diff --git a/.gitattributes b/.gitattributes index a6344aac8c09253b3b630fb776ae94478aa0275b..4b245b9bfccb5a14de955d36d48997cb8fa55531 100644 --- a/.gitattributes +++ b/.gitattributes @@ -1,35 +1,75 @@ *.7z filter=lfs diff=lfs merge=lfs -text *.arrow filter=lfs diff=lfs merge=lfs -text *.bin 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 +*.zstandard filter=lfs diff=lfs merge=lfs -text +*.tfevents* filter=lfs diff=lfs merge=lfs -text +*.db* filter=lfs diff=lfs merge=lfs -text +*.ark* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.gguf* filter=lfs diff=lfs merge=lfs -text +*.ggml filter=lfs diff=lfs merge=lfs -text +*.llamafile* filter=lfs diff=lfs merge=lfs -text +*.pt2 filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text *.zst filter=lfs diff=lfs merge=lfs -text -*tfevents* filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -textassets/minimax_h3_fl2va_dit_4bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_fl2va_dit_5bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_fl2va_dit_6bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_fl2va_dit_7bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_fl2va_dit_8bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_fl2va_dit_extreme_3.0bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_fl2va_dit_fp8_5bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_ref2va_dit_4bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_ref2va_dit_5bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_ref2va_dit_6bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_ref2va_dit_7bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_ref2va_dit_8bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_ref2va_dit_extreme_3.0bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_ref2va_dit_fp8_5bpp.mp4 filter=lfs diff=lfs merge=lfs -text +assets/minimax_h3_ref2va_reference_woman.png filter=lfs diff=lfs merge=lfs -text +assets/qwen_image_2.1_dit_4bpp.png filter=lfs diff=lfs merge=lfs -text +assets/qwen_image_2.1_dit_5bpp.png filter=lfs diff=lfs merge=lfs -text +assets/qwen_image_2.1_dit_6bpp.png filter=lfs diff=lfs merge=lfs -text +assets/qwen_image_2.1_dit_7bpp.png filter=lfs diff=lfs merge=lfs -text +assets/qwen_image_2.1_dit_8bpp.png filter=lfs diff=lfs merge=lfs -text +assets/qwen_image_2.1_dit_extreme_3.0bpp.png filter=lfs diff=lfs merge=lfs -text +assets/qwen_image_2.1_dit_fp8_5bpp.png filter=lfs diff=lfs merge=lfs -text +assets/z_image_turbo_dit_4bpp.jpg filter=lfs diff=lfs merge=lfs -text +assets/z_image_turbo_dit_5bpp.jpg filter=lfs diff=lfs merge=lfs -text +assets/z_image_turbo_dit_6bpp.jpg filter=lfs diff=lfs merge=lfs -text +assets/z_image_turbo_dit_7bpp.jpg filter=lfs diff=lfs merge=lfs -text +assets/z_image_turbo_dit_8bpp.jpg filter=lfs diff=lfs merge=lfs -text +assets/z_image_turbo_dit_extreme_2.3bpp.jpg filter=lfs diff=lfs merge=lfs -text +assets/z_image_turbo_dit_fp8_5bpp.jpg filter=lfs diff=lfs merge=lfs -text diff --git a/README.md b/README.md new file mode 100644 index 0000000000000000000000000000000000000000..9ae08fa38ae9e22a76ca2d599fcf9db3828bc259 --- /dev/null +++ b/README.md @@ -0,0 +1,224 @@ +# EntroPack Pre-Quantized Models + +This repository releases pre-quantized weight packages produced by [**EntroPack**](https://github.com/modelscope/entropack) for representative SOTA diffusion models, to be used with the quantization features of [**DiffSynth-Studio**](https://github.com/modelscope/DiffSynth-Studio): + +- **Z-Image-Turbo** (DiT + text encoder) +- **Qwen-Image-2.1** (DiT + text encoder) +- **MiniMax-H3** (two DiTs, FL2VA / Ref2VA, sharing the text encoder and video VAE) + +For each model, both the **DiT and the text encoder** are provided at five fine-grained compression rates (4 / 5 / 6 / 7 / 8 bpp), plus two kinds of specialized packages: + +- **`fp8@5`** (`dit_fp8_5bpp`): the weights are first **FP8 W8A8-quantized** (8-bit fp8 codes), then entropy-coded by EntroPack down to **5-bit storage** (actual bpp = 5); the forward pass uses **FP8 GEMM** for faster inference; +- **Extreme mixed-bit packages**: bit-rate is allocated per layer by sensitivity (the most sensitive layers keep a high bpp while the rest use a low bpp), pushing the whole model to a much lower bpp (Z-Image-Turbo 2.3bpp / Qwen-Image-2.1 3.0bpp / MiniMax-H3 3.0bpp). + +All packages are built by entropy-coding the original weights directly with EntroPack; the metrics in the tables are weighted measurements on the package artifacts: + +- **bpp** (bits per parameter): average bits per weight, `bpp = 8 × Σ storage bytes / Σ numel`. +- **Linear weight compression ratio**: derived from bpp, `ratio = bpp / 16`. +- **Weight reconstruction error** (DiT only): relative deviation between the dequantized and the original weights, `sqrt(Σ‖Wq−W‖²) / sqrt(Σ‖W‖²)`. +- **Registration entry**: each `*.safetensors` ships with a same-named `*.json`; see the end of this document for loading. + +Validation outputs (images / videos) for each model can be found under `assets/`. + +## Update History + +**2026-10-15 — checkpoints repacked under new key names.** All packages in this repository were repacked +into the single-entry container layout: a compressed layer's whole encoding now lives under one +`.weight.entropack_quant_state` tensor, and the W8A8 row scale under `.weight_quant_scale`. +The **updated** weights require **entropack 0.1.3**; the packages published before this update pair with +**entropack ≤ 0.1.2**. + +## Z-Image-Turbo + +### DiT + +| name | file | actual bpp | weight reconstruction error | size | +|---|---|---|---|---| +| dit_4bpp | `models/Z-Image-Turbo/dit_4bpp.safetensors` | 4.024 | 7.013% | 2.88 GB | +| dit_5bpp | `models/Z-Image-Turbo/dit_5bpp.safetensors` | 5.046 | 3.630% | 3.62 GB | +| dit_6bpp | `models/Z-Image-Turbo/dit_6bpp.safetensors` | 6.055 | 1.854% | 4.34 GB | +| dit_7bpp | `models/Z-Image-Turbo/dit_7bpp.safetensors` | 7.070 | 0.971% | 5.07 GB | +| dit_8bpp | `models/Z-Image-Turbo/dit_8bpp.safetensors` | 8.060 | 0.532% | 5.78 GB | +| dit_fp8_5bpp | `models/Z-Image-Turbo/dit_fp8_5bpp.safetensors` | 5.041 | 4.648% | 3.62 GB | +| dit_extreme_2.3bpp | `models/Z-Image-Turbo/dit_extreme_2.3bpp.safetensors` | 2.319 | 24.007% | 1.66 GB | +| nf4 (bitsandbytes) (for comparison) | - | 4.127 | 9.415% | - | +| torchao int8 w8a16 (for comparison) | - | 8.012 | 1.078% | - | +| torchao fp8 w8a16 (for comparison) | - | 8.010 | 2.650% | - | + +### Text Encoder + +| name | file | actual bpp | size | +|---|---|---|---| +| text_encoder_4bpp | `models/Z-Image-Turbo/text_encoder_4bpp.safetensors` | 4.028 | 2.43 GB | +| text_encoder_5bpp | `models/Z-Image-Turbo/text_encoder_5bpp.safetensors` | 5.051 | 2.86 GB | +| text_encoder_6bpp | `models/Z-Image-Turbo/text_encoder_6bpp.safetensors` | 6.058 | 3.29 GB | +| text_encoder_7bpp | `models/Z-Image-Turbo/text_encoder_7bpp.safetensors` | 7.061 | 3.71 GB | +| text_encoder_8bpp | `models/Z-Image-Turbo/text_encoder_8bpp.safetensors` | 8.041 | 4.13 GB | + +### Extreme Mixed-Bit Package + +The extreme package `dit_extreme_2.3bpp` uses mixed quantization (per-layer-group bit allocation, weighted target = 2.3bpp): modulation / embedding / adaLN / final layers (38 layers) @ **4.0bpp**, attention output projections `*.attention.to_out.0` (34 layers) @ **3.0bpp**, and the remaining 204 layers @ **2.1939bpp**. Measured 2.319bpp / weight reconstruction error 24.0%. + +## Qwen-Image-2.1 + +### DiT + +| name | file | actual bpp | weight reconstruction error | size | +|---|---|---|---|---| +| dit_4bpp | `models/Qwen-Image-2.1/dit_4bpp.safetensors` | 4.027 | 7.064% | 3.34 GB | +| dit_5bpp | `models/Qwen-Image-2.1/dit_5bpp.safetensors` | 5.052 | 3.658% | 4.19 GB | +| dit_6bpp | `models/Qwen-Image-2.1/dit_6bpp.safetensors` | 6.066 | 1.868% | 5.02 GB | +| dit_7bpp | `models/Qwen-Image-2.1/dit_7bpp.safetensors` | 7.085 | 0.978% | 5.87 GB | +| dit_8bpp | `models/Qwen-Image-2.1/dit_8bpp.safetensors` | 8.068 | 0.536% | 6.68 GB | +| dit_fp8_5bpp | `models/Qwen-Image-2.1/dit_fp8_5bpp.safetensors` | 5.048 | 4.662% | 4.19 GB | +| dit_extreme_3.0bpp | `models/Qwen-Image-2.1/dit_extreme_3.0bpp.safetensors` | 3.023 | 14.247% | 2.50 GB | +| nf4 (bitsandbytes) (for comparison) | - | 4.127 | 9.428% | - | +| torchao int8 w8a16 (for comparison) | - | 8.008 | 1.095% | - | +| torchao fp8 w8a16 (for comparison) | - | 8.007 | 2.648% | - | + +### Text Encoder + +| name | file | actual bpp | size | +|---|---|---|---| +| text_encoder_4bpp | `models/Qwen-Image-2.1/text_encoder_4bpp.safetensors` | 4.034 | 4.99 GB | +| text_encoder_5bpp | `models/Qwen-Image-2.1/text_encoder_5bpp.safetensors` | 5.062 | 5.97 GB | +| text_encoder_6bpp | `models/Qwen-Image-2.1/text_encoder_6bpp.safetensors` | 6.082 | 6.93 GB | +| text_encoder_7bpp | `models/Qwen-Image-2.1/text_encoder_7bpp.safetensors` | 7.092 | 7.89 GB | +| text_encoder_8bpp | `models/Qwen-Image-2.1/text_encoder_8bpp.safetensors` | 8.072 | 8.82 GB | + +### Extreme Mixed-Bit Package + +The extreme package `dit_extreme_3.0bpp` uses mixed quantization (weighted target = 3.0bpp): modulation-related layers (`modulation.1`, `norm_out.linear`, `time_text_embed.*`, 4 layers) @ **4.0bpp**, and the remaining 228 layers @ **2.9855bpp**. Measured 3.023bpp / weight reconstruction error 14.2%. + + +## MiniMax-H3 + +### DiT + +| name | file | actual bpp | weight reconstruction error | size | +|---|---|---|---|---| +| fl2va_dit_4bpp | `models/MiniMax-H3/FL2VA/dit_4bpp.safetensors` | 4.038 | 6.567% | 15.58 GB | +| ref2va_dit_4bpp | `models/MiniMax-H3/REF2VA/dit_4bpp.safetensors` | 4.038 | 6.567% | 15.58 GB | +| fl2va_dit_5bpp | `models/MiniMax-H3/FL2VA/dit_5bpp.safetensors` | 5.065 | 3.392% | 19.54 GB | +| ref2va_dit_5bpp | `models/MiniMax-H3/REF2VA/dit_5bpp.safetensors` | 5.065 | 3.392% | 19.54 GB | +| fl2va_dit_6bpp | `models/MiniMax-H3/FL2VA/dit_6bpp.safetensors` | 6.081 | 1.734% | 23.45 GB | +| ref2va_dit_6bpp | `models/MiniMax-H3/REF2VA/dit_6bpp.safetensors` | 6.082 | 1.732% | 23.46 GB | +| fl2va_dit_7bpp | `models/MiniMax-H3/FL2VA/dit_7bpp.safetensors` | 7.093 | 0.911% | 27.36 GB | +| ref2va_dit_7bpp | `models/MiniMax-H3/REF2VA/dit_7bpp.safetensors` | 7.095 | 0.908% | 27.37 GB | +| fl2va_dit_8bpp | `models/MiniMax-H3/FL2VA/dit_8bpp.safetensors` | 8.081 | 0.500% | 31.16 GB | +| ref2va_dit_8bpp | `models/MiniMax-H3/REF2VA/dit_8bpp.safetensors` | 8.079 | 0.500% | 31.16 GB | +| fl2va_dit_fp8_5bpp | `models/MiniMax-H3/FL2VA/dit_fp8_5bpp.safetensors` | 5.058 | 4.510% | 19.54 GB | +| ref2va_dit_fp8_5bpp | `models/MiniMax-H3/REF2VA/dit_fp8_5bpp.safetensors` | 5.058 | 4.510% | 19.54 GB | +| fl2va_dit_extreme_3.0bpp | `models/MiniMax-H3/FL2VA/dit_extreme_3.0bpp.safetensors` | 3.026 | 13.225% | 11.68 GB | +| ref2va_dit_extreme_3.0bpp | `models/MiniMax-H3/REF2VA/dit_extreme_3.0bpp.safetensors` | 3.026 | 13.225% | 11.68 GB | +| nf4 (bitsandbytes) (for comparison) | - | 4.127 | 9.151% | - | +| torchao int8 w8a16 (for comparison) | - | 8.010 | 0.984% | - | +| torchao fp8 w8a16 (for comparison) | - | 8.008 | 2.642% | - | + +### Text Encoder + +| name | file | actual bpp | size | +|---|---|---|---| +| text_encoder_4bpp | `models/MiniMax-H3/text_encoder_4bpp.safetensors` | 4.039 | 13.20 GB | +| text_encoder_5bpp | `models/MiniMax-H3/text_encoder_5bpp.safetensors` | 5.073 | 16.21 GB | +| text_encoder_6bpp | `models/MiniMax-H3/text_encoder_6bpp.safetensors` | 6.102 | 19.20 GB | +| text_encoder_7bpp | `models/MiniMax-H3/text_encoder_7bpp.safetensors` | 7.127 | 22.18 GB | +| text_encoder_8bpp | `models/MiniMax-H3/text_encoder_8bpp.safetensors` | 8.095 | 24.99 GB | + +### Video VAE + +| name | file | actual bpp | size | +|---|---|---|---| +| video_vae_4bpp | `models/MiniMax-H3/video_vae_4bpp.safetensors` | 4.020 | 1.47 GB | +| video_vae_5bpp | `models/MiniMax-H3/video_vae_5bpp.safetensors` | 5.040 | 1.76 GB | +| video_vae_6bpp | `models/MiniMax-H3/video_vae_6bpp.safetensors` | 6.045 | 2.04 GB | +| video_vae_7bpp | `models/MiniMax-H3/video_vae_7bpp.safetensors` | 7.055 | 2.33 GB | +| video_vae_8bpp | `models/MiniMax-H3/video_vae_8bpp.safetensors` | 8.056 | 2.61 GB | + +### Extreme Mixed-Bit Package + +The extreme package `dit_extreme_3.0bpp` (identical configuration for FL2VA / Ref2VA) uses mixed quantization (weighted target = 3.0bpp): low-dimensional patch/final modules (`time_embedder.proj_in/proj_out`, `video/audio_patch_proj`, `condition_proj`, `final_layer.video_out/audio_out`, 7 layers) @ **8.0bpp**, modulation (`*.adaln_proj.linear`, 51 layers) and `attn.out_proj` (52 layers) @ **3.0bpp**, and the remaining 156 layers @ **2.9876bpp**. Measured 3.026bpp / weight reconstruction error 13.2%. + +## Usage + +### Install from Source + +```bash +# DiffSynth-Studio +git clone https://github.com/modelscope/DiffSynth-Studio.git +cd DiffSynth-Studio && pip install -e . + +# EntroPack (pick the extra matching your CUDA version: cuda13 / cuda12) +git clone https://github.com/modelscope/entropack.git +cd entropack && pip install -e '.