks.23.img_norm2.weight",
+    "transformer_blocks.23.txt_norm1.weight",
+    "transformer_blocks.23.txt_norm2.weight",
+    "transformer_blocks.24.attn.norm_q.weight",
+    "transformer_blocks.24.attn.norm_k.weight",
+    "transformer_blocks.24.attn.norm_added_q.weight",
+    "transformer_blocks.24.attn.norm_added_k.weight",
+    "transformer_blocks.24.img_norm1.weight",
+    "transformer_blocks.24.img_norm2.weight",
+    "transformer_blocks.24.txt_norm1.weight",
+    "transformer_blocks.24.txt_norm2.weight",
+    "transformer_blocks.25.attn.norm_q.weight",
+    "transformer_blocks.25.attn.norm_k.weight",
+    "transformer_blocks.25.attn.norm_added_q.weight",
+    "transformer_blocks.25.attn.norm_added_k.weight",
+    "transformer_blocks.25.img_norm1.weight",
+    "transformer_blocks.25.img_norm2.weight",
+    "transformer_blocks.25.txt_norm1.weight",
+    "transformer_blocks.25.txt_norm2.weight",
+    "transformer_blocks.26.attn.norm_q.weight",
+    "transformer_blocks.26.attn.norm_k.weight",
+    "transformer_blocks.26.attn.norm_added_q.weight",
+    "transformer_blocks.26.attn.norm_added_k.weight",
+    "transformer_blocks.26.img_norm1.weight",
+    "transformer_blocks.26.img_norm2.weight",
+    "transformer_blocks.26.txt_norm1.weight",
+    "transformer_blocks.26.txt_norm2.weight",
+    "transformer_blocks.27.attn.norm_q.weight",
+    "transformer_blocks.27.attn.norm_k.weight",
+    "transformer_blocks.27.attn.norm_added_q.weight",
+    "transformer_blocks.27.attn.norm_added_k.weight",
+    "transformer_blocks.27.img_norm1.weight",
+    "transformer_blocks.27.img_norm2.weight",
+    "transformer_blocks.27.txt_norm1.weight",
+    "transformer_blocks.27.txt_norm2.weight",
+    "transformer_blocks.28.attn.norm_q.weight",
+    "transformer_blocks.28.attn.norm_k.weight",
+    "transformer_blocks.28.attn.norm_added_q.weight",
+    "transformer_blocks.28.attn.norm_added_k.weight",
+    "transformer_blocks.28.img_norm1.weight",
+    "transformer_blocks.28.img_norm2.weight",
+    "transformer_blocks.28.txt_norm1.weight",
+    "transformer_blocks.28.txt_norm2.weight",
+    "transformer_blocks.29.attn.norm_q.weight",
+    "transformer_blocks.29.attn.norm_k.weight",
+    "transformer_blocks.29.attn.norm_added_q.weight",
+    "transformer_blocks.29.attn.norm_added_k.weight",
+    "transformer_blocks.29.img_norm1.weight",
+    "transformer_blocks.29.img_norm2.weight",
+    "transformer_blocks.29.txt_norm1.weight",
+    "transformer_blocks.29.txt_norm2.weight",
+    "transformer_blocks.30.attn.norm_q.weight",
+    "transformer_blocks.30.attn.norm_k.weight",
+    "transformer_blocks.30.attn.norm_added_q.weight",
+    "transformer_blocks.30.attn.norm_added_k.weight",
+    "transformer_blocks.30.img_norm1.weight",
+    "transformer_blocks.30.img_norm2.weight",
+    "transformer_blocks.30.txt_norm1.weight",
+    "transformer_blocks.30.txt_norm2.weight",
+    "transformer_blocks.31.attn.norm_q.weight",
+    "transformer_blocks.31.attn.norm_k.weight",
+    "transformer_blocks.31.attn.norm_added_q.weight",
+    "transformer_blocks.31.attn.norm_added_k.weight",
+    "transformer_blocks.31.img_norm1.weight",
+    "transformer_blocks.31.img_norm2.weight",
+    "transformer_blocks.31.txt_norm1.weight",
+    "transformer_blocks.31.txt_norm2.weight",
+    "transformer_blocks.32.attn.norm_q.weight",
+    "transformer_blocks.32.attn.norm_k.weight",
+    "transformer_blocks.32.attn.norm_added_q.weight",
+    "transformer_blocks.32.attn.norm_added_k.weight",
+    "transformer_blocks.32.img_norm1.weight",
+    "transformer_blocks.32.img_norm2.weight",
+    "transformer_blocks.32.txt_norm1.weight",
+    "transformer_blocks.32.txt_norm2.weight",
+    "transformer_blocks.33.attn.norm_q.weight",
