Simplify evaluation notes and student attribution

README.md

@@ -27,7 +27,7 @@ This release contains one **LightningDiT-teacher CDMD student for class-conditio
 | [latents_stats.pt](latents_stats.pt) | 1.8 KB | Matching VA-VAE latent mean and standard deviation, required for image generation. |
 | [fid-50k-256.npz](fid-50k-256.npz) | 33.6 MB | ImageNet reference statistics for FID evaluation with the pinned metric implementation. |
 
-The statistics for this model are `.pt` and `.npz` files; no statistics `.pkl` is required. The compact checkpoint contains no optimizer, teacher, or training history and cannot resume training. Its implementation configuration is `LightningDiT-XL/1`, `dmf_depth=10`, with 32-channel 16×16 VA-VAE latents and 1,000 classes. The student's learned guidance embedding remains fixed at `gw=2.0`; this does not add a second network evaluation.
+The statistics for this model are `.pt` and `.npz` files; no statistics `.pkl` is required. The compact checkpoint contains no optimizer, teacher, or training history and cannot resume training. Its implementation configuration is `LightningDiT-XL/1`, `dmf_depth=10`, with 32-channel 16×16 VA-VAE latents and 1,000 classes.
 
 The required **VA-VAE decoder** is downloaded separately by the code's asset script. The two statistics files are unmodified copies from [Peterande/FACM at revision b7b5fee6d3f43347883d0fe6f9f357e3068375b8](https://huggingface.co/Peterande/FACM/tree/b7b5fee6d3f43347883d0fe6f9f357e3068375b8). [assets.json](assets.json) records their upstream revision, sizes, and SHA-256 hashes, together with the decoder and training-only teacher identities.
 
@@ -71,7 +71,7 @@ Repeat with `--sampling-steps 2` and `4`, using a fresh output directory for eac
 | 4 | **1.3711** |
 | 8 | 1.2764 |
 
-[results.json](results.json) records the full seven-setting evaluation protocol and results. A separate public-code evaluation on October 11, 2026 verified the exact released checkpoint at 1 NFE: **FID 2.03575822379446**, matching the original **2.035758223794346** to numerical precision. The multi-step results are the archived evaluations of the same tensor-identical reporting weights. Hardware and dependency changes can affect exact numerical reproduction.
+[results.json](results.json) contains the evaluation protocol and results.
 
 ## Integrity and provenance
 
@@ -85,7 +85,7 @@ See [consistentdmd-ldit-best.pt.json](consistentdmd-ldit-best.pt.json) for model
 
 ## License and acknowledgements
 
-The CDMD checkpoint follows the project's [CC BY-NC-SA 4.0 license](LICENSE.md). The architecture and decoder build on FACM, Lightning-DiT, and DiT. Upstream assets retain their respective terms; see [THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md).
+The CDMD checkpoint follows the project's [CC BY-NC-SA 4.0 license](LICENSE.md). The architecture and decoder build on Lightning-DiT and DiT. Upstream assets retain their respective terms; see [THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md).
 
 ## Citation
 

SHA256SUMS

@@ -1,6 +1,6 @@
 17b2442405956fe8457ef0e230cbc18ed2cb6444bc86c04b5d61b75df3518335  LICENSE.md
-3058487f3a486c0fdec6320027ef0f392233d65a42205938f360b9e190390dea  README.md
-3cd9a90896fd644a44b78c9473a87e178ddfc843eef29c8b2b41c7c0e040b64c  THIRD_PARTY_NOTICES.md
+4be20b1e9fe72bd56ed7dd521c0105624b686d822ef0579d4db9c5ec05022a15  README.md
+de8c3743bc493d28aa60ecd12baf53e68c6c7620a0d74773894e627c2253686a  THIRD_PARTY_NOTICES.md
 140953e5e44c62e14c5861359697ed511e999cbd16ade9f059133822ce6f7115  assets.json
 0eddca248af6cfaa21e5b9b32d0c02329923e5066d1bb2b140fa6827c061c1d4  consistentdmd-ldit-best.pt
 478c70dd8cb775476e9378e9b68cc1aac4ad9f447d347081089cc525eed6909c  consistentdmd-ldit-best.pt.json

THIRD_PARTY_NOTICES.md

@@ -5,8 +5,8 @@ The CDMD checkpoint follows the project's CC BY-NC-SA 4.0 license in
 available in [felix-yuxiang/CDMD](https://github.com/felix-yuxiang/CDMD), including
 the [code's third-party notices](https://github.com/felix-yuxiang/CDMD/blob/main/THIRD_PARTY_NOTICES.md).
 
-The student builds on [Lightning-DiT](https://github.com/hustvl/LightningDiT),
-[FACM](https://github.com/ali-vilab/FACM), and [DiT](https://github.com/facebookresearch/DiT).
+The student builds on [Lightning-DiT](https://github.com/hustvl/LightningDiT)
+and [DiT](https://github.com/facebookresearch/DiT).
 The reporting average uses [ema-pytorch](https://github.com/lucidrains/ema-pytorch).
 Evaluation uses the pinned [LTH14 torch-fidelity fork](https://github.com/LTH14/torch-fidelity/tree/bfe72b995473e4fee9351705d7b005f17442a4ca).