[cuda13]' +``` + +### Example Script (Z-Image-Turbo) + +```python +""" +Z-Image-Turbo — EntroPack pre-quantized inference example. + +Downloads the pre-quantized packages and their json sidecars from the +`DiffSynth-Studio/EntroPackPreQuants` model repo (each json is a ready-to-use +`MODEL_CONFIGS` entry), then runs inference. + +Available packages (paths inside the prequant model repo): + + dit: + models/Z-Image-Turbo/dit_4bpp.safetensors 4bpp + models/Z-Image-Turbo/dit_5bpp.safetensors 5bpp + models/Z-Image-Turbo/dit_6bpp.safetensors 6bpp + models/Z-Image-Turbo/dit_7bpp.safetensors 7bpp + models/Z-Image-Turbo/dit_8bpp.safetensors 8bpp + models/Z-Image-Turbo/dit_extreme_2.3bpp.safetensors mixed allocation, 2.3bpp + models/Z-Image-Turbo/dit_fp8_5bpp.safetensors fp8 weights, 5bpp + text_encoder: + models/Z-Image-Turbo/text_encoder_4bpp.safetensors 4bpp + models/Z-Image-Turbo/text_encoder_5bpp.safetensors 5bpp + models/Z-Image-Turbo/text_encoder_6bpp.safetensors 6bpp + models/Z-Image-Turbo/text_encoder_7bpp.safetensors 7bpp + models/Z-Image-Turbo/text_encoder_8bpp.safetensors 8bpp + +`dit_package` and `text_encoder_package` are chosen independently: any combination works +(e.g. `dit_8bpp` with `text_encoder_4bpp`). +""" +import json + +import torch + +from diffsynth.configs import MODEL_CONFIGS +from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig + +prequant_model_id = "DiffSynth-Studio/EntroPackPreQuants" +dit_package = "models/Z-Image-Turbo/dit_4bpp.safetensors" +text_encoder_package = "models/Z-Image-Turbo/text_encoder_4bpp.safetensors" + + +def register_prequantized(package_pattern): + """Download the pre-quantized package and its json sidecar (a ready-to-use MODEL_CONFIGS entry), return a loadable ModelConfig.""" + package = ModelConfig(model_id=prequant_model_id, origin_file_pattern=package_pattern) + package.download_if_necessary() + sidecar = ModelConfig(model_id=prequant_model_id, origin_file_pattern=package_pattern.replace(".safetensors", ".json")) + sidecar.download_if_necessary() + with open(sidecar.path) as handle: + MODEL_CONFIGS.append(json.load(handle)) + return package + + +pipe = ZImagePipeline.from_pretrained( + torch_dtype=torch.bfloat16, + device="cuda", + model_configs=[ + register_prequantized(dit_package), + register_prequantized(text_encoder_package), + ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"), + ], + tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"), +) +prompt = "Young Chinese woman in red Hanfu, intricate embroidery. Impeccable makeup, red floral forehead pattern. Elaborate high bun, golden phoenix headdress, red flowers, beads. Holds round folding fan with lady, trees, bird. Neon lightning-bolt lamp (⚡️), bright yellow glow, above extended left palm. Soft-lit outdoor night background, silhouetted tiered pagoda (西安大雁塔), blurred colorful distant lights." +image = pipe(prompt=prompt, seed=42, rand_device="cuda") +image.save("image_Z-Image-Turbo-EntroPack.jpg") +``` + +`dit_package` and `text_encoder_package` can be chosen independently in any combination; the extreme and fp8 packages work in exactly the same way — just point to the corresponding `.safetensors` path (mixed-quantization `configs` and other details live in the same-named `.json`, no script changes needed). For Qwen-Image-2.1 and MiniMax-H3, swap in the corresponding pipeline and package paths. diff --git a/assets/minimax_h3_fl2va_dit_4bpp.mp4 b/assets/minimax_h3_fl2va_dit_4bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..9437b60b71096e7ba4f3a6efae63958ccbc975c7 --- /dev/null +++ b/assets/minimax_h3_fl2va_dit_4bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f3c820132ea422b6c6aeea52ec2133f769bdb1e1307edaa66c0bb6e6d8457db1 +size 20493430 diff --git a/assets/minimax_h3_fl2va_dit_5bpp.mp4 b/assets/minimax_h3_fl2va_dit_5bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..2aedbe3b6bebdc3ba9909a0f5b1b47b44acb6d91 --- /dev/null +++ b/assets/minimax_h3_fl2va_dit_5bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f40aa364480b35a38ee26f856b4f6fdc9161a87362cb78c2ffb572e0519ef8c1 +size 19141428 diff --git a/assets/minimax_h3_fl2va_dit_6bpp.mp4 b/assets/minimax_h3_fl2va_dit_6bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..995d8a4ce252adb56a6ca1315a8733589c23e565 --- /dev/null +++ b/assets/minimax_h3_fl2va_dit_6bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:47df7cc3f4780d25fc5671b275a2bdd1b832feadf55d695b8f1f3dd8f6535723 +size 17279249 diff --git a/assets/minimax_h3_fl2va_dit_7bpp.mp4 b/assets/minimax_h3_fl2va_dit_7bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..88dc268b9a01de8459b819dc0b0dda11435d2099 --- /dev/null +++ b/assets/minimax_h3_fl2va_dit_7bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1544df8e9160996b53b33c777c1f6d9179fbf7129fd0797c07653a4465b23f88 +size 16816039 diff --git a/assets/minimax_h3_fl2va_dit_8bpp.mp4 b/assets/minimax_h3_fl2va_dit_8bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..eb98cbfb9f299f70f6ea7e81d46e4732a220e77d --- /dev/null +++ b/assets/minimax_h3_fl2va_dit_8bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:881b4691e16ac0b5253fb1292fb70a0d2fea43b0254a44e9ac0e02f673f3a415 +size 16122668 diff --git a/assets/minimax_h3_fl2va_dit_extreme_3.0bpp.mp4 b/assets/minimax_h3_fl2va_dit_extreme_3.0bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..ce672fe78782fcf8d8e2a172bb7e100011c36d15 --- /dev/null +++ b/assets/minimax_h3_fl2va_dit_extreme_3.0bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9c7a1986fba2db7fadb0950a2749ee77fe616210ae7fbffb03742b09141de68d +size 29696339 diff --git a/assets/minimax_h3_fl2va_dit_fp8_5bpp.mp4 b/assets/minimax_h3_fl2va_dit_fp8_5bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..cfed2109a85b12589dc82a2b551e4bedbb55de6c --- /dev/null +++ b/assets/minimax_h3_fl2va_dit_fp8_5bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d1897a3d014bee8edd56ac58a76552deb6ec2403026bcd97ddb6a5f1e4619ad4 +size 16393608 diff --git a/assets/minimax_h3_ref2va_dit_4bpp.mp4 b/assets/minimax_h3_ref2va_dit_4bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..854557f8736c32c0fe7006b227f5830c6555c097 --- /dev/null +++ b/assets/minimax_h3_ref2va_dit_4bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fbd52567fa0c8a2c0146ac090e0963f9f0e92c8b24fd6db5bd757bcebcf3cea0 +size 24277888 diff --git a/assets/minimax_h3_ref2va_dit_5bpp.mp4 b/assets/minimax_h3_ref2va_dit_5bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..48317596dcb6c0192c91945b613b1fbfb859a066 --- /dev/null +++ b/assets/minimax_h3_ref2va_dit_5bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7b559bf47f0a89f876df19815f9a02285324cc3791e215e299b3baf320cb1a81 +size 17917329 diff --git a/assets/minimax_h3_ref2va_dit_6bpp.mp4 b/assets/minimax_h3_ref2va_dit_6bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..f850b948fdb758d8745f7579d5e0a3240cf8dfec --- /dev/null +++ b/assets/minimax_h3_ref2va_dit_6bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9fbd1bff45719c41fd9eb9fd70f736899b4b1e571cb5aa56cb4cf9b0b5e16a43 +size 13943028 diff --git a/assets/minimax_h3_ref2va_dit_7bpp.mp4 b/assets/minimax_h3_ref2va_dit_7bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..8ea6bdc20a706ea743d21792f8029a87fcafcef2 --- /dev/null +++ b/assets/minimax_h3_ref2va_dit_7bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:40b63d9cbd1041c46a483eefc1350d91ccffdc2c7407a630c3c32ea35988df22 +size 12921232 diff --git a/assets/minimax_h3_ref2va_dit_8bpp.mp4 b/assets/minimax_h3_ref2va_dit_8bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..1b2e4c3c264e5b6ba16f3589f3ded0f8ac3132f0 --- /dev/null +++ b/assets/minimax_h3_ref2va_dit_8bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9b240f3b569cb1d575154ff05512f103ba6f3a266603f22fc0912d4cd14589fa +size 12527535 diff --git a/assets/minimax_h3_ref2va_dit_extreme_3.0bpp.mp4 b/assets/minimax_h3_ref2va_dit_extreme_3.0bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..a7542b98c93a18f641e67093ca1bac1a5401465a --- /dev/null +++ b/assets/minimax_h3_ref2va_dit_extreme_3.0bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2cfb1350ffd0cb8ecbaa67b6f65a879a872be28646aa8a8cb7dedb0e00fa2098 +size 24864847 diff --git a/assets/minimax_h3_ref2va_dit_fp8_5bpp.mp4 b/assets/minimax_h3_ref2va_dit_fp8_5bpp.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..7b3a2a4a8eb1e677b14f6fc5ad98e45e2054ecb5 --- /dev/null +++ b/assets/minimax_h3_ref2va_dit_fp8_5bpp.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1b600c61f47a9493a5292247920dd02dad85fd0e2603a0fa77c75ce5d6615646 +size 17608131 diff --git a/assets/minimax_h3_ref2va_reference_woman.png b/assets/minimax_h3_ref2va_reference_woman.png new file mode 100644 index 0000000000000000000000000000000000000000..c39ee8592320cfac22abcde6e6080d8c3c7ce958 --- /dev/null +++ b/assets/minimax_h3_ref2va_reference_woman.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a81d41f025468153d1edf0c1193f2c896d73409818d8eaca7cd52bc6fc96508 +size 1351278 diff --git a/assets/qwen_image_2.1_dit_4bpp.png b/assets/qwen_image_2.1_dit_4bpp.png new file mode 100644 index 0000000000000000000000000000000000000000..fd91f0b301fc45378ded71cdde388749ae6538e9 --- /dev/null +++ b/assets/qwen_image_2.1_dit_4bpp.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:63f4710605afa1d78892d87a9dabf9a18abfef83721de6e174f330c0ca15724e +size 1731993 diff --git a/assets/qwen_image_2.1_dit_5bpp.png b/assets/qwen_image_2.1_dit_5bpp.png new file mode 100644 index 0000000000000000000000000000000000000000..9808e5c62d8de2a09543b4286da2150c3c8f4a81 --- /dev/null +++ b/assets/qwen_image_2.1_dit_5bpp.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:11a01295fe7def8c5dc4917e45f20a3a493772d7b8b7ae0f0a39e5a2db9f2b0a +size 1719908 diff --git a/assets/qwen_image_2.1_dit_6bpp.png b/assets/qwen_image_2.1_dit_6bpp.png new file mode 100644 index 0000000000000000000000000000000000000000..dc5eef713539656c271b983d3fc508f63cdefd8c --- /dev/null +++ b/assets/qwen_image_2.1_dit_6bpp.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fedadc53e35c230f97f1b25e9d52617ddbfc50f998d33fd13faa5ccc94269ff4 +size 1739844 diff --git a/assets/qwen_image_2.1_dit_7bpp.png b/assets/qwen_image_2.1_dit_7bpp.png new file mode 100644 index 0000000000000000000000000000000000000000..8793ad381fe1d7c0371410ed1f857a30cbf75117 --- /dev/null +++ b/assets/qwen_image_2.1_dit_7bpp.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b76863098e63d53bc0932ba87037a22f8fd63e5a2be85fd25b3850de086fb926 +size 1713653 diff --git a/assets/qwen_image_2.1_dit_8bpp.png b/assets/qwen_image_2.1_dit_8bpp.png new file mode 100644 index 0000000000000000000000000000000000000000..00e0903d7c2c33a6dc0840c2be36a8672d1a0fe5 --- /dev/null +++ b/assets/qwen_image_2.1_dit_8bpp.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:faf1010e26eebea85e4a2f102566da3714daac2506bada7915b13089f0180474 +size 1703392 diff --git a/assets/qwen_image_2.1_dit_extreme_3.0bpp.png b/assets/qwen_image_2.1_dit_extreme_3.0bpp.png new file mode 100644 index 0000000000000000000000000000000000000000..e17722bd3def119aa157724cd2f4e19f571c35c5 --- /dev/null +++ b/assets/qwen_image_2.1_dit_extreme_3.0bpp.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f228f75f736a22a9c8f19f84a3e796516d7ba733efd318592e1b02733c109125 +size 1739484 diff --git a/assets/qwen_image_2.1_dit_fp8_5bpp.png b/assets/qwen_image_2.1_dit_fp8_5bpp.png new file mode 100644 index 0000000000000000000000000000000000000000..960bd3568fd3f6c503bac87d3f6bebfdcf536a0d --- /dev/null +++ b/assets/qwen_image_2.1_dit_fp8_5bpp.