Update README

.gitattributes

@@ -44,3 +44,7 @@ zimage_quanto_bench_results/images/int8/portrait_01_seed46.png filter=lfs diff=l
 zimage_quanto_bench_results/images/int8/portrait_02_seed123.png filter=lfs diff=lfs merge=lfs -text
 zimage_quanto_bench_results/images/int8/scene_01_seed777.png filter=lfs diff=lfs merge=lfs -text
 tokenizer/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+zimage_quanto_bench_results/images/baseline/landscape_01_seed123.png filter=lfs diff=lfs merge=lfs -text
+zimage_quanto_bench_results/images/baseline/portrait_02_seed111.png filter=lfs diff=lfs merge=lfs -text
+zimage_quanto_bench_results/images/int8/landscape_01_seed123.png filter=lfs diff=lfs merge=lfs -text
+zimage_quanto_bench_results/images/int8/portrait_02_seed111.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -33,6 +33,12 @@ This repository provides an INT8-quantized variant of [Tongyi-MAI/Z-Image](https
 - **Pipeline**: `diffusers.ZImagePipeline`
 - **Negative prompt support**: Yes (same pipeline API as the base model)
 
+## Platform Support
+
+- ✅ Supported: Linux/Windows with NVIDIA CUDA
+- ⚠️ Limited support: macOS Apple Silicon (MPS, usually much slower than CUDA)
+- ❌ Not supported: macOS Intel
+
 ## Files
 
 Key files in this repository:
@@ -61,14 +67,11 @@ python -m pip install --upgrade pip
 # PyTorch (NVIDIA CUDA, example)
 pip install torch --index-url https://download.pytorch.org/whl/cu128
 
-# PyTorch (macOS / CPU-only example)
+# PyTorch (macOS Apple Silicon, MPS)
 # pip install torch
 
 # Inference dependencies
 pip install diffusers transformers accelerate safetensors sentencepiece optimum-quanto pillow
-
-# Recommended minimum versions (helps avoid backend compatibility issues)
-pip install -U "torch>=2.4" "diffusers>=0.36.0" "accelerate>=0.33"
 ```
 
 ## Quick Start (Diffusers)
@@ -77,6 +80,7 @@ This repo already stores quantized weights, so you do **not** need to re-run qua
 
 ```python
 import torch
+
 from diffusers import ZImagePipeline
 
 model_id = "ixim/Z-Image-INT8"
@@ -89,7 +93,7 @@ elif torch.backends.mps.is_available():
     device = "mps"
     dtype = torch.float16
 else:
-    # Intel Mac / CPU-only
+    # CPU fallback (functional but very slow for this model)
     device = "cpu"
     dtype = torch.float32
 
@@ -105,7 +109,7 @@ else:
     pipe = pipe.to(device)
 
 prompt = "A cinematic portrait of a young woman, soft lighting, high detail"
-negative_prompt = "blurry, low quality, distorted face, extra limbs, artifacts"
+negative_prompt = "blurry, sad, low quality, distorted face, extra limbs, artifacts"
 # Use CPU generator for best cross-device compatibility (cpu/mps/cuda)
 generator = torch.Generator(device="cpu").manual_seed(42)
 
@@ -125,9 +129,8 @@ print("Saved: zimage_int8_sample.png")
 
 ## macOS Notes & Troubleshooting
 
-- `AttributeError: module 'torch' has no attribute 'xpu'` is usually a backend/version compatibility issue in the local environment, not a model issue.
-- Fix it by upgrading to recent versions:
-    - `pip install -U "torch>=2.4" "diffusers>=0.36.0" "accelerate>=0.33"`
+- macOS Intel is no longer supported for this model in this repository.
+- If you need macOS inference, use Apple Silicon (`mps`) only.
 - On Apple Silicon, warnings like `CUDA not available` and `Disabling autocast` are expected in non-CUDA execution paths.
 - Slow speed on Mac is expected compared with high-end NVIDIA GPUs. To improve speed on Apple Silicon:
     - Ensure the script uses `mps` (as in the example above), not `cpu`.
@@ -154,15 +157,15 @@ These two images are generated with this quantized model:
 Test environment:
 - GPU: NVIDIA GeForce RTX 5090
 - Framework: PyTorch 2.10.0+cu130
-- Inference setting: 1024×1024, 28 steps, guidance=4.0, CPU offload enabled
-- Cases: 4 prompts (`portrait_01`, `portrait_02`, `scene_01`, `night_01`)
+- Inference setting: 1024×1024, 50 steps, guidance=4.0, CPU offload enabled
+- Cases: 5 prompts (`portrait_01`, `portrait_02`, `landscape_01`, `scene_01`, `night_01`)
 
 ### Aggregate Comparison (Baseline vs INT8)
 
 | Metric | Baseline | INT8 | Delta |
 |---|---:|---:|---:|
-| Avg elapsed / image (s) | 51.7766 | 39.5662 | **-23.6%** |
-| Avg sec / step | 1.8492 | 1.4131 | **-23.6%** |
+| Avg elapsed / image (s) | 49.0282 | 46.7867 | **-4.6%** |
+| Avg sec / step | 0.980564 | 0.935733 | **-4.6%** |
 | Avg peak CUDA alloc (GB) | 12.5195 | 7.7470 | **-38.1%** |
 
 
@@ -172,10 +175,11 @@ Test environment:
 
 | Case | Baseline (s) | INT8 (s) | Speedup |
 |---|---:|---:|---:|
-| portrait_01 | 99.9223 | 60.6768 | 1.65x |
-| portrait_02 | 37.4116 | 32.8863 | 1.14x |
-| scene_01 | 34.9946 | 32.2035 | 1.09x |
-| night_01 | 34.7780 | 32.4981 | 1.07x |
+| portrait_01 | 56.9943 | 50.1124 | 1.14x |
+| portrait_02 | 50.3810 | 46.0371 | 1.09x |
+| landscape_01 | 46.0286 | 46.0192 | 1.00x |
+| scene_01 | 45.9097 | 45.8291 | 1.00x |
+| night_01 | 45.8275 | 45.9356 | 1.00x |
 
 ## Visual Comparison (Baseline vs INT8)
 
@@ -184,7 +188,8 @@ Left: Baseline. Right: INT8. (Same prompt/seed/steps.)
 | Case | Base | INT8 |
 |---|---|---|
 | portrait_01 | ![](zimage_quanto_bench_results/images/baseline/portrait_01_seed46.png) | ![](zimage_quanto_bench_results/images/int8/portrait_01_seed46.png) |
-| portrait_02 | ![](zimage_quanto_bench_results/images/baseline/portrait_02_seed123.png) | ![](zimage_quanto_bench_results/images/int8/portrait_02_seed123.png) |
+| portrait_02 | ![](zimage_quanto_bench_results/images/baseline/portrait_02_seed111.png) | ![](zimage_quanto_bench_results/images/int8/portrait_02_seed111.png) |
+| landscape_01 | ![](zimage_quanto_bench_results/images/baseline/landscape_01_seed123.png) | ![](zimage_quanto_bench_results/images/int8/landscape_01_seed123.png) |
 | scene_01 | ![](zimage_quanto_bench_results/images/baseline/scene_01_seed777.png) | ![](zimage_quanto_bench_results/images/int8/scene_01_seed777.png) |
 | night_01 | ![](zimage_quanto_bench_results/images/baseline/night_01_seed2026.png) | ![](zimage_quanto_bench_results/images/int8/night_01_seed2026.png) |