Mirror model files from microsoft/SciForma-Base

README.md

@@ -0,0 +1,76 @@
+---
+license: other
+license_name: flux-non-commercial-license-v2.1
+license_link: https://huggingface.co/black-forest-labs/FLUX.2-klein-base-9B/blob/main/LICENSE.md
+language:
+- en
+base_model:
+- black-forest-labs/FLUX.2-klein-base-9B
+tags:
+- text-to-image
+- scientific-diagrams
+- methodology-figures
+- flux
+- diffusers
+pipeline_tag: text-to-image
+---
+
+# SciForma-Base
+
+**SciForma-Base** is the Stage 1+2 SFT checkpoint of [SciForma: Structure-Faithful Generation of Scientific Diagrams](https://github.com/microsoft/SciForma), fine-tuned from [FLUX.2-klein-base-9B](https://huggingface.co/black-forest-labs/FLUX.2-klein-base-9B).
+
+| Model | Overall ↑ | Comp. ↑ | Arrow ↑ | Text ↑ |
+|-------|-----------|---------|---------|--------|
+| **SciForma-Base** | **67.59** | 73.52 | 64.64 | 63.84 |
+| SciForma-9B | 69.51 | 74.49 | 66.46 | 67.00 |
+| FLUX.2-klein-base-9B (no fine-tuning) | 33.87 | 42.80 | 25.20 | 23.60 |
+
+Scores on SciFormaBench-2K (GPT-5.4 judge, split_dims=True).
+
+## Usage
+
+This repository contains only the fine-tuned **transformer** weights. The text encoder, tokenizer, VAE, and scheduler are inherited from the base model.
+
+```python
+import torch
+from diffusers import Flux2KleinPipeline, Flux2Transformer2DModel
+
+# Load fine-tuned transformer
+transformer = Flux2Transformer2DModel.from_pretrained(
+    "microsoft/SciForma-Base",
+    subfolder="transformer",
+    torch_dtype=torch.bfloat16,
+)
+
+# Load full pipeline with base model components
+pipe = Flux2KleinPipeline.from_pretrained(
+    "black-forest-labs/FLUX.2-klein-base-9B",
+    transformer=transformer,
+    torch_dtype=torch.bfloat16,
+)
+pipe.enable_model_cpu_offload()
+
+# Generate
+prompt = "The figure illustrates a transformer architecture with encoder and decoder..."
+image = pipe(
+    prompt,
+    height=576,
+    width=1008,
+    num_inference_steps=28,
+    guidance_scale=4.0,
+).images[0]
+image.save("output.png")
+```
+
+## Training Details
+
+- **Base model**: FLUX.2-klein-base-9B (9.08B parameters)
+- **Stage 1**: Full fine-tuning on 661K arXiv methodology diagrams at 768px (200K steps, 8× B200)
+- **Stage 2**: Joint generation + editing fine-tuning on 244K high-quality 1024px diagrams + 70K editing triplets (120K steps, 8× B200)
+- **Training data**: [microsoft/SciFormaData-700K](https://huggingface.co/datasets/microsoft/SciFormaData-700K)
+
+## License
+
+Released under the [FLUX Non-Commercial License v2.1](https://huggingface.co/black-forest-labs/FLUX.2-klein-base-9B/blob/main/LICENSE.md). Non-commercial research use only.
+
+> This FLUX Model is licensed by Black Forest Labs Inc. under the FLUX Non-Commercial License. Copyright Black Forest Labs Inc. This model has been modified by Microsoft Research Asia.

