Add SDMatte model files and scripts

README.md

@@ -0,0 +1,65 @@
+# SDMatte_plus-fp16-and-bf16
+
+This repository provides optimized, inference-only versions of the original [SDMatte model by LongfeiHuang](https://huggingface.co/LongfeiHuang/SDMatte).
+
+The models here have been specifically processed to be lightweight and efficient for deployment and use in applications like ComfyUI, without compromising the quality of the matting results.
+
+### What is this?
+
+This repository contains inference-only weights for the **SDMatte** model. The original checkpoint file (`.pth`) was a full training checkpoint, which included not only the model weights but also ~6.5 GB of trainer states (like optimizer states). These trainer states are crucial for resuming training but are unnecessary for performing inference (i.e., actually using the model for matting).
+
+### Optimizations Performed
+
+1.  **Removal of Trainer States**: The largest optimization was stripping the `trainer` key from the original checkpoint. This removes all unnecessary data related to the training process, significantly reducing the file size without affecting the model's output.
+
+2.  **16-bit Precision Quantization**: The model weights have been converted from their original 32-bit floating-point precision (FP32) to 16-bit precision. We provide two popular formats:
+    * **FP16 (half-precision)**: Offers a great balance of speed, reduced memory usage, and high quality. It is supported by most modern NVIDIA GPUs (10-series and newer).
+    * **BF16 (bfloat16)**: Offers a dynamic range identical to FP32, making it more resilient to overflow/underflow issues. It provides the best performance on newer NVIDIA GPUs (RTX 30-series and newer).
+
+### DIY Quantization with `convert_precision.py`
+
+This repository also includes the Python script, `convert_precision.py`, which was used to create these fp16 and bf16 models. You can use this script to convert the original FP32 checkpoint yourself.
+
+1.  Place the original FP32 `SDMatte_plus.pth` file in the same folder as the script.
+2.  Open the `convert_precision.py` file with a text editor.
+3.  Modify the `TARGET_PRECISION` variable at the top to either `'fp16'` or `'bf16'`.
+4.  Run the script from your terminal: `python convert_precision.py`.
+
+### Acknowledgements
+
+Huge thanks to **LongfeiHuang** for creating and open-sourcing the original SDMatte model. This repository is merely an optimized packaging of their incredible work. Please visit the original repository for more details on the model's architecture and training.
+
+* **Original Model**: [https://huggingface.co/LongfeiHuang/SDMatte](https://huggingface.co/LongfeiHuang/SDMatte)
+
+---
+
+## 中文版本
+
+### 这是什么？
+
+本仓库提供原始 [SDMatte 模型（作者 LongfeiHuang）](https://huggingface.co/LongfeiHuang/SDMatte) 的优化版，仅用于推理。
+
+这里的模型都经过了专门处理，旨在使其轻量化且高效，以方便在 ComfyUI 等应用中部署和使用，同时不影响抠图效果的质量。
+
+### 执行的优化
+
+1.  **移除 Trainer 状态**：最大的优化是剥离了原始检查点中的 `trainer` 键。这移除了所有与训练过程相关的非必要数据（如优化器状态），在完全不影响模型输出质量的前提下，极大地减小了文件体积。原始文件中约 6.5 GB 的数据都属于此类。
+
+2.  **16 位精度量化**：模型权重已从原始的 32 位浮点精度 (FP32) 转换为 16 位精度。我们提供了两种主流格式：
+    * **FP16 (半精度)**：在速度、显存占用和高质量之间取得了绝佳的平衡。它被大多数现代 NVIDIA GPU（10 系及更新）所支持。
+    * **BF16 (bfloat16)**：拥有与 FP32 相同的动态范围，使其在处理数据溢出/下溢问题时更具弹性。它在较新的 NVIDIA GPU（RTX 30 系及更新）上能提供最佳性能。
+
+### 使用 `convert_precision.py` 自行量化
+
+本仓库同样包含了用于创建这些 fp16 和 bf16 模型的 Python 脚本 `convert_precision.py`。您可以使用此脚本，自行将原始的 FP32 检查点进行转换。
+
+1.  将原始的 FP32 `SDMatte_plus.pth` 文件放置于脚本所在的同一个文件夹内。
+2.  使用文本编辑器打开 `convert_precision.py` 文件。
+3.  修改文件顶部的 `TARGET_PRECISION` 变量，将其设置为 `'fp16'` 或 `'bf16'`。
+4.  在您的终端中运行脚本：`python convert_precision.py`。
+
+### 致谢
+
+非常感谢 **LongfeiHuang** 创建并开源了卓越的 SDMatte 模型。本仓库的工作仅仅是对其出色成果的优化和打包。有关模型架构和训练的更多详细信息，请访问原始仓库。
+
+* **原始模型仓库**：[https://huggingface.co/LongfeiHuang/SDMatte](https://huggingface.co/LongfeiHuang/SDMatte)

