Describe the current release without archive or migration history

README.md

@@ -31,23 +31,22 @@ AudioLDM-ControlNet is a dance-conditioned music generation model. It combines:
 3. a ControlNet-style conditioning module built on an AudioLDM backbone.
 
 This repository includes the **full model**, the **w/o MotionBERT** checkpoint,
-and the complete artifact collection previously shared on
-[Google Drive](https://drive.google.com/drive/folders/1rhDqPitcFlthedr5vS6kG2PoNO_PW2it?usp=sharing).
+along with generated audio, evaluation metrics, training logs, and comparison
+videos.
 
 ## Repository Contents
 
-The original Google Drive directory structure is preserved under `models/` and
-`testsets_video_results/`. Only macOS `.DS_Store` metadata files are omitted.
+Model checkpoints and evaluation outputs are organized under `models/` and
+`testsets_video_results/`.
 
 | Path | Contents |
 | --- | --- |
-| [`models/full/`](https://huggingface.co/kmraven/AudioLDM-ControlNet/tree/main/models/full) | Full-model checkpoint, YAML configuration, 36 generated WAV files and `metrics.txt` in `300ksteps/`, and the original W&B training run in `wandb/` |
+| [`models/full/`](https://huggingface.co/kmraven/AudioLDM-ControlNet/tree/main/models/full) | Full-model checkpoint, YAML configuration, 36 generated WAV files and `metrics.txt` in `300ksteps/`, and the W&B training run in `wandb/` |
 | [`models/without_motioinbert/`](https://huggingface.co/kmraven/AudioLDM-ControlNet/tree/main/models/without_motioinbert) | w/o MotionBERT checkpoint, YAML configuration, and 36 generated WAV files and `metrics.txt` in `300ksteps/` |
 | [`testsets_video_results/`](https://huggingface.co/kmraven/AudioLDM-ControlNet/tree/main/testsets_video_results) | 216 MP4 comparison videos: 36 clips for each of the six conditions listed below |
-| `checkpoint-global_step=299999.ckpt` | Original root-level full-model checkpoint, retained for existing download links |
-| `audioldm_original_medium_stretch_pretrained.yaml` | Original root-level full-model configuration, retained for existing download links |
+| `checkpoint-global_step=299999.ckpt` | Full-model checkpoint, also available under `models/full/checkpoints/` |
+| `audioldm_original_medium_stretch_pretrained.yaml` | Full-model configuration, also available under `models/full/` |
 
-The spelling `without_motioinbert` is retained from the original archive.
 The root-level checkpoint and configuration are identical to their counterparts
 under `models/full/`.
 
@@ -61,7 +60,7 @@ configurations are:
 The six video folders are `AudioLDM (default)`, `Ground Truth`, `Li et al.`,
 `Ours`, `Ours (w-o Contrastive Pretraining)`, and `Ours (w-o MotionBERT)`.
 The w/o Contrastive Pretraining variant is represented by comparison videos;
-a checkpoint for that variant was not included in the Google Drive archive.
+a checkpoint for that variant is not included in this release.
 
 ## Download and Use
 
@@ -71,8 +70,7 @@ Install the Hugging Face CLI if needed:
 pip install -U huggingface_hub
 ```
 
-Download the complete archived collection (approximately 16.5 GB), preserving
-its directory structure and avoiding a second local copy of the root checkpoint:
+Download all model and evaluation files (approximately 16.5 GB):
 
 ```bash
 hf download kmraven/AudioLDM-ControlNet \
@@ -92,13 +90,13 @@ hf download kmraven/AudioLDM-ControlNet \
 Run preprocessing and evaluation with the
 [official source code and setup instructions](https://github.com/kmraven/AudioLDM-ControlNet#data-and-checkpoints).
 These are project-specific PyTorch Lightning checkpoints; use the repository's
-evaluation scripts. The archived YAML files preserve the original training
-settings and relative paths. Dataset inputs and separately referenced upstream
-weights must be prepared as described in the source repository. The archive
+evaluation scripts. The YAML files specify the training settings and use
+relative paths. Dataset inputs and separately referenced upstream weights must
+be prepared as described in the source repository. This release
 does not include a standalone Stage 1 `model_best.pth` checkpoint.
 
 Exact training resume, including optimizer state and global step, may be
-incompatible with the refactored modules. If exact resume fails, the source
+incompatible with the current training code. If exact resume fails, the source
 repository recommends commit `0c1f4e3197a7c339586ce5769c6a4510dfbc6f30`.
 
 ## Checkpoint Details