+    "transformer_blocks.33.attn.norm_k.weight",
+    "transformer_blocks.33.attn.norm_added_q.weight",
+    "transformer_blocks.33.attn.norm_added_k.weight",
+    "transformer_blocks.33.img_norm1.weight",
+    "transformer_blocks.33.img_norm2.weight",
+    "transformer_blocks.33.txt_norm1.weight",
+    "transformer_blocks.33.txt_norm2.weight",
+    "transformer_blocks.34.attn.norm_q.weight",
+    "transformer_blocks.34.attn.norm_k.weight",
+    "transformer_blocks.34.attn.norm_added_q.weight",
+    "transformer_blocks.34.attn.norm_added_k.weight",
+    "transformer_blocks.34.img_norm1.weight",
+    "transformer_blocks.34.img_norm2.weight",
+    "transformer_blocks.34.txt_norm1.weight",
+    "transformer_blocks.34.txt_norm2.weight",
+    "transformer_blocks.35.attn.norm_q.weight",
+    "transformer_blocks.35.attn.norm_k.weight",
+    "transformer_blocks.35.attn.norm_added_q.weight",
+    "transformer_blocks.35.attn.norm_added_k.weight",
+    "transformer_blocks.35.img_norm1.weight",
+    "transformer_blocks.35.img_norm2.weight",
+    "transformer_blocks.35.txt_norm1.weight",
+    "transformer_blocks.35.txt_norm2.weight",
+    "transformer_blocks.36.attn.norm_q.weight",
+    "transformer_blocks.36.attn.norm_k.weight",
+    "transformer_blocks.36.attn.norm_added_q.weight",
+    "transformer_blocks.36.attn.norm_added_k.weight",
+    "transformer_blocks.36.img_norm1.weight",
+    "transformer_blocks.36.img_norm2.weight",
+    "transformer_blocks.36.txt_norm1.weight",
+    "transformer_blocks.36.txt_norm2.weight",
+    "transformer_blocks.37.attn.norm_q.weight",
+    "transformer_blocks.37.attn.norm_k.weight",
+    "transformer_blocks.37.attn.norm_added_q.weight",
+    "transformer_blocks.37.attn.norm_added_k.weight",
+    "transformer_blocks.37.img_norm1.weight",
+    "transformer_blocks.37.img_norm2.weight",
+    "transformer_blocks.37.txt_norm1.weight",
+    "transformer_blocks.37.txt_norm2.weight",
+    "transformer_blocks.38.attn.norm_q.weight",
+    "transformer_blocks.38.attn.norm_k.weight",
+    "transformer_blocks.38.attn.norm_added_q.weight",
+    "transformer_blocks.38.attn.norm_added_k.weight",
+    "transformer_blocks.38.img_norm1.weight",
+    "transformer_blocks.38.img_norm2.weight",
+    "transformer_blocks.38.txt_norm1.weight",
+    "transformer_blocks.38.txt_norm2.weight",
+    "transformer_blocks.39.attn.norm_q.weight",
+    "transformer_blocks.39.attn.norm_k.weight",
+    "transformer_blocks.39.attn.norm_added_q.weight",
+    "transformer_blocks.39.attn.norm_added_k.weight",
+    "transformer_blocks.39.img_norm1.weight",
+    "transformer_blocks.39.img_norm2.weight",
+    "transformer_blocks.39.txt_norm1.weight",
+    "transformer_blocks.39.txt_norm2.weight",
+    "transformer_blocks.40.attn.norm_q.weight",
+    "transformer_blocks.40.attn.norm_k.weight",
+    "transformer_blocks.40.attn.norm_added_q.weight",
+    "transformer_blocks.40.attn.norm_added_k.weight",
+    "transformer_blocks.40.img_norm1.weight",
+    "transformer_blocks.40.img_norm2.weight",
+    "transformer_blocks.40.txt_norm1.weight",
+    "transformer_blocks.40.txt_norm2.weight",
+    "transformer_blocks.41.attn.norm_q.weight",
+    "transformer_blocks.41.attn.norm_k.weight",
+    "transformer_blocks.41.attn.norm_added_q.weight",
+    "transformer_blocks.41.attn.norm_added_k.weight",
+    "transformer_blocks.41.img_norm1.weight",
+    "transformer_blocks.41.img_norm2.weight",
+    "transformer_blocks.41.txt_norm1.weight",
+    "transformer_blocks.41.txt_norm2.weight",
+    "transformer_blocks.42.attn.norm_q.weight",
+    "transformer_blocks.42.attn.norm_k.weight",
+    "transformer_blocks.42.attn.norm_added_q.weight",
+    "transformer_blocks.42.attn.norm_added_k.weight",