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ca4618ef2db88e00e6f5ea59ab46711bfadc229ce16e7228c175b44257c1b48 +size 1713289 diff --git a/assets/z_image_turbo_dit_4bpp.jpg b/assets/z_image_turbo_dit_4bpp.jpg new file mode 100644 index 0000000000000000000000000000000000000000..4a8ef85c655a123ded3c8c9b6f99cdc5fe9ee683 --- /dev/null +++ b/assets/z_image_turbo_dit_4bpp.jpg @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:059d37f0e58e288b9ea574bd5daeaac576e88bb549d09dd6ec03030660c81041 +size 100927 diff --git a/assets/z_image_turbo_dit_5bpp.jpg b/assets/z_image_turbo_dit_5bpp.jpg new file mode 100644 index 0000000000000000000000000000000000000000..8c21553b1d235f1a6468f487a14c607a79ab2ca4 --- /dev/null +++ b/assets/z_image_turbo_dit_5bpp.jpg @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e0630873fcd953cf6f722d978c39a44cfb3cda9acba82f67eb876bb2af961a60 +size 102477 diff --git a/assets/z_image_turbo_dit_6bpp.jpg b/assets/z_image_turbo_dit_6bpp.jpg new file mode 100644 index 0000000000000000000000000000000000000000..8b6fb84586877f8d59ba7ab282f8fa469d807512 --- /dev/null +++ b/assets/z_image_turbo_dit_6bpp.jpg @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:279ddd20935b3b6a4dd8cf7d51c7c9d4a1c70622250f87b5942ed9d27ae35944 +size 102866 diff --git a/assets/z_image_turbo_dit_7bpp.jpg b/assets/z_image_turbo_dit_7bpp.jpg new file mode 100644 index 0000000000000000000000000000000000000000..8f51a9537671fce693e202fe2566935c719e49f4 --- /dev/null +++ b/assets/z_image_turbo_dit_7bpp.jpg @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:55d5a34c0d9d695f6b450c019e621af0d18dcf2cce24494396e778ad72d52847 +size 101914 diff --git a/assets/z_image_turbo_dit_8bpp.jpg b/assets/z_image_turbo_dit_8bpp.jpg new file mode 100644 index 0000000000000000000000000000000000000000..1bac46a87066e3e1b438daf9139d77eccf503b2b --- /dev/null +++ b/assets/z_image_turbo_dit_8bpp.jpg @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:38892ee05fd57c0d4a83d80189cd950a8f9b3939aa7b02931f6b3ec29ad4838b +size 102158 diff --git a/assets/z_image_turbo_dit_extreme_2.3bpp.jpg b/assets/z_image_turbo_dit_extreme_2.3bpp.jpg new file mode 100644 index 0000000000000000000000000000000000000000..f3e83d7177ca873a1e0c47104bfce44970f3c3f1 --- /dev/null +++ b/assets/z_image_turbo_dit_extreme_2.3bpp.jpg @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eb95bcb85fa86d85b8e304b8bbb90d78c8514789f81d938e298249820fd7db05 +size 108174 diff --git a/assets/z_image_turbo_dit_fp8_5bpp.jpg b/assets/z_image_turbo_dit_fp8_5bpp.jpg new file mode 100644 index 0000000000000000000000000000000000000000..02afb346b83518ad6cd319665ffa6bda9eac9e73 --- /dev/null +++ b/assets/z_image_turbo_dit_fp8_5bpp.jpg @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:95ba92acaa83d3ff5b536d770a59df61619df0baa19934b1807e314ad3268a3b +size 103442 diff --git a/configuration.json b/configuration.json new file mode 100644 index 0000000000000000000000000000000000000000..63572e1f551ba544fc517a662b76c063f7618d4d --- /dev/null +++ b/configuration.json @@ -0,0 +1 @@ +{"task":"text-to-image-synthesis"} \ No newline at end of file diff --git a/models/MiniMax-H3/FL2VA/dit_4bpp.json b/models/MiniMax-H3/FL2VA/dit_4bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..6db2214b34eba71b09cc5e3dda58afe1be96e5dd --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_4bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "612cc4c6e1da912c8ce01e63e113215f", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/FL2VA/dit_4bpp.safetensors b/models/MiniMax-H3/FL2VA/dit_4bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..606242e45a2066939f4d67ce7e5b4b93d45ceda7 --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_4bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0bc847b6205a2efdaa30e8895c71433a6b3ed677563adff1c43a577e774dce29 +size 16726248296 diff --git a/models/MiniMax-H3/FL2VA/dit_5bpp.json b/models/MiniMax-H3/FL2VA/dit_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..cc1b82a010fe1a20df9c764fd9a03d8bb028bf84 --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_5bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "d3c067cb3761cd4dfc3d245a241062d4", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/FL2VA/dit_5bpp.safetensors b/models/MiniMax-H3/FL2VA/dit_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..505dd6f5ae5231351a94af816b3011c0c4b0342a --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cb44baaf9a1e0fa98b9ed1cca900cc855de03ab1b8eb3dbfe069899820067842 +size 20978626656 diff --git a/models/MiniMax-H3/FL2VA/dit_6bpp.json b/models/MiniMax-H3/FL2VA/dit_6bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..51eecb01fd7454ca6da8c9d7b001cc16f3202716 --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_6bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "dcbfbfa6dfa5228b6b207879687c58d8", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 6.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/FL2VA/dit_6bpp.safetensors b/models/MiniMax-H3/FL2VA/dit_6bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b4d3731b9457b615e7e7ca0d2e3ac27a1a7e8368 --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_6bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:10ef6ef8eb8e7cee584f6ce92d2a790d66147d1e96c167206b977f939a5d3bae +size 25183663296 diff --git a/models/MiniMax-H3/FL2VA/dit_7bpp.json b/models/MiniMax-H3/FL2VA/dit_7bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..2d8971f584d1cd850295365b2fe4c4af5e8e9792 --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_7bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "19f12bd0e985a00a192cac32fd605833", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 7.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/FL2VA/dit_7bpp.safetensors b/models/MiniMax-H3/FL2VA/dit_7bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b7dda85fbea4f584d9046a314861cf92661e3d71 --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_7bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:149a454b25635b82d373b7fc7f276b3bf51271571313510c90a47928366b3c48 +size 29372100480 diff --git a/models/MiniMax-H3/FL2VA/dit_8bpp.json b/models/MiniMax-H3/FL2VA/dit_8bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..cf022928b5e00dd6d8e696e5e07e26c4299ebec8 --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_8bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "601fbcb5f5e5b3531a46754b93d5750b", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/FL2VA/dit_8bpp.safetensors b/models/MiniMax-H3/FL2VA/dit_8bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1057a8d094f8b9f06ad920e0ff6b8bf4d34a31f7 --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_8bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e32600e8280b7e87aff01e80f0357ecd16e203b87fe69bf74b976dc78e1498cf +size 33462978544 diff --git a/models/MiniMax-H3/FL2VA/dit_extreme_3.0bpp.json b/models/MiniMax-H3/FL2VA/dit_extreme_3.0bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..9fe026b2ad24699153864c070c5ab5525d8d93ea --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_extreme_3.0bpp.json @@ -0,0 +1,306 @@ +{ + "model_hash": "29795a310c0b91df83abc8990d2ea8fd", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 2.9875992063492065, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "blocks.0.attn.qkv_proj", + "blocks.0.mlp.fc2", + "blocks.0.mlp.fc1", + "blocks.1.attn.qkv_proj", + "blocks.1.mlp.fc2", + "blocks.1.mlp.fc1", + "blocks.2.mlp.fc2", + "blocks.2.mlp.fc1", + "token_refiner.blocks.0.attn.qkv_proj", + "token_refiner.blocks.0.mlp.fc2", + "token_refiner.blocks.0.mlp.fc1", + "token_refiner.blocks.1.attn.qkv_proj", + "token_refiner.blocks.1.mlp.fc2", + "token_refiner.blocks.1.mlp.fc1", + "blocks.2.attn.qkv_proj", + "blocks.3.attn.qkv_proj", + "blocks.3.mlp.fc2", + "blocks.3.mlp.fc1", + "blocks.4.attn.qkv_proj", + "blocks.4.mlp.fc2", + "blocks.4.mlp.fc1", + "blocks.5.attn.qkv_proj", + "blocks.5.mlp.fc2", + "blocks.5.mlp.fc1", + "blocks.6.mlp.fc2", + "blocks.6.mlp.fc1", + "blocks.10.mlp.fc2", + "blocks.10.mlp.fc1", + "blocks.6.attn.qkv_proj", + "blocks.7.attn.qkv_proj", + "blocks.7.mlp.fc2", + "blocks.7.mlp.fc1", + "blocks.8.attn.qkv_proj", + "blocks.8.mlp.fc2", + "blocks.8.mlp.fc1", + "blocks.9.attn.qkv_proj", + "blocks.9.mlp.fc2", + "blocks.9.mlp.fc1", + "blocks.10.attn.qkv_proj", + "blocks.11.attn.qkv_proj", + "blocks.11.mlp.fc2", + "blocks.11.mlp.fc1", + "blocks.12.attn.qkv_proj", + "blocks.12.mlp.fc2", + "blocks.12.mlp.fc1", + "blocks.13.attn.qkv_proj", + "blocks.13.mlp.fc2", + "blocks.13.mlp.fc1", + "blocks.14.mlp.fc2", + "blocks.14.mlp.fc1", + "blocks.14.attn.qkv_proj", + "blocks.15.attn.qkv_proj", + "blocks.15.mlp.fc2", + "blocks.15.mlp.fc1", + "blocks.16.attn.qkv_proj", + "blocks.16.mlp.fc2", + "blocks.16.mlp.fc1", + "blocks.17.attn.qkv_proj", + "blocks.17.mlp.fc2", + "blocks.17.mlp.fc1", + "blocks.18.mlp.fc2", + "blocks.18.mlp.fc1", + "blocks.18.attn.qkv_proj", + "blocks.19.attn.qkv_proj", + "blocks.19.mlp.fc2", + "blocks.19.mlp.fc1", + "blocks.20.attn.qkv_proj", + "blocks.20.mlp.fc2", + "blocks.20.mlp.fc1", + "blocks.21.attn.qkv_proj", + "blocks.21.mlp.fc2", + "blocks.21.mlp.fc1", + "blocks.22.mlp.fc2", + "blocks.22.mlp.fc1", + "blocks.22.attn.qkv_proj", + "blocks.23.attn.qkv_proj", + "blocks.23.mlp.fc2", + "blocks.23.mlp.fc1", + "blocks.24.attn.qkv_proj", + "blocks.24.mlp.fc2", + "blocks.24.mlp.fc1", + "blocks.25.attn.qkv_proj", + "blocks.25.mlp.fc2", + "blocks.25.mlp.fc1", + "blocks.26.mlp.fc2", + "blocks.26.mlp.fc1", + "blocks.26.attn.qkv_proj", + "blocks.27.attn.qkv_proj", + "blocks.27.mlp.fc2", + "blocks.27.mlp.fc1", + "blocks.28.attn.qkv_proj", + "blocks.28.mlp.fc2", + "blocks.28.mlp.fc1", + "blocks.29.attn.qkv_proj", + "blocks.29.mlp.fc2", + "blocks.29.mlp.fc1", + "blocks.30.mlp.fc2", + "blocks.30.mlp.fc1", + "blocks.30.attn.qkv_proj", + "blocks.31.attn.qkv_proj", + "blocks.31.mlp.fc2", + "blocks.31.mlp.fc1", + "blocks.32.attn.qkv_proj", + "blocks.32.mlp.fc2", + "blocks.32.mlp.fc1", + "blocks.33.attn.qkv_proj", + "blocks.33.mlp.fc2", + "blocks.33.mlp.fc1", + "blocks.34.mlp.fc2", + "blocks.34.mlp.fc1", + "blocks.34.attn.qkv_proj", + "blocks.35.attn.qkv_proj", + "blocks.35.mlp.fc2", + "blocks.35.mlp.fc1", + "blocks.36.attn.qkv_proj", + "blocks.36.mlp.fc2", + "blocks.36.mlp.fc1", + "blocks.37.attn.qkv_proj", + "blocks.37.mlp.fc2", + "blocks.37.mlp.fc1", + "blocks.38.mlp.fc2", + "blocks.38.mlp.fc1", + "blocks.38.attn.qkv_proj", + "blocks.39.attn.qkv_proj", + "blocks.39.mlp.fc2", + "blocks.39.mlp.fc1", + "blocks.40.attn.qkv_proj", + "blocks.40.mlp.fc2", + "blocks.40.mlp.fc1", + "blocks.41.attn.qkv_proj", + "blocks.41.mlp.fc2", + "blocks.41.mlp.fc1", + "blocks.42.mlp.fc2", + "blocks.42.mlp.fc1", + "blocks.42.attn.qkv_proj", + "blocks.43.attn.qkv_proj", + "blocks.43.mlp.fc2", + "blocks.43.mlp.fc1", + "blocks.44.attn.qkv_proj", + "blocks.44.mlp.fc2", + "blocks.44.mlp.fc1", + "blocks.45.attn.qkv_proj", + "blocks.45.mlp.fc2", + "blocks.45.mlp.fc1", + "blocks.46.mlp.fc2", + "blocks.46.mlp.fc1", + "blocks.46.attn.qkv_proj", + "blocks.47.attn.qkv_proj", + "blocks.47.mlp.fc2", + "blocks.47.mlp.fc1", + "blocks.48.attn.qkv_proj", + "blocks.48.mlp.fc2", + "blocks.48.mlp.fc1", + "blocks.49.attn.qkv_proj", + "blocks.49.mlp.fc2", + "blocks.49.mlp.fc1" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 3.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "blocks.0.adaln_proj.linear", + "blocks.0.attn.out_proj", + "blocks.1.adaln_proj.linear", + "blocks.1.attn.out_proj", + "blocks.2.adaln_proj.linear", + "token_refiner.blocks.0.attn.out_proj", + "token_refiner.blocks.1.attn.out_proj", + "blocks.2.attn.out_proj", + "blocks.3.adaln_proj.linear", + "blocks.3.attn.out_proj", + "blocks.4.adaln_proj.linear", + "blocks.4.attn.out_proj", + "blocks.5.adaln_proj.linear", + "blocks.5.attn.out_proj", + "blocks.6.adaln_proj.linear", + "blocks.10.adaln_proj.linear", + "blocks.6.attn.out_proj", + "blocks.7.adaln_proj.linear", + "blocks.7.attn.out_proj", + "blocks.8.adaln_proj.linear", + "blocks.8.attn.out_proj", + "blocks.9.adaln_proj.linear", + "blocks.9.attn.out_proj", + "blocks.10.attn.out_proj", + "blocks.11.adaln_proj.linear", + "blocks.11.attn.out_proj", + "blocks.12.adaln_proj.linear", + "blocks.12.attn.out_proj", + "blocks.13.adaln_proj.linear", + "blocks.13.attn.out_proj", + "blocks.14.adaln_proj.linear", + "blocks.14.attn.out_proj", + "blocks.15.adaln_proj.linear", + "blocks.15.attn.out_proj", + "blocks.16.adaln_proj.linear", + "blocks.16.attn.out_proj", + "blocks.17.adaln_proj.linear", + "blocks.17.attn.out_proj", + "blocks.18.adaln_proj.linear", + "blocks.18.attn.out_proj", + "blocks.19.adaln_proj.linear", + "blocks.19.attn.out_proj", + "blocks.20.adaln_proj.linear", + "blocks.20.attn.out_proj", + "blocks.21.adaln_proj.linear", + "blocks.21.attn.out_proj", + "blocks.22.adaln_proj.linear", + "blocks.22.attn.out_proj", + "blocks.23.adaln_proj.linear", + "blocks.23.attn.out_proj", + "blocks.24.adaln_proj.linear", + "blocks.24.attn.out_proj", + "blocks.25.adaln_proj.linear", + "blocks.25.attn.out_proj", + "blocks.26.adaln_proj.linear", + "blocks.26.attn.out_proj", + "blocks.27.adaln_proj.linear", + "blocks.27.attn.out_proj", + "blocks.28.adaln_proj.linear", + "blocks.28.attn.out_proj", + "blocks.29.adaln_proj.linear", + "blocks.29.attn.out_proj", + "blocks.30.adaln_proj.linear", + "blocks.30.attn.out_proj", + "blocks.31.adaln_proj.linear", + "blocks.31.attn.out_proj", + "blocks.32.adaln_proj.linear", + "blocks.32.attn.out_proj", + "blocks.33.adaln_proj.linear", + "blocks.33.attn.out_proj", + "blocks.34.adaln_proj.linear", + "blocks.34.attn.out_proj", + "blocks.35.adaln_proj.linear", + "blocks.35.attn.out_proj", + "blocks.36.adaln_proj.linear", + "blocks.36.attn.out_proj", + "blocks.37.adaln_proj.linear", + "blocks.37.attn.out_proj", + "blocks.38.adaln_proj.linear", + "blocks.38.attn.out_proj", + "blocks.39.adaln_proj.linear", + "blocks.39.attn.out_proj", + "blocks.40.adaln_proj.linear", + "blocks.40.attn.out_proj", + "blocks.41.adaln_proj.linear", + "blocks.41.attn.out_proj", + "blocks.42.adaln_proj.linear", + "blocks.42.attn.out_proj", + "blocks.43.adaln_proj.linear", + "blocks.43.attn.out_proj", + "blocks.44.adaln_proj.linear", + "blocks.44.attn.out_proj", + "blocks.45.adaln_proj.linear", + "blocks.45.attn.out_proj", + "blocks.46.adaln_proj.linear", + "blocks.46.attn.out_proj", + "blocks.47.adaln_proj.linear", + "blocks.47.attn.out_proj", + "blocks.48.adaln_proj.linear", + "blocks.48.attn.out_proj", + "blocks.49.adaln_proj.linear", + "blocks.49.attn.out_proj", + "final_layer.adaln_proj.linear" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "audio_patch_proj", + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "condition_proj", + "final_layer.audio_out", + "final_layer.video_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/FL2VA/dit_extreme_3.0bpp.safetensors