Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,34 @@
+---
+license: apache-2.0
+base_model: Qwen/Qwen-Image-Edit-2511
+tags:
+- qwen-image
+- comfyui
+- int8
+- quantized
+- image-editing
+---
+
+# Qwen-Image-Edit-2511 — INT8 tensorwise (+ConvRot), stock-ComfyUI native
+
+Offline INT8 quantization of the `Qwen/Qwen-Image-Edit-2511` transformer, produced
+with [Comfy-Org/comfy-quants](https://github.com/Comfy-Org/comfy-quants)
+(`export-model-int8-tensorwise`). Loads **natively in stock ComfyUI >= v0.27.0**
+(`QUANT_ALGOS["int8_tensorwise"]`, SM >= 7.5 / Turing+) — no custom node required.
+
+Transformer-only (for `models/diffusion_models/`, load via `UNETLoader`); pair with
+the usual Qwen-Image-Edit text encoder and VAE.
+
+## Quantization contract
+
+- 839 quantized Linears (selection mirrors the family's official recipe — same
+  membership as the mxfp8/nvfp4 exports; `time_text_embed`/`img_in`/`norm_out`/
+  `proj_out`/`txt_in` kept bf16); all 839 ConvRot-rotated (regular Hadamard,
+  group 256).
+- Per layer: `int8` weight + `float32 [out,1]` scale + `comfy_quant` marker
+  `{"format": "int8_tensorwise", "convrot": true, "convrot_groupsize": 256}`
+  (byte-exact to stock ComfyUI's save path).
+- Includes the `__index_timestep_zero__` sentinel required for correct 2511
+  reference-image conditioning.
+- Weight SQNR (comfy-kitchen dequant, original basis): ~40.6 dB.
+- Quant math bit-faithful to comfy-kitchen >= 0.2.15.