+    "transformer_blocks.42.img_norm1.weight",
+    "transformer_blocks.42.img_norm2.weight",
+    "transformer_blocks.42.txt_norm1.weight",
+    "transformer_blocks.42.txt_norm2.weight",
+    "transformer_blocks.43.attn.norm_q.weight",
+    "transformer_blocks.43.attn.norm_k.weight",
+    "transformer_blocks.43.attn.norm_added_q.weight",
+    "transformer_blocks.43.attn.norm_added_k.weight",
+    "transformer_blocks.43.img_norm1.weight",
+    "transformer_blocks.43.img_norm2.weight",
+    "transformer_blocks.43.txt_norm1.weight",
+    "transformer_blocks.43.txt_norm2.weight",
+    "transformer_blocks.44.attn.norm_q.weight",
+    "transformer_blocks.44.attn.norm_k.weight",
+    "transformer_blocks.44.attn.norm_added_q.weight",
+    "transformer_blocks.44.attn.norm_added_k.weight",
+    "transformer_blocks.44.img_norm1.weight",
+    "transformer_blocks.44.img_norm2.weight",
+    "transformer_blocks.44.txt_norm1.weight",
+    "transformer_blocks.44.txt_norm2.weight",
+    "transformer_blocks.45.attn.norm_q.weight",
+    "transformer_blocks.45.attn.norm_k.weight",
+    "transformer_blocks.45.attn.norm_added_q.weight",
+    "transformer_blocks.45.attn.norm_added_k.weight",
+    "transformer_blocks.45.img_norm1.weight",
+    "transformer_blocks.45.img_norm2.weight",
+    "transformer_blocks.45.txt_norm1.weight",
+    "transformer_blocks.45.txt_norm2.weight",
+    "transformer_blocks.46.attn.norm_q.weight",
+    "transformer_blocks.46.attn.norm_k.weight",
+    "transformer_blocks.46.attn.norm_added_q.weight",
+    "transformer_blocks.46.attn.norm_added_k.weight",
+    "transformer_blocks.46.img_norm1.weight",
+    "transformer_blocks.46.img_norm2.weight",
+    "transformer_blocks.46.txt_norm1.weight",
+    "transformer_blocks.46.txt_norm2.weight",
+    "transformer_blocks.47.attn.norm_q.weight",
+    "transformer_blocks.47.attn.norm_k.weight",
+    "transformer_blocks.47.attn.norm_added_q.weight",
+    "transformer_blocks.47.attn.norm_added_k.weight",
+    "transformer_blocks.47.img_norm1.weight",
+    "transformer_blocks.47.img_norm2.weight",
+    "transformer_blocks.47.txt_norm1.weight",
+    "transformer_blocks.47.txt_norm2.weight"
+  ],
+  "modules_to_not_use_matmul": [],
+  "non_blocking": false,
+  "quant_conv": false,
+  "quant_embedding": false,
+  "quant_method": "sdnq",
+  "quantization_device": null,
+  "quantized_matmul_dtype": "int8",
+  "return_device": null,
+  "sdnq_version": "0.2.0",
+  "svd_rank": 32,
+  "svd_steps": 8,
+  "use_dynamic_quantization": false,
+  "use_grad_ckpt": true,
+  "use_hadamard": false,
+  "use_quantized_matmul": true,
+  "use_quantized_matmul_conv": false,
+  "use_static_quantization": true,
+  "use_stochastic_rounding": false,
+  "use_svd": false,
+  "weights_dtype": "uint4"
+}

vae/config.json

@@ -0,0 +1,40 @@
+{
+  "_class_name": "AutoencoderKLFlux2",
+  "_diffusers_version": "0.37.1",
+  "_name_or_path": "black-forest-labs/FLUX.2-dev",
+  "act_fn": "silu",
+  "batch_norm_eps": 0.0001,
+  "batch_norm_momentum": 0.1,
+  "block_out_channels": [
+    128,
+    256,
+    512,
+    512
+  ],
+  "down_block_types": [
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D"
+  ],
+  "force_upcast": true,
+  "in_channels": 3,
+  "latent_channels": 32,
+  "layers_per_block": 2,
+  "mid_block_add_attention": true,
+  "norm_num_groups": 32,
+  "out_channels": 3,
+  "patch_size": [
+    2,
+    2
+  ],
+  "sample_size": 1024,
+  "up_block_types": [
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D"
+  ],
+  "use_post_quant_conv": true,
+  "use_quant_conv": true
+}

vae/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d64f3a68e1cc4f9f4e29b6e0da38a0204fe9a49f2d4053f0ec1fa1ca02f9c4b5
+size 336213556