Duplicate from realrebelai/Rebels_w4a8s

Co-authored-by: Rebel AI <realrebelai@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz 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

FLUX/Flux2-Klein-4B-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddb9566f8ff12df0b72794901b252e28c938bb9abf8ade716d804a21c33aff2e
+size 2464248722

FLUX/Flux2-Klein-9B-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de83a64b5617325a9ca39a112cb72dd131127dc2174dfe2c5bbdc9e7200d1839
+size 5623109192

KREA/Krea-2-Turbo-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:468bf97902559bfab6566b71fe10601a6098b221b27a9cd1aaaef49c14cef056
+size 7226824858

LTX/ENCODERS/gemma4-12b-ltx25-w4a8-v2.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94dc5813f8ea6863bcb28fa055fa5e39324bc67389e5dfaa25bd257f38f6ba84
+size 8441692051

LTX/ENCODERS/gemma4-12b-ltx25-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:323a6294aef6e87f1f0af3d79ae94b9420a341d8c8c273e1400528c3e76330c3
+size 8441694015

LTX/LTX-2.5-Distilled-w4a8-audiofix.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4adc072ff3bc4bbf0daf7f186edc1d6c954678d953e03ebd5dd94b396ab996c7
+size 15153292760

LTX/LTX-2.5-Distilled-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b07a9df6a2f16cec62012850aa805ee84708827e94481dec82183c416b5e89a
+size 12520544200

MiniMax/MiniMax-H3-REF2VA-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1475de170900a18c2a5881d378d695ae4bb11a3a296f5c42132c7bf583bc7b76
+size 24482659432

MiniMax/MiniMax-Music-3-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:501ac0f83e3df90157909b2cebfa9fda8115640ff897f78963c273bdf5f65e5e
+size 1510463731

QWEN/Qwen-Image-2512-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb97a56152e6ea6bd03ec1f14c318cc815f38b5483741aec75a680e13e31c367
+size 14545127544

