docs: honest model card (validation status + fleet-specific perf caveat)

README.md

@@ -17,57 +17,29 @@ tags:
   - 4-bit
 ---
 
-# Qwen-Image — fni8 (int4)
+# Qwen-Image — fni8 (int8/W4A8 dp4a, Volta sm_70)
 
-An int8/int4 quantization of [`Qwen/Qwen-Image`](https://huggingface.co/Qwen/Qwen-Image)
-to the `.fni8` format, for the [fni8](https://github.com/jajmangold/fni8) DP4A kernels
-on NVIDIA Volta (sm_70) GPUs (Tesla V100 and CMP 100-210). It is a derivative of the
-parent model; its license and acceptable uses follow the parent, linked above.
+Diffusion transformer (DiT), quantized from [`Qwen/Qwen-Image`](https://huggingface.co/Qwen/Qwen-Image). Repackaged to the **`.fni8`** resident format (~31.4 GB (int4 DiT)) for the [fni8](https://github.com/jajmangold/fni8) **W8A8/W4A8 DP4A** kernels on **NVIDIA Volta (sm_70)** — Tesla V100 / CMP 100-210.
 
-## Files
+## Status
 
-| file | precision | scheme | size |
-|------|-----------|--------|------|
-| `Qwen__Qwen-Image.dit.b4.fni8` | int4 | int4 per-group W4A8 (int8 activations) | 10.92 GB |
+**Validated:** end-to-end ComfyUI-fni8 sanity pass (a one-step output that is finite and non-constant — a sanity check, not a quality benchmark). **Performance is fleet-specific.** All fni8 speedups are measured on the CMP 100-210 mining-card fleet, where the fp16 tensor cores are firmware-gimped. These numbers do **not** transfer to a real Tesla V100 (whose fp16 tensor cores would beat dp4a).
 
-`.b8.` files are int8, `.b4.` files are int4. Download the one you want. The bytes are
-the resident dp4a VRAM layout, so loading is a memory-map and copy with no dequantize
-or repack step.
+## Format
 
-## How to use
+- **Weights:** int4 per-group W4A8 (int8 activations), fp32 scales, resident dp4a VRAM layout.
+- **Why dp4a:** sm_70 has no int8 tensor cores, so the matmul contraction runs on the `__dp4a` CUDA-core intrinsic. On the CMP 100-210 fleet (whose fp16 tensor cores are firmware-limited) dp4a is the fast path, not a compromise.
 
-Load with [ComfyUI-fni8](https://github.com/jajmangold/ComfyUI-fni8), which runs the diffusion transformer through the dp4a kernels inside ComfyUI (`UnetLoaderFNI8`). The text encoder and VAE are unchanged.
+## How to run
 
-## What was quantized
-
-- Linear and attention weights go to int8 (per-row) or int4 (per-group), with fp32
-  scales. Norms, embeddings, and the MoE router are kept in fp16.
-- Source dtype: `bfloat16`. Non-quantized tensors are stored fp16, upcast to fp32 only where fp16 would overflow.
-- Target hardware is sm_70, where the fp16 tensor cores are firmware-limited, so the
-  integer `__dp4a` path is used for the matmuls.
-
-## Intended use and scope
-
-- For inference with the fni8 runtimes above, on Volta (sm_70) GPUs.
-- Out of scope: other GPU architectures (the kernels require sm_70), and anything the
-  parent model's license does not permit. It is a derivative, not a new model.
+[ComfyUI-fni8](https://github.com/jajmangold/ComfyUI-fni8) is the runtime (`UnetLoaderFNI8` runs the diffusion transformer through the dp4a kernels; the text encoder and VAE are unchanged). fni8-serve is for LLMs only and does not load this model.
 
 ## Limitations
 
-- Quantization is lossy. int8 and especially int4 outputs differ from the fp16/bf16
-  parent, and the difference varies by model and task.
-- This repository does not include per-model accuracy or benchmark measurements.
-  Evaluate on your own task before relying on it.
-- Any capabilities, biases, and risks of the parent model carry over. See the parent
-  model card for those.
-
-## License
-
-Follows the parent model (`apache-2.0`). This is a
-derivative quantization, not a relicense.
+- Quantization is lossy: int8 (and especially int4) outputs differ from the fp16/bf16 parent, and the difference varies by task.
+- Capabilities, biases, and risks of the parent model carry over — see the parent card.
+- This is a derivative quantization, not a relicense; the parent model's license and acceptable uses apply.
 
 ---
 
-Part of the fni8 stack: [kernels](https://github.com/jajmangold/fni8) ·
-[LLM serving](https://github.com/jajmangold/fni8-serve) ·
-[ComfyUI DiTs](https://github.com/jajmangold/ComfyUI-fni8).
+Part of the fni8 stack: [kernels](https://github.com/jajmangold/fni8) · [LLM serving](https://github.com/jajmangold/fni8-serve) · [ComfyUI DiTs](https://github.com/jajmangold/ComfyUI-fni8).