b/models/MiniMax-H3/FL2VA/dit_extreme_3.0bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..91961b77eedb7aa8a10959efbbb1d524bce03ccf --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_extreme_3.0bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ce5d26f13ad42a80bd5b515096ecef1ca6b062ed8c5765ac8950383a45ed5e91 +size 12535981160 diff --git a/models/MiniMax-H3/FL2VA/dit_fp8_5bpp.json b/models/MiniMax-H3/FL2VA/dit_fp8_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..8cbae45e2c6f3fed8825677131f061ded1007688 --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_fp8_5bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "16ca1fa6f82125d160f4b185a3ba2096", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant_fp8", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/FL2VA/dit_fp8_5bpp.safetensors b/models/MiniMax-H3/FL2VA/dit_fp8_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..681c9a0a97eb823b9d5e4de48404b25f308d698c --- /dev/null +++ b/models/MiniMax-H3/FL2VA/dit_fp8_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c017b2f9118486aaedafd57797826bd1108a83fbd237074f62e0000eaed52c05 +size 20981618128 diff --git a/models/MiniMax-H3/REF2VA/dit_4bpp.json b/models/MiniMax-H3/REF2VA/dit_4bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..6feadb6fb53acd5926aa663ad4738027063b254c --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_4bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "49df57910799fd05dca1e79ec1fbf0e9", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/REF2VA/dit_4bpp.safetensors b/models/MiniMax-H3/REF2VA/dit_4bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..58b1b68c4fb43546f03d9597eff6ff8ff1e597a1 --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_4bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b6f974fc0d9ba9286c23fd9764b1ee2cda679b507139a8b2b69e9cbb04dc4a08 +size 16726551560 diff --git a/models/MiniMax-H3/REF2VA/dit_5bpp.json b/models/MiniMax-H3/REF2VA/dit_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..84ac58d3927b6f1948ee198548e3b2454bf55a10 --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_5bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "e6e4edc4382aac92c772ff56625eb108", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/REF2VA/dit_5bpp.safetensors b/models/MiniMax-H3/REF2VA/dit_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fc6553438284031cdf88555674b66be2cd7af76f --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dcb400d282ed59bf54388b0a684e2556a33a6a66869d90255b7de1e0cc998f9f +size 20978285280 diff --git a/models/MiniMax-H3/REF2VA/dit_6bpp.json b/models/MiniMax-H3/REF2VA/dit_6bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..32a8ff435691b5447eea2c7c61098fc4e9b85605 --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_6bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "292821084a3e56e01df4b29e3abb84d1", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 6.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/REF2VA/dit_6bpp.safetensors b/models/MiniMax-H3/REF2VA/dit_6bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0a84c9a1303d4a5e8580db227cabf654e81e83df --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_6bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8e868a65f8f3c7c94ff347143bb47de91fdef9fdfb5bc4977706788ea6f13b6d +size 25187787440 diff --git a/models/MiniMax-H3/REF2VA/dit_7bpp.json b/models/MiniMax-H3/REF2VA/dit_7bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..36b169667b763bb8c5f7555e4cd104d35741648f --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_7bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "d1d5f217965889e5e59dee810e3ce118", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 7.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/REF2VA/dit_7bpp.safetensors b/models/MiniMax-H3/REF2VA/dit_7bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..705c7f362d83a8a3ecae2455c6785349c5e7bf40 --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_7bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6e42b740c60ef903f773ffcca8a3ee99ce38cf6ff020fe38d1b42f35922e8d8f +size 29383627664 diff --git a/models/MiniMax-H3/REF2VA/dit_8bpp.json b/models/MiniMax-H3/REF2VA/dit_8bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..e56c4879579577056d0bd60ff49c9a558124485a --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_8bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "26745a544a73acfe2a13f37a53e4b9fa", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/REF2VA/dit_8bpp.safetensors b/models/MiniMax-H3/REF2VA/dit_8bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..22de99cd2ebd9437fe4a89c6817d801d8a5918f3 --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_8bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:589aa373ac4db72f5d9a81ebf367e55f63881ea1499154378ff051c4ab448d6e +size 33456926544 diff --git a/models/MiniMax-H3/REF2VA/dit_extreme_3.0bpp.json b/models/MiniMax-H3/REF2VA/dit_extreme_3.0bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..dcbc14bcf564ef2822709e37252b8fc21a3406b7 --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_extreme_3.0bpp.json @@ -0,0 +1,306 @@ +{ + "model_hash": "a7409c6124dce3e1c1e6e80c5814a54e", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 2.9875992063492065, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "blocks.0.attn.qkv_proj", + "blocks.0.mlp.fc2", + "blocks.0.mlp.fc1", + "blocks.1.attn.qkv_proj", + "blocks.1.mlp.fc2", + "blocks.1.mlp.fc1", + "blocks.2.mlp.fc2", + "blocks.2.mlp.fc1", + "token_refiner.blocks.0.attn.qkv_proj", + "token_refiner.blocks.0.mlp.fc2", + "token_refiner.blocks.0.mlp.fc1", + "token_refiner.blocks.1.attn.qkv_proj", + "token_refiner.blocks.1.mlp.fc2", + "token_refiner.blocks.1.mlp.fc1", + "blocks.2.attn.qkv_proj", + "blocks.3.attn.qkv_proj", + "blocks.3.mlp.fc2", + "blocks.3.mlp.fc1", + "blocks.4.attn.qkv_proj", + "blocks.4.mlp.fc2", + "blocks.4.mlp.fc1", + "blocks.5.attn.qkv_proj", + "blocks.5.mlp.fc2", + "blocks.5.mlp.fc1", + "blocks.6.mlp.fc2", + "blocks.6.mlp.fc1", + "blocks.10.mlp.fc2", + "blocks.10.mlp.fc1", + "blocks.6.attn.qkv_proj", + "blocks.7.attn.qkv_proj", + "blocks.7.mlp.fc2", + "blocks.7.mlp.fc1", + "blocks.8.attn.qkv_proj", + "blocks.8.mlp.fc2", + "blocks.8.mlp.fc1", + "blocks.9.attn.qkv_proj", + "blocks.9.mlp.fc2", + "blocks.9.mlp.fc1", + "blocks.10.attn.qkv_proj", + "blocks.11.attn.qkv_proj", + "blocks.11.mlp.fc2", + "blocks.11.mlp.fc1", + "blocks.12.attn.qkv_proj", + "blocks.12.mlp.fc2", + "blocks.12.mlp.fc1", + "blocks.13.attn.qkv_proj", + "blocks.13.mlp.fc2", + "blocks.13.mlp.fc1", + "blocks.14.mlp.fc2", + "blocks.14.mlp.fc1", + "blocks.14.attn.qkv_proj", + "blocks.15.attn.qkv_proj", + "blocks.15.mlp.fc2", + "blocks.15.mlp.fc1", + "blocks.16.attn.qkv_proj", + "blocks.16.mlp.fc2", + "blocks.16.mlp.fc1", + "blocks.17.attn.qkv_proj", + "blocks.17.mlp.fc2", + "blocks.17.mlp.fc1", + "blocks.18.mlp.fc2", + "blocks.18.mlp.fc1", + "blocks.18.attn.qkv_proj", + "blocks.19.attn.qkv_proj", + "blocks.19.mlp.fc2", + "blocks.19.mlp.fc1", + "blocks.20.attn.qkv_proj", + "blocks.20.mlp.fc2", + "blocks.20.mlp.fc1", + "blocks.21.attn.qkv_proj", + "blocks.21.mlp.fc2", + "blocks.21.mlp.fc1", + "blocks.22.mlp.fc2", + "blocks.22.mlp.fc1", + "blocks.22.attn.qkv_proj", + "blocks.23.attn.qkv_proj", + "blocks.23.mlp.fc2", + "blocks.23.mlp.fc1", + "blocks.24.attn.qkv_proj", + "blocks.24.mlp.fc2", + "blocks.24.mlp.fc1", + "blocks.25.attn.qkv_proj", + "blocks.25.mlp.fc2", + "blocks.25.mlp.fc1", + "blocks.26.mlp.fc2", + "blocks.26.mlp.fc1", + "blocks.26.attn.qkv_proj", + "blocks.27.attn.qkv_proj", + "blocks.27.mlp.fc2", + "blocks.27.mlp.fc1", + "blocks.28.attn.qkv_proj", + "blocks.28.mlp.fc2", + "blocks.28.mlp.fc1", + "blocks.29.attn.qkv_proj", + "blocks.29.mlp.fc2", + "blocks.29.mlp.fc1", + "blocks.30.mlp.fc2", + "blocks.30.mlp.fc1", + "blocks.30.attn.qkv_proj", + "blocks.31.attn.qkv_proj", + "blocks.31.mlp.fc2", + "blocks.31.mlp.fc1", + "blocks.32.attn.qkv_proj", + "blocks.32.mlp.fc2", + "blocks.32.mlp.fc1", + "blocks.33.attn.qkv_proj", + "blocks.33.mlp.fc2", + "blocks.33.mlp.fc1", + "blocks.34.mlp.fc2", + "blocks.34.mlp.fc1", + "blocks.34.attn.qkv_proj", + "blocks.35.attn.qkv_proj", + "blocks.35.mlp.fc2", + "blocks.35.mlp.fc1", + "blocks.36.attn.qkv_proj", + "blocks.36.mlp.fc2", + "blocks.36.mlp.fc1", + "blocks.37.attn.qkv_proj", + "blocks.37.mlp.fc2", + "blocks.37.mlp.fc1", + "blocks.38.mlp.fc2", + "blocks.38.mlp.fc1", + "blocks.38.attn.qkv_proj", + "blocks.39.attn.qkv_proj", + "blocks.39.mlp.fc2", + "blocks.39.mlp.fc1", + "blocks.40.attn.qkv_proj", + "blocks.40.mlp.fc2", + "blocks.40.mlp.fc1", + "blocks.41.attn.qkv_proj", + "blocks.41.mlp.fc2", + "blocks.41.mlp.fc1", + "blocks.42.mlp.fc2", + "blocks.42.mlp.fc1", + "blocks.42.attn.qkv_proj", + "blocks.43.attn.qkv_proj", + "blocks.43.mlp.fc2", + "blocks.43.mlp.fc1", + "blocks.44.attn.qkv_proj", + "blocks.44.mlp.fc2", + "blocks.44.mlp.fc1", + "blocks.45.attn.qkv_proj", + "blocks.45.mlp.fc2", + "blocks.45.mlp.fc1", + "blocks.46.mlp.fc2", + "blocks.46.mlp.fc1", + "blocks.46.attn.qkv_proj", + "blocks.47.attn.qkv_proj", + "blocks.47.mlp.fc2", + "blocks.47.mlp.fc1", + "blocks.48.attn.qkv_proj", + "blocks.48.mlp.fc2", + "blocks.48.mlp.fc1", + "blocks.49.attn.qkv_proj", + "blocks.49.mlp.fc2", + "blocks.49.mlp.fc1" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 3.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "blocks.0.adaln_proj.linear", + "blocks.0.attn.out_proj", + "blocks.1.adaln_proj.linear", + "blocks.1.attn.out_proj", + "blocks.2.adaln_proj.linear", + "token_refiner.blocks.0.attn.out_proj", + "token_refiner.blocks.1.attn.out_proj", + "blocks.2.attn.out_proj", + "blocks.3.adaln_proj.linear", + "blocks.3.attn.out_proj", + "blocks.4.adaln_proj.linear", + "blocks.4.attn.out_proj", + "blocks.5.adaln_proj.linear", + "blocks.5.attn.out_proj", + "blocks.6.adaln_proj.linear", + "blocks.10.adaln_proj.linear", + "blocks.6.attn.out_proj", + "blocks.7.adaln_proj.linear", + "blocks.7.attn.out_proj", + "blocks.8.adaln_proj.linear", + "blocks.8.attn.out_proj", + "blocks.9.adaln_proj.linear", + "blocks.9.attn.out_proj", + "blocks.10.attn.out_proj", + "blocks.11.adaln_proj.linear", + "blocks.11.attn.out_proj", + "blocks.12.adaln_proj.linear", + "blocks.12.attn.out_proj", + "blocks.13.adaln_proj.linear", + "blocks.13.attn.out_proj", + "blocks.14.adaln_proj.linear", + "blocks.14.attn.out_proj", + "blocks.15.adaln_proj.linear", + "blocks.15.attn.out_proj", + "blocks.16.adaln_proj.linear", + "blocks.16.attn.out_proj", + "blocks.17.adaln_proj.linear", + "blocks.17.attn.out_proj", + "blocks.18.adaln_proj.linear", + "blocks.18.attn.out_proj", + "blocks.19.adaln_proj.linear", + "blocks.19.attn.out_proj", + "blocks.20.adaln_proj.linear", + "blocks.20.attn.out_proj", + "blocks.21.adaln_proj.linear", + "blocks.21.attn.out_proj", + "blocks.22.adaln_proj.linear", + "blocks.22.attn.out_proj", + "blocks.23.adaln_proj.linear", + "blocks.23.attn.out_proj", + "blocks.24.adaln_proj.linear", + "blocks.24.attn.out_proj", + "blocks.25.adaln_proj.linear", + "blocks.25.attn.out_proj", + "blocks.26.adaln_proj.linear", + "blocks.26.attn.out_proj", + "blocks.27.adaln_proj.linear", + "blocks.27.attn.out_proj", + "blocks.28.adaln_proj.linear", + "blocks.28.attn.out_proj", + "blocks.29.adaln_proj.linear", + "blocks.29.attn.out_proj", + "blocks.30.adaln_proj.linear", + "blocks.30.attn.out_proj", + "blocks.31.adaln_proj.linear", + "blocks.31.attn.out_proj", + "blocks.32.adaln_proj.linear", + "blocks.32.attn.out_proj", + "blocks.33.adaln_proj.linear", + "blocks.33.attn.out_proj", + "blocks.34.adaln_proj.linear", + "blocks.34.attn.out_proj", + "blocks.35.adaln_proj.linear", + "blocks.35.attn.out_proj", + "blocks.36.adaln_proj.linear", + "blocks.36.attn.out_proj", + "blocks.37.adaln_proj.linear", + "blocks.37.attn.out_proj", + "blocks.38.adaln_proj.linear", + "blocks.38.attn.out_proj", + "blocks.39.adaln_proj.linear", + "blocks.39.attn.out_proj", + "blocks.40.adaln_proj.linear", + "blocks.40.attn.out_proj", + "blocks.41.adaln_proj.linear", + "blocks.41.attn.out_proj", + "blocks.42.adaln_proj.linear", + "blocks.42.attn.out_proj", + "blocks.43.adaln_proj.linear", + "blocks.43.attn.out_proj", + "blocks.44.adaln_proj.linear", + "blocks.44.attn.out_proj", + "blocks.45.adaln_proj.linear", + "blocks.45.attn.out_proj", + "blocks.46.adaln_proj.linear", + "blocks.46.attn.out_proj", + "blocks.47.adaln_proj.linear", + "blocks.47.attn.out_proj", + "blocks.48.adaln_proj.linear", + "blocks.48.attn.out_proj", + "blocks.49.adaln_proj.linear", + "blocks.49.attn.out_proj", + "final_layer.adaln_proj.linear" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "audio_patch_proj", + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "condition_proj", + "final_layer.audio_out", + "final_layer.video_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/REF2VA/dit_extreme_3.0bpp.safetensors