model_index.json

@@ -0,0 +1,25 @@
+{
+  "_class_name": "Flux2KleinPipeline",
+  "_diffusers_version": "0.37.0.dev0",
+  "scheduler": [
+    "diffusers",
+    "FlowMatchEulerDiscreteScheduler"
+  ],
+  "text_encoder": [
+    "transformers",
+    "Qwen3ForCausalLM"
+  ],
+  "tokenizer": [
+    "transformers",
+    "Qwen2TokenizerFast"
+  ],
+  "transformer": [
+    "diffusers",
+    "Flux2Transformer2DModel"
+  ],
+  "vae": [
+    "diffusers",
+    "AutoencoderKLFlux2"
+  ],
+  "_note": "text_encoder/tokenizer/vae/scheduler shared with black-forest-labs/FLUX.2-klein-base-9B"
+}

transformer/config.json

@@ -0,0 +1,24 @@
+{
+  "_class_name": "Flux2Transformer2DModel",
+  "_diffusers_version": "0.37.0.dev0",
+  "_name_or_path": "/raid/yiyi/klein-9b-diffusers/transformer",
+  "attention_head_dim": 128,
+  "axes_dims_rope": [
+    32,
+    32,
+    32,
+    32
+  ],
+  "eps": 1e-06,
+  "guidance_embeds": false,
+  "in_channels": 128,
+  "joint_attention_dim": 12288,
+  "mlp_ratio": 3.0,
+  "num_attention_heads": 32,
+  "num_layers": 8,
+  "num_single_layers": 24,
+  "out_channels": null,
+  "patch_size": 1,
+  "rope_theta": 2000,
+  "timestep_guidance_channels": 256
+}

transformer/diffusion_pytorch_model-00001-of-00005.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96b7694d9199d8aaae3e75539e683eff23a029079d1e934c7ebcba809373422f
+size 3828363264

transformer/diffusion_pytorch_model-00002-of-00005.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c46f4163bc91c8a446b78efda0bb13d6f013bc866dd8d43e2ab70fd77e49cce5
+size 3791662800

transformer/diffusion_pytorch_model-00003-of-00005.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77a8b1fa50ddda38146aa38baa53f7d8caff1c063d31335588a7c7c6cd13efc3
+size 3925877600

transformer/diffusion_pytorch_model-00004-of-00005.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1457409322102210e4d82971eddc73918f9fcd395043f5da2ef6d76fe6d9534
+size 3925877632

transformer/diffusion_pytorch_model-00005-of-00005.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5474e684693b45d4bdf9532fb7d83a420e07d6c0ea430224ea0911e8daca5b28
+size 2685409392