README.md

@@ -0,0 +1,148 @@
+---
+license: other
+license_name: see-individual-model-repos
+pipeline_tag: text-to-image
+tags:
+  - w4a8
+  - quantized
+  - int4
+  - comfyui
+  - comfy-kitchen
+  - convrot
+  - low-vram
+  - safetensors
+---
+
+# Rebels W4A8 Collection
+
+ 
+<table align="center">
+ <tr>
+    <td><img src="https://cdn-uploads.huggingface.co/production/uploads/68761990332d15464ccc8dee/SNhCJ0PXAiQ9eAD18Ywtt.png" width="300"/></td>
+    <td><img src="https://cdn-uploads.huggingface.co/production/uploads/68761990332d15464ccc8dee/89-J55SmjsVOtPH4gaUz-.png" width="300"/></td>
+    <td><img src="https://cdn-uploads.huggingface.co/production/uploads/68761990332d15464ccc8dee/UkEPjN6KJUsomhZiD8oc7.png" width="300"/></td>
+  </tr>
+</table>
+
+
+
+## Models
+
+| Model | Params | Size | Notes |
+|---|---|---|---|
+| **Flux2-Klein-4B-w4a8** | 3.9 B | **2.46 GB** | Apache-2.0. Smallest here — runs on 6 GB cards |
+| **Z-Image-Turbo-w4a8** | 6.2 B | **3.67 GB** | Distilled — 8 steps, CFG 1.0 |
+| **Flux2-Klein-9B-w4a8** | 9.1 B | **5.62 GB** | Flux License |
+| **Krea-2-Turbo-w4a8** | 12.8 B | **7.23 GB** | Turbo — CFG 1.0, not 4.0 |
+| **SCAIL-2-14B-w4a8** | ~14 B | **_.__ GB** | Wan2.1-based video, NEEDS WORK |
+| **Wan-Animate-2-TURBO-w4a8** | 16.4 B | **9.59 GB** | Video. **Reference image must match the driving video's opening pose** |
+| **Qwen-Image-2512-w4a8** | ~20 B | **14.5 GB** | Full model — normal steps and CFG, not the Flash recipe, NEEDS WORK |
+| **MiniMax-H3-REF2VA-w4a8** | 33.1 B | **24.5 GB** | Reference-to-video with audio. Mixed: `adaln_proj` at int8 — its 2688-wide rows can't use the w4a8 kernel |
+
+All load with the stock **Load Diffusion Model** node. Base licenses carry over — check each source model before commercial use.
+
+All load with the stock **Load Diffusion Model** node. Base licenses carry over — check each source model before commercial use.
+
+**Four-bit weights. Eight-bit math. Native ComfyUI kernels. No custom nodes.**
+
+W4A8 conversions of current diffusion and video models, quantized to run on ComfyUI's native `asym_w4a8_int8` path — where int8 tensor cores do the work instead of a dequantize-then-fp16 fallback.
+
+Roughly **0.56 bytes per parameter**, and it *runs* at that size rather than merely storing at it.
+
+---
+<table align="center">
+  <tr>
+    <td><img src="https://cdn-uploads.huggingface.co/production/uploads/68761990332d15464ccc8dee/khVfxQbfqNsXy7ceyD8Y9.png" alt="INT8" width="150"/></td>
+    <td><img src="https://cdn-uploads.huggingface.co/production/uploads/68761990332d15464ccc8dee/hVnkIGDU8p8hh90cVhowJ.png" alt="w4a8" width="150"/></td>
+  </tr>
+  <tr>
+    <td align="center"><b>INT8</b></td>
+    <td align="center"><b>w4a8</b></td>
+  </tr>
+</table>
+
+
+## Why W4A8 instead of GGUF
+
+GGUF is excellent and I ship plenty of it. But every GGUF forward pass unpacks weights back to fp16 before the matmul — the file is small, the math is not. W4A8 keeps compute in int8 end to end.
+
+| | GGUF Q4_K_M | W4A8 |
+|---|---|---|
+| Storage | ~0.60 B/elem | ~0.56 B/elem |
+| Compute path | dequant → fp16 GEMM | int8 GEMM |
+| Loader | ComfyUI-GGUF node | **stock Load Diffusion Model** |
+| Weight error (measured) | varies by tensor | ~7% relL2 |
+
+The format comes from Kijai's `AsymW4A8Int8Layout` work in comfy-kitchen. This collection is about applying it correctly to models nobody has converted yet, and being explicit about what was verified.
+
+---
+
+
+
+
+## What's inside a file
+
+Each quantized Linear stores five pieces:
+
+| Tensor | Purpose |
+|---|---|
+| `weight` | int4 codes, two per byte |
+| `weight_s_rel` | fp8 scale, one per group of 16 |
+| `weight_s_channel` | one scale per output channel |
+| `weight_codebook` | 16 Lloyd-Max levels, fit to the tensor |
+| `comfy_quant` | layout config the loader reads |
+
+Three ideas stacked: a **ConvRot Hadamard rotation** that flattens outliers so four bits go further, a **codebook** of non-uniform levels fit to the actual weight distribution instead of an even grid, and **per-group fp8 scales** preserving local dynamic range. Calibration-free — no activation dataset, so nothing in the conversion biases the model toward one kind of prompt.
+
+---
+
+## What I do differently
+
+**Sensitive layers are never quantized.** Timestep embeddings, conditioning projections, patch projections, final output layers and rotary tables stay high precision. On a few-step model the timestep embedder has only a handful of sigma values to distinguish — crushing it to four bits corrupts every step of the schedule. Every file is checked after conversion to confirm those layers really are stored at F16/F32, because quantizers do not preserve them automatically.
+
+**Mixed formats where the kernel demands it.** The fused W4A8 kernel accepts a ConvRot group of exactly 256, so any layer whose input dimension isn't divisible by 256 cannot use it. Rather than silently shipping a file that errors on load, those layers are written as `int8_tensorwise` — also native, no group constraint, ~1% error. Each model card states which layers took that path.
+
+**Every file is measured.** Conversion reports per-layer reconstruction error against the original bf16 weights. Anything that doesn't land where it should doesn't get uploaded.
+
+---
+
+## Requirements
+
+- **ComfyUI 0.30.0+** with `asym_w4a8_int8` in its native quant registry
+- **comfy-kitchen** installed (ships the kernels)
+- An NVIDIA GPU or AMD GPU.
+
+On startup ComfyUI prints its available formats. You want `asym_w4a8_int8` in the **Native ops** list — under *emulated* it still runs, without the int8 speed advantage.
+
+---
+
+## Usage
+
+1. Drop the `.safetensors` in `ComfyUI/models/diffusion_models`
+2. Load it with **Load Diffusion Model** — the stock node, no custom loader
+3. Text encoder, VAE and sampler settings are unchanged from the base model
+
+Per-model notes (step counts, CFG, resolution) live in each model's card. Distilled models have fixed schedules that must be respected — base-model settings on them produce poor results regardless of quantization.
+
+---
+
+## Models
+
+*Collection in progress. Each conversion has its own repo with exact sizes, measured error, and the list of layers kept at high precision.*
+
+---
+
+## Licensing
+
+These are quantized derivatives. **Every original license and usage restriction carries over unchanged**, and each model repo states the license of its base model. Check the specific model's card before commercial use — several bases in this collection are not permissive.
+
+---
+
+## Credits
+
+- **Kijai** — the W4A8 int8-codebook layout and kernels
+- **Comfy-Org / comfyanonymous** — comfy-kitchen and the native quantization registry
+- **city96** — ComfyUI-GGUF, which taught most of us how quantized loading works in ComfyUI
+- Original model authors — all base licenses apply
+
+Quantized by [RealRebelAI](https://huggingface.co/realrebelai) · [GitHub](https://github.com/RealRebelAI) · [X](https://x.com/realrebelai)

SCAIL/SCAIL-2-14B-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b9f3df6bdd644b18a94eccd5c98d197fa2184de3ba9f37b679ce3d0effc91fa
+size 9592737504

Wan/Wan-Animate-2-TURBO-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd39bf0df36c4656f0a05b299d82e8c0cb1e3521e9e35da0ed8606e2235702b2
+size 9588609960

Z-Image/Z-Image-Turbo-w4a8.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e35a61d1ebe2e4f8f6c059c6f5bb45b20cc0701208a1c3d51a394d8d17605d9
+size 3667086492

Z-Image/z_image_base_w4a8-mixed.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e0847681fbfc3154b2233244a61e18928a48f859043d5609d89c68eda91a8bd7
+size 4339705712

config.json

@@ -0,0 +1,8 @@
+{
+  "quantization": "asym_w4a8_int8",
+  "format": "w4a8",
+  "weight_bits": 4,
+  "activation_bits": 8,
+  "loader": "comfyui",
+  "library": "comfy-kitchen"
+}