b/models/MiniMax-H3/REF2VA/dit_extreme_3.0bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1a4c8ed7583e1e156b9a6efc4a808bcdafc30e19 --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_extreme_3.0bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f2abefcb27f683f1d416017faab1ec8818974266d2547fb6fd28bccacdb12714 +size 12536255960 diff --git a/models/MiniMax-H3/REF2VA/dit_fp8_5bpp.json b/models/MiniMax-H3/REF2VA/dit_fp8_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..89f2ccc74be7f3ffa1f62c02388aedd21b33b7e3 --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_fp8_5bpp.json @@ -0,0 +1,44 @@ +{ + "model_hash": "c62f95564a2e0f9bee3e21ce1b1cbb93", + "model_name": "minimax_h3_dit", + "model_class": "diffsynth.models.minimax_h3_dit.MiniMaxH3DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant_fp8", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "time_embedder.proj_in", + "time_embedder.proj_out", + "video_patch_proj", + "audio_patch_proj", + "condition_proj", + "final_layer.video_out", + "final_layer.audio_out" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/REF2VA/dit_fp8_5bpp.safetensors b/models/MiniMax-H3/REF2VA/dit_fp8_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..82d7e7cdfb4a5eae8d8b6431adac1e4c9f406443 --- /dev/null +++ b/models/MiniMax-H3/REF2VA/dit_fp8_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cdd3986da27261a7a08cacb1c6d15f07bc076cff00defa9f88078ec21f12fa4c +size 20981984128 diff --git a/models/MiniMax-H3/text_encoder_4bpp.json b/models/MiniMax-H3/text_encoder_4bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..6fac22aacf8ab99a90638eeee1f576781f891083 --- /dev/null +++ b/models/MiniMax-H3/text_encoder_4bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "8c6f5ccfe9ffe861675ca0da95bb9295", + "model_name": "minimax_h3_text_encoder", + "model_class": "diffsynth.models.minimax_h3_text_encoder.MiniMaxH3TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/text_encoder_4bpp.safetensors b/models/MiniMax-H3/text_encoder_4bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0eea87b7c466afca5048a08887be0214ee4a66bd --- /dev/null +++ b/models/MiniMax-H3/text_encoder_4bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:207b73dd37c2a0888965a45b5cdb4456255e8f1ed7f528b4fe32b1da2183d5ba +size 14174263896 diff --git a/models/MiniMax-H3/text_encoder_5bpp.json b/models/MiniMax-H3/text_encoder_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..e3da34dc23aca64c01fec6fa41cbe70bd9149445 --- /dev/null +++ b/models/MiniMax-H3/text_encoder_5bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "0bed581070b396883a96632b22848aa2", + "model_name": "minimax_h3_text_encoder", + "model_class": "diffsynth.models.minimax_h3_text_encoder.MiniMaxH3TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/text_encoder_5bpp.safetensors b/models/MiniMax-H3/text_encoder_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a733a07131698945767ca8547519aab191a94bdb --- /dev/null +++ b/models/MiniMax-H3/text_encoder_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0e2bf2011c462b1aaa6e9285d8fba41386d186ee364972794ad74d1e809aabc0 +size 17400593792 diff --git a/models/MiniMax-H3/text_encoder_6bpp.json b/models/MiniMax-H3/text_encoder_6bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..68920be8b189b78015d3163f142c437eb44315e4 --- /dev/null +++ b/models/MiniMax-H3/text_encoder_6bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "84374e496c86bc34989eaa893b50a9a8", + "model_name": "minimax_h3_text_encoder", + "model_class": "diffsynth.models.minimax_h3_text_encoder.MiniMaxH3TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 6.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/text_encoder_6bpp.safetensors b/models/MiniMax-H3/text_encoder_6bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4bbb0257a664b9898efe596f59f13576ba6af18e --- /dev/null +++ b/models/MiniMax-H3/text_encoder_6bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6c4390126bfc61c88bab5111e8529b78dccde216a2e744d0f48dbc8cb8666254 +size 20613948104 diff --git a/models/MiniMax-H3/text_encoder_7bpp.json b/models/MiniMax-H3/text_encoder_7bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..f5fb55077a1f8a0a092a485832f9bb89005a4df8 --- /dev/null +++ b/models/MiniMax-H3/text_encoder_7bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "455ee7ef05d5bbb3c1fb1e7a8147cb61", + "model_name": "minimax_h3_text_encoder", + "model_class": "diffsynth.models.minimax_h3_text_encoder.MiniMaxH3TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 7.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/text_encoder_7bpp.safetensors b/models/MiniMax-H3/text_encoder_7bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4fa299f1116ace689dee318805c6aec2d238fa84 --- /dev/null +++ b/models/MiniMax-H3/text_encoder_7bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2febecd4a646d3f11530a49385f31f06623df17d59eef8fc6773e58da9b9d7a5 +size 23812932752 diff --git a/models/MiniMax-H3/text_encoder_8bpp.json b/models/MiniMax-H3/text_encoder_8bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..c2d5a0ad2c9ef6813f679fe295cf08d636452362 --- /dev/null +++ b/models/MiniMax-H3/text_encoder_8bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "5e498d14ba9ab4f230685ca419f43949", + "model_name": "minimax_h3_text_encoder", + "model_class": "diffsynth.models.minimax_h3_text_encoder.MiniMaxH3TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/text_encoder_8bpp.safetensors b/models/MiniMax-H3/text_encoder_8bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..09b1c3b7b852222d51a9a60cba30aa58b9f1a500 --- /dev/null +++ b/models/MiniMax-H3/text_encoder_8bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2ee2925c90ac5bee29a7b11e954e534055d785b480c8b7da51ab05f0d51b528d +size 26831739344 diff --git a/models/MiniMax-H3/video_vae_4bpp.json b/models/MiniMax-H3/video_vae_4bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..ebd12694c55ed85317898845c954a4cd7df95d95 --- /dev/null +++ b/models/MiniMax-H3/video_vae_4bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "611a5d0f0c1d9bae8d609f8ddc568dc5", + "model_name": "minimax_h3_video_vae", + "model_class": "diffsynth.models.minimax_h3_video_vae.MiniMaxH3VideoVAE", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/video_vae_4bpp.safetensors b/models/MiniMax-H3/video_vae_4bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..422df36c176c447eba490108edee9ba9c38333c7 --- /dev/null +++ b/models/MiniMax-H3/video_vae_4bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bec714e3d04ffb8a79f89298316525630861e0af55e54a4125bf3d2308618f5e +size 1580398808 diff --git a/models/MiniMax-H3/video_vae_5bpp.json b/models/MiniMax-H3/video_vae_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..579dfe1d9857da513c96ff208b11ad6cb6421962 --- /dev/null +++ b/models/MiniMax-H3/video_vae_5bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "64ec5c91ae690f7162a3bdb3d7517298", + "model_name": "minimax_h3_video_vae", + "model_class": "diffsynth.models.minimax_h3_video_vae.MiniMaxH3VideoVAE", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/video_vae_5bpp.safetensors b/models/MiniMax-H3/video_vae_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..081a2e77b12b5eabc96c86df74355fe485d872b9 --- /dev/null +++ b/models/MiniMax-H3/video_vae_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1d97ebf5e6e6277ddb52b1bb8f30b494bb6ead32b4570e07f18acb9cac08889a +size 1889253216 diff --git a/models/MiniMax-H3/video_vae_6bpp.json b/models/MiniMax-H3/video_vae_6bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..3424db6fdb703058d1aa2c3179ac100d9df3ab00 --- /dev/null +++ b/models/MiniMax-H3/video_vae_6bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "f76db8c08e3a042f3d1d8deb8adf61e2", + "model_name": "minimax_h3_video_vae", + "model_class": "diffsynth.models.minimax_h3_video_vae.MiniMaxH3VideoVAE", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 6.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/video_vae_6bpp.safetensors b/models/MiniMax-H3/video_vae_6bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..40ac726663fe0c58796e09aedf02f4c83321c1bf --- /dev/null +++ b/models/MiniMax-H3/video_vae_6bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f988a4f059e269e2e4bf4659b5eeeb272ba786c2a2fc5bc1eb535b65aa643105 +size 2193711120 diff --git a/models/MiniMax-H3/video_vae_7bpp.json b/models/MiniMax-H3/video_vae_7bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..5c740aae8e8722af649c88a0874228802f88fb37 --- /dev/null +++ b/models/MiniMax-H3/video_vae_7bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "b9974b020d52eb1e25066eb95a95d365", + "model_name": "minimax_h3_video_vae", + "model_class": "diffsynth.models.minimax_h3_video_vae.MiniMaxH3VideoVAE", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 7.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/video_vae_7bpp.safetensors b/models/MiniMax-H3/video_vae_7bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..18b6c32f4a5c59812e4f2a3e686dca82ad1ff335 --- /dev/null +++ b/models/MiniMax-H3/video_vae_7bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cebea5fbde78583d150cb515f6a9e2f29bfd42207231f9b5e74a90351757ead0 +size 2499535784 diff --git a/models/MiniMax-H3/video_vae_8bpp.json b/models/MiniMax-H3/video_vae_8bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..e46c5342ee017764a6a12f8ef98063605b0b423d --- /dev/null +++ b/models/MiniMax-H3/video_vae_8bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "f47f645fa291cad82be5dd09064288d2", + "model_name": "minimax_h3_video_vae", + "model_class": "diffsynth.models.minimax_h3_video_vae.MiniMaxH3VideoVAE", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/MiniMax-H3/video_vae_8bpp.safetensors b/models/MiniMax-H3/video_vae_8bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..02cd1722d7b26fb386a8b5843989d32b2dd8af98 --- /dev/null +++ b/models/MiniMax-H3/video_vae_8bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c548dee1698a518708ff07fb01deadee289f95a3b53819a8d0953714427db92e +size 2802642112 diff --git a/models/Qwen-Image-2.1/dit_4bpp.json b/models/Qwen-Image-2.1/dit_4bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..7bc33eeba965593a84833d454ed3f9539913d0c1 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_4bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "8379a4f9c705a948cd405e41da5a731b", + "model_name": "qwen_image_21_dit", + "model_class": "diffsynth.models.qwen_image_21_dit.QwenImage21DiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/dit_4bpp.safetensors b/models/Qwen-Image-2.1/dit_4bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c21da636204692c70fbf05f8b930a880d1676d4b --- /dev/null +++ b/models/Qwen-Image-2.1/dit_4bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fdf18574fda82bd2f0f0c9914887676288f3a3960095cc6304da5496c23e1589 +size 3581973192 diff --git a/models/Qwen-Image-2.1/dit_5bpp.json b/models/Qwen-Image-2.1/dit_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..5179b0e49c12f951084503c82937710bf8503a8e --- /dev/null +++ b/models/Qwen-Image-2.1/dit_5bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "e6858b051174a3592d2954ed9142b18a", + "model_name": "qwen_image_21_dit", + "model_class": "diffsynth.models.qwen_image_21_dit.QwenImage21DiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/dit_5bpp.safetensors b/models/Qwen-Image-2.1/dit_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f9c700014f11007c90a12c80a4741b89b5498ac6 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:efe7851a66122b7eb251dbb27c4f2b3a2975149fb1de46ce6f4c0890076f23ea +size 4493605336 diff --git a/models/Qwen-Image-2.1/dit_6bpp.json b/models/Qwen-Image-2.1/dit_6bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..d0ae16132c8fdfcdfabdcbff8b1f0c30037d7ba6 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_6bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "0267340b4f952a934b068c54297ca4fe", + "model_name": "qwen_image_21_dit", + "model_class": "diffsynth.models.qwen_image_21_dit.QwenImage21DiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 6.