transformer/diffusion_pytorch_model.safetensors.index.json

@@ -0,0 +1,240 @@
+{
+  "metadata": {
+    "total_size": 18157162496
+  },
+  "weight_map": {
+    "time_guidance_embed.timestep_embedder.linear_1.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "time_guidance_embed.timestep_embedder.linear_2.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "double_stream_modulation_img.linear.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "double_stream_modulation_txt.linear.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "single_stream_modulation.linear.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "x_embedder.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "context_embedder.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.to_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.to_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.to_v.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.norm_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.norm_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.to_out.0.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.norm_added_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.norm_added_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.add_q_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.add_k_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.add_v_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.attn.to_add_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.ff.linear_in.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.ff.linear_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.ff_context.linear_in.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.0.ff_context.linear_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.to_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.to_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.to_v.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.norm_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.norm_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.to_out.0.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.norm_added_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.norm_added_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.add_q_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.add_k_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.add_v_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.attn.to_add_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.ff.linear_in.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.ff.linear_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.ff_context.linear_in.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.1.ff_context.linear_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.to_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.to_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.to_v.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.norm_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.norm_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.to_out.0.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.norm_added_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.norm_added_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.add_q_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.add_k_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.add_v_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.attn.to_add_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.ff.linear_in.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.ff.linear_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.ff_context.linear_in.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.2.ff_context.linear_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.to_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.to_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.to_v.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.norm_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.norm_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.to_out.0.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.norm_added_q.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.norm_added_k.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.add_q_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.add_k_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.add_v_proj.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.attn.to_add_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.ff.linear_in.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.ff.linear_out.weight": "diffusion_pytorch_model-00001-of-00005.safetensors",
+    "transformer_blocks.3.ff_context.linear_in.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.3.ff_context.linear_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.to_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.to_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.to_v.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.norm_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.norm_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.to_out.0.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.norm_added_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.norm_added_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.add_q_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.add_k_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.add_v_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.attn.to_add_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.ff.linear_in.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.ff.linear_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.ff_context.linear_in.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.4.ff_context.linear_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.to_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.to_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.to_v.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.norm_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.norm_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.to_out.0.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.norm_added_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.norm_added_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.add_q_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.add_k_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.add_v_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.attn.to_add_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.ff.linear_in.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.ff.linear_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.ff_context.linear_in.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.5.ff_context.linear_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.to_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.to_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.to_v.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.norm_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.norm_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.to_out.0.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.norm_added_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.norm_added_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.add_q_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.add_k_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.add_v_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.attn.to_add_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.ff.linear_in.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.ff.linear_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.ff_context.linear_in.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.6.ff_context.linear_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.to_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.to_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.to_v.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.norm_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.norm_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.to_out.0.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.norm_added_q.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.norm_added_k.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.add_q_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.add_k_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.add_v_proj.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.attn.to_add_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.ff.linear_in.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.ff.linear_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.ff_context.linear_in.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "transformer_blocks.7.ff_context.linear_out.weight": "diffusion_pytorch_model-00002-of-00005.safetensors",
+    "single_transformer_blocks.0.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.0.attn.norm_q.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.0.attn.norm_k.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.0.attn.to_out.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.1.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.1.attn.norm_q.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.1.attn.norm_k.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.1.attn.to_out.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.2.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.2.attn.norm_q.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.2.attn.norm_k.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.2.attn.to_out.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.3.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.3.attn.norm_q.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.3.attn.norm_k.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.3.attn.to_out.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.4.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.4.attn.norm_q.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.4.attn.norm_k.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.4.attn.to_out.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.5.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.5.attn.norm_q.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.5.attn.norm_k.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.5.attn.to_out.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.6.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.6.attn.norm_q.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.6.attn.norm_k.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.6.attn.to_out.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.7.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.7.attn.norm_q.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.7.attn.norm_k.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.7.attn.to_out.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.8.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.8.attn.norm_q.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.8.attn.norm_k.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.8.attn.to_out.weight": "diffusion_pytorch_model-00003-of-00005.safetensors",
+    "single_transformer_blocks.9.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.9.attn.norm_q.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.9.attn.norm_k.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.9.attn.to_out.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.10.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.10.attn.norm_q.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.10.attn.norm_k.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.10.attn.to_out.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.11.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.11.attn.norm_q.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.11.attn.norm_k.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.11.attn.to_out.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.12.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.12.attn.norm_q.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.12.attn.norm_k.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.12.attn.to_out.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.13.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.13.attn.norm_q.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.13.attn.norm_k.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.13.attn.to_out.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.14.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.14.attn.norm_q.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.14.attn.norm_k.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.14.attn.to_out.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.15.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.15.attn.norm_q.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.15.attn.norm_k.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.15.attn.to_out.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.16.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.16.attn.norm_q.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.16.attn.norm_k.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.16.attn.to_out.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.17.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.17.attn.norm_q.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.17.attn.norm_k.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.17.attn.to_out.weight": "diffusion_pytorch_model-00004-of-00005.safetensors",
+    "single_transformer_blocks.18.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.18.attn.norm_q.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.18.attn.norm_k.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.18.attn.to_out.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.19.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.19.attn.norm_q.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.19.attn.norm_k.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.19.attn.to_out.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.20.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.20.attn.norm_q.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.20.attn.norm_k.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.20.attn.to_out.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.21.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.21.attn.norm_q.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.21.attn.norm_k.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.21.attn.to_out.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.22.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.22.attn.norm_q.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.22.attn.norm_k.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.22.attn.to_out.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.23.attn.to_qkv_mlp_proj.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.23.attn.norm_q.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.23.attn.norm_k.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "single_transformer_blocks.23.attn.to_out.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "norm_out.linear.weight": "diffusion_pytorch_model-00005-of-00005.safetensors",
+    "proj_out.weight": "diffusion_pytorch_model-00005-of-00005.safetensors"
+  }
+}