SDMatte_plus_bf16_inference.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f04e5cb6a50af6516cea8908ed2de77d5ae2753b33b6e4ea0cbb61146a27f55
+size 2594696326

SDMatte_plus_fp16_inference.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f407176bc081bd58636c472fa150891559d9bb24ae00399e9d86254c4624572
+size 2594696326

convert_precision.py

@@ -0,0 +1,56 @@
+import torch
+import os
+
+# --- 配置 ---
+# 1. 设置目标精度: 'fp16' 或 'bf16'
+TARGET_PRECISION = 'bf16' 
+
+# 2. 设置原始32位训练检查点文件的路径
+fp32_checkpoint_path = r"E:\comfyui\ComfyUI-aki-v1.3\models\SDMatte\SDMatte_plus.pth"
+
+# --------------------------------------------------------------------
+
+# 自动生成输出文件名
+output_filename = fp32_checkpoint_path.replace('.pth', f'_{TARGET_PRECISION}_inference.pth')
+
+if not os.path.exists(fp32_checkpoint_path):
+    print(f"[错误] 文件不存在: {fp32_checkpoint_path}")
+else:
+    try:
+        print(f"--- 开始处理训练检查点: {fp32_checkpoint_path} ---")
+        full_checkpoint = torch.load(fp32_checkpoint_path, map_location="cpu", weights_only=False)
+
+        # 检查 'model' 键是否存在，这是包含我们所需权重的部分
+        if 'model' in full_checkpoint:
+            # 明确提取出模型的 state_dict
+            state_dict = full_checkpoint['model']
+            print("成功提取到 'model' 键中的权重字典。")
+        else:
+            # 如果没有 'model' 键，则假定整个文件就是 state_dict
+            print("[警告] 未在顶层找到 'model' 键，将尝试转换整个文件。")
+            state_dict = full_checkpoint
+        
+        print(f"开始将权重转换为 {TARGET_PRECISION} ...")
+        
+        target_dtype = torch.float16 if TARGET_PRECISION == 'fp16' else torch.bfloat16
+
+        # 遍历权重字典中的每一项并进行转换
+        for key in state_dict:
+            if isinstance(state_dict[key], torch.Tensor) and state_dict[key].is_floating_point():
+                state_dict[key] = state_dict[key].to(target_dtype)
+
+        print(f"正在保存纯推理模型到: {output_filename} ...")
+        # 直接保存处理后的 state_dict，不包含任何训练相关的附加信息
+        torch.save(state_dict, output_filename)
+        
+        # 打印大小对比
+        original_size_gb = os.path.getsize(fp32_checkpoint_path) / (1024**3)
+        final_size_gb = os.path.getsize(output_filename) / (1024**3)
+
+        print("\n--- 转换成功 ---")
+        print(f"原始训练检查点大小: {original_size_gb:.2f} GB")
+        print(f"生成的纯推理模型大小 ({TARGET_PRECISION.upper()}): {final_size_gb:.2f} GB")
+        print("说明: 新文件只包含用于推理的核心模型权重，已移除训练相关的优化器状态。")
+
+    except Exception as e:
+        print(f"\n[错误] 处理过程中发生错误: {e}")