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/dit_6bpp.safetensors b/models/Qwen-Image-2.1/dit_6bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d1da8526d6e64fc03941f925d54748414ad49a98 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_6bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c1a80c526fdc815acfc3ca876176e6e55c845e8c220f62bf58a180685f9907f4 +size 5394764344 diff --git a/models/Qwen-Image-2.1/dit_7bpp.json b/models/Qwen-Image-2.1/dit_7bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..04effae1591e1600e236d2ff9acfe1f18f731ad4 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_7bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "64154d85e7844d12fc6302905e3de4b2", + "model_name": "qwen_image_21_dit", + "model_class": "diffsynth.models.qwen_image_21_dit.QwenImage21DiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 7.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/dit_7bpp.safetensors b/models/Qwen-Image-2.1/dit_7bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d6edc6c5e464cbbbd5401e1a39ce01cb746ff7b5 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_7bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0beb9a3c2228d24f7571e0fd41e968dffae080e9c24946586b0c3cf5b267cac6 +size 6301372920 diff --git a/models/Qwen-Image-2.1/dit_8bpp.json b/models/Qwen-Image-2.1/dit_8bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..571c87eed2b4cb9ac54cf859a31829379907bc48 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_8bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "64a9bff895a4936af133c8fe40642c24", + "model_name": "qwen_image_21_dit", + "model_class": "diffsynth.models.qwen_image_21_dit.QwenImage21DiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/dit_8bpp.safetensors b/models/Qwen-Image-2.1/dit_8bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d7a078b21f411748c60466712cac4a842f0b2bc1 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_8bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2d380d5ab75eb9c614df9b636143ba0508b3cbf94fa6cb9280459d40bd77fd36 +size 7175703336 diff --git a/models/Qwen-Image-2.1/dit_extreme_3.0bpp.json b/models/Qwen-Image-2.1/dit_extreme_3.0bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..d8b268cc4af6128caef4cf7d3e5fc7d11a7eb853 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_extreme_3.0bpp.json @@ -0,0 +1,262 @@ +{ + "model_hash": "41bac1ce260e3572cf0ba1ebab255a09", + "model_name": "qwen_image_21_dit", + "model_class": "diffsynth.models.qwen_image_21_dit.QwenImage21DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 2.985497495701577, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "img_in", + "transformer_blocks.0.attn.to_k", + "transformer_blocks.0.attn.to_out.0", + "transformer_blocks.0.attn.to_q", + "transformer_blocks.0.attn.to_v", + "transformer_blocks.0.img_mlp.gate_layer", + "transformer_blocks.0.img_mlp.out", + "transformer_blocks.0.img_mlp.proj", + "transformer_blocks.1.attn.to_k", + "transformer_blocks.1.attn.to_out.0", + "transformer_blocks.1.attn.to_q", + "transformer_blocks.1.attn.to_v", + "transformer_blocks.1.img_mlp.gate_layer", + "transformer_blocks.1.img_mlp.out", + "transformer_blocks.1.img_mlp.proj", + "transformer_blocks.10.attn.to_k", + "transformer_blocks.10.attn.to_out.0", + "transformer_blocks.10.attn.to_q", + "transformer_blocks.10.attn.to_v", + "transformer_blocks.10.img_mlp.gate_layer", + "transformer_blocks.10.img_mlp.out", + "transformer_blocks.10.img_mlp.proj", + "transformer_blocks.11.attn.to_k", + "transformer_blocks.11.attn.to_out.0", + "transformer_blocks.11.attn.to_q", + "transformer_blocks.11.attn.to_v", + "transformer_blocks.11.img_mlp.gate_layer", + "transformer_blocks.11.img_mlp.out", + "transformer_blocks.11.img_mlp.proj", + "transformer_blocks.12.attn.to_k", + "transformer_blocks.12.attn.to_out.0", + "transformer_blocks.12.attn.to_q", + "transformer_blocks.12.attn.to_v", + "transformer_blocks.12.img_mlp.gate_layer", + "transformer_blocks.12.img_mlp.out", + "transformer_blocks.12.img_mlp.proj", + "transformer_blocks.13.attn.to_k", + "transformer_blocks.13.attn.to_out.0", + "transformer_blocks.13.attn.to_q", + "transformer_blocks.13.attn.to_v", + "transformer_blocks.13.img_mlp.gate_layer", + "transformer_blocks.13.img_mlp.out", + "transformer_blocks.13.img_mlp.proj", + "transformer_blocks.14.attn.to_k", + "transformer_blocks.14.attn.to_out.0", + "transformer_blocks.14.attn.to_q", + "transformer_blocks.14.attn.to_v", + "transformer_blocks.14.img_mlp.gate_layer", + "transformer_blocks.14.img_mlp.out", + "transformer_blocks.14.img_mlp.proj", + "transformer_blocks.15.attn.to_k", + "transformer_blocks.15.attn.to_out.0", + "transformer_blocks.15.attn.to_q", + "transformer_blocks.15.attn.to_v", + "transformer_blocks.15.img_mlp.gate_layer", + "transformer_blocks.15.img_mlp.out", + "transformer_blocks.15.img_mlp.proj", + "transformer_blocks.16.attn.to_k", + "transformer_blocks.16.attn.to_out.0", + "transformer_blocks.16.attn.to_q", + "transformer_blocks.16.attn.to_v", + "transformer_blocks.16.img_mlp.gate_layer", + "transformer_blocks.16.img_mlp.out", + "transformer_blocks.16.img_mlp.proj", + "transformer_blocks.17.attn.to_k", + "transformer_blocks.17.attn.to_out.0", + "transformer_blocks.17.attn.to_q", + "transformer_blocks.17.attn.to_v", + "transformer_blocks.17.img_mlp.gate_layer", + "transformer_blocks.17.img_mlp.out", + "transformer_blocks.17.img_mlp.proj", + "transformer_blocks.18.attn.to_k", + "transformer_blocks.18.attn.to_out.0", + "transformer_blocks.18.attn.to_q", + "transformer_blocks.18.attn.to_v", + "transformer_blocks.18.img_mlp.gate_layer", + "transformer_blocks.18.img_mlp.out", + "transformer_blocks.18.img_mlp.proj", + "transformer_blocks.19.attn.to_k", + "transformer_blocks.19.attn.to_out.0", + "transformer_blocks.19.attn.to_q", + "transformer_blocks.19.attn.to_v", + "transformer_blocks.19.img_mlp.gate_layer", + "transformer_blocks.19.img_mlp.out", + "transformer_blocks.19.img_mlp.proj", + "transformer_blocks.2.attn.to_k", + "transformer_blocks.2.attn.to_out.0", + "transformer_blocks.2.attn.to_q", + "transformer_blocks.2.attn.to_v", + "transformer_blocks.2.img_mlp.gate_layer", + "transformer_blocks.2.img_mlp.out", + "transformer_blocks.2.img_mlp.proj", + "transformer_blocks.20.attn.to_k", + "transformer_blocks.20.attn.to_out.0", + "transformer_blocks.20.attn.to_q", + "transformer_blocks.20.attn.to_v", + "transformer_blocks.20.img_mlp.gate_layer", + "transformer_blocks.20.img_mlp.out", + "transformer_blocks.20.img_mlp.proj", + "transformer_blocks.21.attn.to_k", + "transformer_blocks.21.attn.to_out.0", + "transformer_blocks.21.attn.to_q", + "transformer_blocks.21.attn.to_v", + "transformer_blocks.21.img_mlp.gate_layer", + "transformer_blocks.21.img_mlp.out", + "transformer_blocks.21.img_mlp.proj", + "transformer_blocks.22.attn.to_k", + "transformer_blocks.22.attn.to_out.0", + "transformer_blocks.22.attn.to_q", + "transformer_blocks.22.attn.to_v", + "transformer_blocks.3.attn.to_k", + "transformer_blocks.3.attn.to_out.0", + "transformer_blocks.3.attn.to_q", + "transformer_blocks.3.attn.to_v", + "transformer_blocks.3.img_mlp.gate_layer", + "transformer_blocks.3.img_mlp.out", + "transformer_blocks.3.img_mlp.proj", + "transformer_blocks.4.attn.to_k", + "transformer_blocks.4.attn.to_out.0", + "transformer_blocks.4.attn.to_q", + "transformer_blocks.4.attn.to_v", + "transformer_blocks.4.img_mlp.gate_layer", + "transformer_blocks.4.img_mlp.out", + "transformer_blocks.4.img_mlp.proj", + "transformer_blocks.5.attn.to_k", + "transformer_blocks.5.attn.to_out.0", + "transformer_blocks.5.attn.to_q", + "transformer_blocks.5.attn.to_v", + "transformer_blocks.5.img_mlp.gate_layer", + "transformer_blocks.5.img_mlp.out", + "transformer_blocks.5.img_mlp.proj", + "transformer_blocks.6.attn.to_k", + "transformer_blocks.6.attn.to_out.0", + "transformer_blocks.6.attn.to_q", + "transformer_blocks.6.attn.to_v", + "transformer_blocks.6.img_mlp.gate_layer", + "transformer_blocks.6.img_mlp.out", + "transformer_blocks.6.img_mlp.proj", + "transformer_blocks.7.attn.to_k", + "transformer_blocks.7.attn.to_out.0", + "transformer_blocks.7.attn.to_q", + "transformer_blocks.7.attn.to_v", + "transformer_blocks.7.img_mlp.gate_layer", + "transformer_blocks.7.img_mlp.out", + "transformer_blocks.7.img_mlp.proj", + "transformer_blocks.8.attn.to_k", + "transformer_blocks.8.attn.to_out.0", + "transformer_blocks.8.attn.to_q", + "transformer_blocks.8.attn.to_v", + "transformer_blocks.8.img_mlp.gate_layer", + "transformer_blocks.8.img_mlp.out", + "transformer_blocks.8.img_mlp.proj", + "transformer_blocks.9.attn.to_k", + "transformer_blocks.9.attn.to_out.0", + "transformer_blocks.9.attn.to_q", + "transformer_blocks.9.attn.to_v", + "transformer_blocks.9.img_mlp.gate_layer", + "transformer_blocks.9.img_mlp.out", + "transformer_blocks.9.img_mlp.proj", + "txt_in.in_layer", + "txt_in.out_layer", + "proj_out", + "transformer_blocks.22.img_mlp.gate_layer", + "transformer_blocks.22.img_mlp.out", + "transformer_blocks.22.img_mlp.proj", + "transformer_blocks.23.attn.to_k", + "transformer_blocks.23.attn.to_out.0", + "transformer_blocks.23.attn.to_q", + "transformer_blocks.23.attn.to_v", + "transformer_blocks.23.img_mlp.gate_layer", + "transformer_blocks.23.img_mlp.out", + "transformer_blocks.23.img_mlp.proj", + "transformer_blocks.24.attn.to_k", + "transformer_blocks.24.attn.to_out.0", + "transformer_blocks.24.attn.to_q", + "transformer_blocks.24.attn.to_v", + "transformer_blocks.24.img_mlp.gate_layer", + "transformer_blocks.24.img_mlp.out", + "transformer_blocks.24.img_mlp.proj", + "transformer_blocks.25.attn.to_k", + "transformer_blocks.25.attn.to_out.0", + "transformer_blocks.25.attn.to_q", + "transformer_blocks.25.attn.to_v", + "transformer_blocks.25.img_mlp.gate_layer", + "transformer_blocks.25.img_mlp.out", + "transformer_blocks.25.img_mlp.proj", + "transformer_blocks.26.attn.to_k", + "transformer_blocks.26.attn.to_out.0", + "transformer_blocks.26.attn.to_q", + "transformer_blocks.26.attn.to_v", + "transformer_blocks.26.img_mlp.gate_layer", + "transformer_blocks.26.img_mlp.out", + "transformer_blocks.26.img_mlp.proj", + "transformer_blocks.27.attn.to_k", + "transformer_blocks.27.attn.to_out.0", + "transformer_blocks.27.attn.to_q", + "transformer_blocks.27.attn.to_v", + "transformer_blocks.27.img_mlp.gate_layer", + "transformer_blocks.27.img_mlp.out", + "transformer_blocks.27.img_mlp.proj", + "transformer_blocks.28.attn.to_k", + "transformer_blocks.28.attn.to_out.0", + "transformer_blocks.28.attn.to_q", + "transformer_blocks.28.attn.to_v", + "transformer_blocks.28.img_mlp.gate_layer", + "transformer_blocks.28.img_mlp.out", + "transformer_blocks.28.img_mlp.proj", + "transformer_blocks.29.attn.to_k", + "transformer_blocks.29.attn.to_out.0", + "transformer_blocks.29.attn.to_q", + "transformer_blocks.29.attn.to_v", + "transformer_blocks.29.img_mlp.gate_layer", + "transformer_blocks.29.img_mlp.out", + "transformer_blocks.29.img_mlp.proj", + "transformer_blocks.30.attn.to_k", + "transformer_blocks.30.attn.to_out.0", + "transformer_blocks.30.attn.to_q", + "transformer_blocks.30.attn.to_v", + "transformer_blocks.30.img_mlp.gate_layer", + "transformer_blocks.30.img_mlp.out", + "transformer_blocks.30.img_mlp.proj", + "transformer_blocks.31.attn.to_k", + "transformer_blocks.31.attn.to_out.0", + "transformer_blocks.31.attn.to_q", + "transformer_blocks.31.attn.to_v", + "transformer_blocks.31.img_mlp.gate_layer", + "transformer_blocks.31.img_mlp.out", + "transformer_blocks.31.img_mlp.proj" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "modulation.1", + "time_text_embed.timestep_embedder.linear_1", + "time_text_embed.timestep_embedder.linear_2", + "norm_out.linear" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/dit_extreme_3.0bpp.safetensors b/models/Qwen-Image-2.1/dit_extreme_3.0bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6fc2455b40766fc027e15a8f2e9aaeb9790c574f --- /dev/null +++ b/models/Qwen-Image-2.1/dit_extreme_3.0bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5f6ce6ae4d431622be7024d8d019d3e01d90d64a7e13259af328f53bc2c59216 +size 2688642472 diff --git a/models/Qwen-Image-2.1/dit_fp8_5bpp.json b/models/Qwen-Image-2.1/dit_fp8_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..d7a80c57ee8c070e94d28ba42221de2624632b0c --- /dev/null +++ b/models/Qwen-Image-2.1/dit_fp8_5bpp.json @@ -0,0 +1,32 @@ +{ + "model_hash": "643cda205bc74f9f7fb5c0b8452b6073", + "model_name": "qwen_image_21_dit", + "model_class": "diffsynth.models.qwen_image_21_dit.QwenImage21DiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant_fp8", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "exclude_modules": [ + "img_in" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "img_in" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/dit_fp8_5bpp.safetensors b/models/Qwen-Image-2.1/dit_fp8_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3ab10e42c689a4cc654e8ff64cbc8fd6607e8bc1 --- /dev/null +++ b/models/Qwen-Image-2.1/dit_fp8_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d489ef801052b9384ddcedfead8b8eaafe6b720054f1ae46df087175616b61a0 +size 4495548712 diff --git a/models/Qwen-Image-2.1/text_encoder_4bpp.json b/models/Qwen-Image-2.1/text_encoder_4bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..4d1d83386c0f6f4a7ea5bf46252bc213076f100a --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_4bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "a6ade35feb1d907538ab41b458321e9b", + "model_name": "qwen_image_21_text_encoder", + "model_class": "diffsynth.models.qwen_image_21_text_encoder.QwenImage21TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/text_encoder_4bpp.safetensors b/models/Qwen-Image-2.1/text_encoder_4bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5d9ba0963b06ebf72b9852698c148b33005dce55 --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_4bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:22f6863ae59f2e925048cfb6d69f5538f259e56d0c6b1f0a84eb87494c867715 +size 5359260472 diff --git a/models/Qwen-Image-2.1/text_encoder_5bpp.json b/models/Qwen-Image-2.1/text_encoder_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..7075f981144f316fddbbfe45e6209e0ef4a2bc15 --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_5bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "e579ef014278d930f2634c1f7adb6dcf", + "model_name": "qwen_image_21_text_encoder", + "model_class": "diffsynth.models.qwen_image_21_text_encoder.QwenImage21TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/text_encoder_5bpp.safetensors b/models/Qwen-Image-2.1/text_encoder_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1aaa9b0ce009f961b23899f02089b50ad8131403 --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:47c9481afeda0558cbd2f5f901f016f61e0fbb6bb04ddd83d557a631523ec436 +size 6405854632 diff --git a/models/Qwen-Image-2.1/text_encoder_6bpp.json b/models/Qwen-Image-2.1/text_encoder_6bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..d001d9894673e6fb515d432e880bc91ecd2a6278 --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_6bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "2236acb9951a522384a2a393521221fd", + "model_name": "qwen_image_21_text_encoder", + "model_class": "diffsynth.models.qwen_image_21_text_encoder.QwenImage21TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 6.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/text_encoder_6bpp.safetensors b/models/Qwen-Image-2.1/text_encoder_6bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0759dc3d3b8b7cf35fc8cf6bff97624e17b6628d --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_6bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4cdf18465eb6078830ddbe0ade641b205bcb39b34359497af2e814b8651b7293 +size 7443117456 diff --git a/models/Qwen-Image-2.1/text_encoder_7bpp.json b/models/Qwen-Image-2.1/text_encoder_7bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..2c70c35e091225b9ab14e37724075edee0e91eae --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_7bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "0dea9a7b8f2ef23713e949f6a4c871ce", + "model_name": "qwen_image_21_text_encoder", + "model_class": "diffsynth.models.qwen_image_21_text_encoder.QwenImage21TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 7.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/text_encoder_7bpp.safetensors b/models/Qwen-Image-2.1/text_encoder_7bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8360602d137d3689c3dca4412e7f3d1ded76d951 --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_7bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4dd9f797a1d24f18d900fed4b4b77b036748bbd416186ccdf1a8210658bfc26c +size 8470737008 diff --git a/models/Qwen-Image-2.1/text_encoder_8bpp.json b/models/Qwen-Image-2.1/text_encoder_8bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..16e12b90fbdf2489ffb9ce4276d1f548f2b78459 --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_8bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "6f2b057baeaab46e5cd89a00858592c6", + "model_name": "qwen_image_21_text_encoder", + "model_class": "diffsynth.models.qwen_image_21_text_encoder.QwenImage21TextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Qwen-Image-2.1/text_encoder_8bpp.safetensors b/models/Qwen-Image-2.1/text_encoder_8bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..10262d5126e3900beeb86e8a2c863f32f1fd9325 --- /dev/null +++ b/models/Qwen-Image-2.1/text_encoder_8bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:df72cc818599a87d1922b581ac0bf796e6c85bd50e8695a46727ae42afc96e0e +size 9467907376 diff --git a/models/Z-Image-Turbo/dit_4bpp.json b/models/Z-Image-Turbo/dit_4bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..ed6bf0cbe97e994dbfe82eb58200b42f54e4786f --- /dev/null +++ b/models/Z-Image-Turbo/dit_4bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "1a5ac048b70fc00c06e2d39096f376c0", + "model_name": "z_image_dit", + "model_class": "diffsynth.models.z_image_dit.ZImageDiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/dit_4bpp.safetensors b/models/Z-Image-Turbo/dit_4bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..70c145227d6e5f5ea5fd066cd68d5bdf604e501e --- /dev/null +++ b/models/Z-Image-Turbo/dit_4bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:751daf1d3ac013421d801b707527f38904838189b2bd64d2ff222b6b683457ba +size 3097388928 diff --git a/models/Z-Image-Turbo/dit_5bpp.json b/models/Z-Image-Turbo/dit_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..efaf38d4a4e6e1a2f1ad98803864b741102ad5a3 --- /dev/null +++ b/models/Z-Image-Turbo/dit_5bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "f13192c745eae7e663460179937a647f", + "model_name": "z_image_dit", + "model_class": "diffsynth.models.z_image_dit.ZImageDiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/dit_5bpp.safetensors b/models/Z-Image-Turbo/dit_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..7b599604601f4d5f3e2b834607105a6b20d648bf --- /dev/null +++ b/models/Z-Image-Turbo/dit_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f84a44f130f657c6364a8a40de9c4a23f029e7dfc9afe80c6d7bb10feb56fb66 +size 3883934968 diff --git a/models/Z-Image-Turbo/dit_6bpp.json b/models/Z-Image-Turbo/dit_6bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..3a98e32fe06d7469028947109ba456a572df7578 --- /dev/null +++ b/models/Z-Image-Turbo/dit_6bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "b0750ddd71937c9cfec9f25a49887711", + "model_name": "z_image_dit", + "model_class": "diffsynth.models.z_image_dit.ZImageDiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 6.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/dit_6bpp.safetensors b/models/Z-Image-Turbo/dit_6bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8a9b2a0c1ea97c655fddf8ebdcf5a13032893b4e --- /dev/null +++ b/models/Z-Image-Turbo/dit_6bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:41aa8039609d1401f03ff0436c10e48198348fe9d5779d601b3803bb27487180 +size 4660190824 diff --git a/models/Z-Image-Turbo/dit_7bpp.json b/models/Z-Image-Turbo/dit_7bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..0a429c1eed1a5901d34f4adcf62d3ab25c0a3fe6 --- /dev/null +++ b/models/Z-Image-Turbo/dit_7bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "f89cfa0412d58e27fbeeafa7e63caffd", + "model_name": "z_image_dit", + "model_class": "diffsynth.models.z_image_dit.ZImageDiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 7.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/dit_7bpp.safetensors b/models/Z-Image-Turbo/dit_7bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6cf66c90f17abb967087f347cbbacbc0d2554dd2 --- /dev/null +++ b/models/Z-Image-Turbo/dit_7bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e15d301fdac4053550faddd41652cf3301f43a7167df0a0b3e832f34c7af6565 +size 5440907736 diff --git a/models/Z-Image-Turbo/dit_8bpp.json b/models/Z-Image-Turbo/dit_8bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..f11b6104a6b2bcb0e4b1ebf83fc4d323a1eaabf4 --- /dev/null +++ b/models/Z-Image-Turbo/dit_8bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "3af3efca34bc019f0ff7f98598e87e36", + "model_name": "z_image_dit", + "model_class": "diffsynth.models.z_image_dit.ZImageDiT", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/dit_8bpp.safetensors b/models/Z-Image-Turbo/dit_8bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..7d9674e77115ab4c6ea4abb5cbb3e0799a26b800 --- /dev/null +++ b/models/Z-Image-Turbo/dit_8bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06a054997e2adf5c4d42a82a0e2ae469d176a4b8647851e7016a44909710ea08 +size 6202326376 diff --git a/models/Z-Image-Turbo/dit_extreme_2.3bpp.json b/models/Z-Image-Turbo/dit_extreme_2.3bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..6bbbd4137a5075c1cde5813945b5b61cc1ccfb53 --- /dev/null +++ b/models/Z-Image-Turbo/dit_extreme_2.3bpp.json @@ -0,0 +1,316 @@ +{ + "model_hash": "e576b74953084f4b29de87857e612650", + "model_name": "z_image_dit", + "model_class": "diffsynth.models.z_image_dit.ZImageDiT", + "quant_config": { + "load_prequantized": true, + "configs": [ + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 2.193929292929293, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "noise_refiner.0.attention.to_q", + "noise_refiner.0.attention.to_k", + "noise_refiner.0.attention.to_v", + "noise_refiner.0.feed_forward.w1", + "noise_refiner.0.feed_forward.w2", + "noise_refiner.0.feed_forward.w3", + "noise_refiner.1.attention.to_q", + "noise_refiner.1.attention.to_k", + "noise_refiner.1.attention.to_v", + "noise_refiner.1.feed_forward.w1", + "noise_refiner.1.feed_forward.w2", + "noise_refiner.1.feed_forward.w3", + "context_refiner.0.attention.to_q", + "context_refiner.0.attention.to_k", + "context_refiner.0.attention.to_v", + "context_refiner.0.feed_forward.w1", + "context_refiner.0.feed_forward.w2", + "context_refiner.0.feed_forward.w3", + "context_refiner.1.attention.to_q", + "context_refiner.1.attention.to_k", + "context_refiner.1.attention.to_v", + "context_refiner.1.feed_forward.w1", + "context_refiner.1.feed_forward.w2", + "context_refiner.1.feed_forward.w3", + "layers.0.attention.to_q", + "layers.0.attention.to_k", + "layers.0.attention.to_v", + "layers.0.feed_forward.w1", + "layers.0.feed_forward.w2", + "layers.0.feed_forward.w3", + "layers.1.attention.to_q", + "layers.1.attention.to_k", + "layers.1.attention.to_v", + "layers.1.feed_forward.w1", + "layers.1.feed_forward.w2", + "layers.1.feed_forward.w3", + "layers.2.attention.to_q", + "layers.2.attention.to_k", + "layers.2.attention.to_v", + "layers.2.feed_forward.w1", + "layers.2.feed_forward.w2", + "layers.2.feed_forward.w3", + "layers.3.attention.to_q", + "layers.3.attention.to_k", + "layers.3.attention.to_v", + "layers.3.feed_forward.w1", + "layers.3.feed_forward.w2", + "layers.3.feed_forward.w3", + "layers.4.attention.to_q", + "layers.4.attention.to_k", + "layers.4.attention.to_v", + "layers.4.feed_forward.w1", + "layers.4.feed_forward.w2", + "layers.4.feed_forward.w3", + "layers.5.attention.to_q", + "layers.5.attention.to_k", + "layers.5.attention.to_v", + "layers.5.feed_forward.w1", + "layers.5.feed_forward.w2", + "layers.5.feed_forward.w3", + "layers.6.attention.to_q", + "layers.6.attention.to_k", + "layers.6.attention.to_v", + "layers.6.feed_forward.w1", + "layers.6.feed_forward.w2", + "layers.6.feed_forward.w3", + "layers.7.attention.to_q", + "layers.7.attention.to_k", + "layers.7.attention.to_v", + "layers.7.feed_forward.w1", + "layers.7.feed_forward.w2", + "layers.7.feed_forward.w3", + "layers.8.attention.to_q", + "layers.8.attention.to_k", + "layers.8.attention.to_v", + "layers.8.feed_forward.w1", + "layers.8.feed_forward.w2", + "layers.8.feed_forward.w3", + "layers.9.attention.to_q", + "layers.9.attention.to_k", + "layers.9.attention.to_v", + "layers.9.feed_forward.w1", + "layers.9.feed_forward.w2", + "layers.9.feed_forward.w3", + "layers.10.attention.to_q", + "layers.10.attention.to_k", + "layers.10.attention.to_v", + "layers.10.feed_forward.w1", + "layers.10.feed_forward.w2", + "layers.10.feed_forward.w3", + "layers.11.attention.to_q", + "layers.11.attention.to_k", + "layers.11.attention.to_v", + "layers.11.feed_forward.w1", + "layers.11.feed_forward.w2", + "layers.11.feed_forward.w3", + "layers.12.attention.to_q", + "layers.12.attention.to_k", + "layers.12.attention.to_v", + "layers.12.feed_forward.w1", + "layers.12.feed_forward.w2", + "layers.12.feed_forward.w3", + "layers.13.attention.to_q", + "layers.13.attention.to_k", + "layers.13.attention.to_v", + "layers.13.feed_forward.w1", + "layers.13.feed_forward.w2", + "layers.13.feed_forward.w3", + "layers.14.attention.to_q", + "layers.14.attention.to_k", + "layers.14.attention.to_v", + "layers.14.feed_forward.w1", + "layers.14.feed_forward.w2", + "layers.14.feed_forward.w3", + "layers.15.attention.to_q", + "layers.15.attention.to_k", + "layers.15.attention.to_v", + "layers.15.feed_forward.w1", + "layers.15.feed_forward.w2", + "layers.15.feed_forward.w3", + "layers.16.attention.to_q", + "layers.16.attention.to_k", + "layers.16.attention.to_v", + "layers.16.feed_forward.w1", + "layers.16.feed_forward.w2", + "layers.16.feed_forward.w3", + "layers.17.attention.to_q", + "layers.17.attention.to_k", + "layers.17.attention.to_v", + "layers.17.feed_forward.w1", + "layers.17.feed_forward.w2", + "layers.17.feed_forward.w3", + "layers.18.attention.to_q", + "layers.18.attention.to_k", + "layers.18.attention.to_v", + "layers.18.feed_forward.w1", + "layers.18.feed_forward.w2", + "layers.18.feed_forward.w3", + "layers.19.attention.to_q", + "layers.19.attention.to_k", + "layers.19.attention.to_v", + "layers.19.feed_forward.w1", + "layers.19.feed_forward.w2", + "layers.19.feed_forward.w3", + "layers.20.attention.to_q", + "layers.20.attention.to_k", + "layers.20.attention.to_v", + "layers.20.feed_forward.w1", + "layers.20.feed_forward.w2", + "layers.20.feed_forward.w3", + "layers.21.attention.to_q", + "layers.21.attention.to_k", + "layers.21.attention.to_v", + "layers.21.feed_forward.w1", + "layers.21.feed_forward.w2", + "layers.21.feed_forward.w3", + "layers.22.attention.to_q", + "layers.22.attention.to_k", + "layers.22.attention.to_v", + "layers.22.feed_forward.w1", + "layers.22.feed_forward.w2", + "layers.22.feed_forward.w3", + "layers.23.attention.to_q", + "layers.23.attention.to_k", + "layers.23.attention.to_v", + "layers.23.feed_forward.w1", + "layers.23.feed_forward.w2", + "layers.23.feed_forward.w3", + "layers.24.attention.to_q", + "layers.24.attention.to_k", + "layers.24.attention.to_v", + "layers.24.feed_forward.w1", + "layers.24.feed_forward.w2", + "layers.24.feed_forward.w3", + "layers.25.attention.to_q", + "layers.25.attention.to_k", + "layers.25.attention.to_v", + "layers.25.feed_forward.w1", + "layers.25.feed_forward.w2", + "layers.25.feed_forward.w3", + "layers.26.attention.to_q", + "layers.26.attention.to_k", + "layers.26.attention.to_v", + "layers.26.feed_forward.w1", + "layers.26.feed_forward.w2", + "layers.26.feed_forward.w3", + "layers.27.attention.to_q", + "layers.27.attention.to_k", + "layers.27.attention.to_v", + "layers.27.feed_forward.w1", + "layers.27.feed_forward.w2", + "layers.27.feed_forward.w3", + "layers.28.attention.to_q", + "layers.28.attention.to_k", + "layers.28.attention.to_v", + "layers.28.feed_forward.w1", + "layers.28.feed_forward.w2", + "layers.28.feed_forward.w3", + "layers.29.attention.to_q", + "layers.29.attention.to_k", + "layers.29.attention.to_v", + "layers.29.feed_forward.w1", + "layers.29.feed_forward.w2", + "layers.29.feed_forward.w3" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 3.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "noise_refiner.0.attention.to_out.0", + "noise_refiner.1.attention.to_out.0", + "context_refiner.0.attention.to_out.0", + "context_refiner.1.attention.to_out.0", + "layers.0.attention.to_out.0", + "layers.1.attention.to_out.0", + "layers.2.attention.to_out.0", + "layers.3.attention.to_out.0", + "layers.4.attention.to_out.0", + "layers.5.attention.to_out.0", + "layers.6.attention.to_out.0", + "layers.7.attention.to_out.0", + "layers.8.attention.to_out.0", + "layers.9.attention.to_out.0", + "layers.10.attention.to_out.0", + "layers.11.attention.to_out.0", + "layers.12.attention.to_out.0", + "layers.13.attention.to_out.0", + "layers.14.attention.to_out.0", + "layers.15.attention.to_out.0", + "layers.16.attention.to_out.0", + "layers.17.attention.to_out.0", + "layers.18.attention.to_out.0", + "layers.19.attention.to_out.0", + "layers.20.attention.to_out.0", + "layers.21.attention.to_out.0", + "layers.22.attention.to_out.0", + "layers.23.attention.to_out.0", + "layers.24.attention.to_out.0", + "layers.25.attention.to_out.0", + "layers.26.attention.to_out.0", + "layers.27.attention.to_out.0", + "layers.28.attention.to_out.0", + "layers.29.attention.to_out.0" + ] + }, + { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "target_modules": [ + "all_x_embedder.2-1", + "all_final_layer.2-1.linear", + "all_final_layer.2-1.adaLN_modulation.1", + "noise_refiner.0.adaLN_modulation.0", + "noise_refiner.1.adaLN_modulation.0", + "t_embedder.mlp.0", + "t_embedder.mlp.2", + "cap_embedder.1", + "layers.0.adaLN_modulation.0", + "layers.1.adaLN_modulation.0", + "layers.2.adaLN_modulation.0", + "layers.3.adaLN_modulation.0", + "layers.4.adaLN_modulation.0", + "layers.5.adaLN_modulation.0", + "layers.6.adaLN_modulation.0", + "layers.7.adaLN_modulation.0", + "layers.8.adaLN_modulation.0", + "layers.9.adaLN_modulation.0", + "layers.10.adaLN_modulation.0", + "layers.11.adaLN_modulation.0", + "layers.12.adaLN_modulation.0", + "layers.13.adaLN_modulation.0", + "layers.14.adaLN_modulation.0", + "layers.15.adaLN_modulation.0", + "layers.16.adaLN_modulation.0", + "layers.17.adaLN_modulation.0", + "layers.18.adaLN_modulation.0", + "layers.19.adaLN_modulation.0", + "layers.20.adaLN_modulation.0", + "layers.21.adaLN_modulation.0", + "layers.22.adaLN_modulation.0", + "layers.23.adaLN_modulation.0", + "layers.24.adaLN_modulation.0", + "layers.25.adaLN_modulation.0", + "layers.26.adaLN_modulation.0", + "layers.27.adaLN_modulation.0", + "layers.28.adaLN_modulation.0", + "layers.29.adaLN_modulation.0" + ] + } + ] + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/dit_extreme_2.3bpp.safetensors b/models/Z-Image-Turbo/dit_extreme_2.3bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..def5be3897cdde4329142b3804f5ad1abcd47d01 --- /dev/null +++ b/models/Z-Image-Turbo/dit_extreme_2.3bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f7e73b79de4151cf01f66834faf0f670605420f899b22b0f27523c63a7e2889b +size 1786234528 diff --git a/models/Z-Image-Turbo/dit_fp8_5bpp.json b/models/Z-Image-Turbo/dit_fp8_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..93dc221427a4073e2963b8d767f7d5f48afb3380 --- /dev/null +++ b/models/Z-Image-Turbo/dit_fp8_5bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "df4d215bbfe3dc7a8d4a55162c42aba0", + "model_name": "z_image_dit", + "model_class": "diffsynth.models.z_image_dit.ZImageDiT", + "quant_config": { + "method": "entropack_lossy_quant_fp8", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/dit_fp8_5bpp.safetensors b/models/Z-Image-Turbo/dit_fp8_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4377704741b92ae46c8d58a0fd7b300b12e2d337 --- /dev/null +++ b/models/Z-Image-Turbo/dit_fp8_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5ae320750c36b2cd17707e6b0710747fe82e28e6dff4c0cdb94088f946a8f63 +size 3887473936 diff --git a/models/Z-Image-Turbo/text_encoder_4bpp.json b/models/Z-Image-Turbo/text_encoder_4bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..1691560a5f41c7418e2a3561654c983b0a831dc7 --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_4bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "7322a84f0b4eadef9054d8554b9a81d0", + "model_name": "z_image_text_encoder", + "model_class": "diffsynth.models.z_image_text_encoder.ZImageTextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 4.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/text_encoder_4bpp.safetensors b/models/Z-Image-Turbo/text_encoder_4bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..879ec0e0f913d1b51d0e2e9cc3987f942728e0b8 --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_4bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e6957e1f5de901b95ce9272fb2bf0ffee841d9c33e0e2bfa72bbf7c258668906 +size 2607910416 diff --git a/models/Z-Image-Turbo/text_encoder_5bpp.json b/models/Z-Image-Turbo/text_encoder_5bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..37cfd7e3816729d748ce8d9dc4b674679ca7005d --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_5bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "75f7ab81e4fb8c867a35e7f07fda289e", + "model_name": "z_image_text_encoder", + "model_class": "diffsynth.models.z_image_text_encoder.ZImageTextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 5.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/text_encoder_5bpp.safetensors b/models/Z-Image-Turbo/text_encoder_5bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..862410dcf8cfa0359197bbcd5c7d70cb82fd2086 --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_5bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:30817126e2856f6736cd6d0898432eec2005ec97d5a73780b833f1a5dd5dc0a4 +size 3072497392 diff --git a/models/Z-Image-Turbo/text_encoder_6bpp.json b/models/Z-Image-Turbo/text_encoder_6bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..856be4e17a5358f01f8ba38ba0655ba27681e73c --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_6bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "56d854d1df20191413652f557a567785", + "model_name": "z_image_text_encoder", + "model_class": "diffsynth.models.z_image_text_encoder.ZImageTextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 6.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/text_encoder_6bpp.safetensors b/models/Z-Image-Turbo/text_encoder_6bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..16c1d444cacdb1f7b9da0bece36b1099f01ad727 --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_6bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:071dbad102d29b78712285e80f39cd4af975ed7bae9e6eaf348612ba4c390af3 +size 3529676480 diff --git a/models/Z-Image-Turbo/text_encoder_7bpp.json b/models/Z-Image-Turbo/text_encoder_7bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..1c22f16eef87029f5f258508f81a5ccb11617972 --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_7bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "c65ff5fa3889a162d01896906bb41127", + "model_name": "z_image_text_encoder", + "model_class": "diffsynth.models.z_image_text_encoder.ZImageTextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 7.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/text_encoder_7bpp.safetensors b/models/Z-Image-Turbo/text_encoder_7bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..571433b4543a682aeec16d0abd9216bee5dbe269 --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_7bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:347ff52b3647acbbd9ffc0c290f54fc307bd4218f288c4b0ad39caba347a3be4 +size 3985048888 diff --git a/models/Z-Image-Turbo/text_encoder_8bpp.json b/models/Z-Image-Turbo/text_encoder_8bpp.json new file mode 100644 index 0000000000000000000000000000000000000000..2cd92cef8a0d927d3a454ac4c05fdb82c49da9ae --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_8bpp.json @@ -0,0 +1,14 @@ +{ + "model_hash": "102562327096b07f58cfff3e5e2223f4", + "model_name": "z_image_text_encoder", + "model_class": "diffsynth.models.z_image_text_encoder.ZImageTextEncoder", + "quant_config": { + "method": "entropack_lossy_quant", + "backend_config_kwargs": { + "target_bpp": 8.0, + "row_rdo_iterations": 2, + "row_rdo_candidates": 20 + }, + "load_prequantized": true + } +} \ No newline at end of file diff --git a/models/Z-Image-Turbo/text_encoder_8bpp.safetensors b/models/Z-Image-Turbo/text_encoder_8bpp.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c0b21d35d8028541073b31f0855f0fdf4f558c0c --- /dev/null +++ b/models/Z-Image-Turbo/text_encoder_8bpp.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:219c3d0c90549966d7ad352d0857c79b327f72029904f83fc11d6aec915cb00a +size 4430436232