Update dataset card (README)

README.md

@@ -8,7 +8,8 @@ tags:
 
 # BridgeVLA++
 
-Pre-training data and checkpoints for **BridgeVLA++** and its predecessor **BridgeVLA**.
+Pre-training data, checkpoints and benchmark keyframe data for **BridgeVLA++** and its
+predecessor **BridgeVLA**.
 
 BridgeVLA++ is a 3D vision-language-action framework that preserves the input-output
 alignment of a pre-trained VLM during 3D action learning — point clouds are projected
@@ -19,60 +20,91 @@ original benchmarks without sacrificing data efficiency or generalization, reach
 state of the art on two memory-dependent benchmarks, and extends to bimanual
 manipulation and a new real-world embodiment.
 
-BridgeVLA++ 是一个 3D 视觉-语言-动作框架:沿用 BridgeVLA「把点云投影成多视角图像、
-先预测热力图再生成动作」的输入输出对齐设计,并引入统一的**时空记忆**,同时建模持久的
-空间上下文与时序交互历史。它在原有基准上追平或超过 BridgeVLA(数据效率与泛化性不降),
-在两个记忆依赖基准上达到 SOTA,并可扩展到双臂操作与新的真机本体。
-
-## Contents 内容
+## Contents
 
 ```
 checkpoints/
-├── pretrain/              # grounding pre-training weights, shared by both models
-│                          #   两代共用的 grounding 预训练权重(finetune 暖启点)
-├── bridgevla/             # BridgeVLA (original) 原版权重
+├── pretrain/              # grounding pre-training weights, shared by both models (finetune warm-start)
+├── bridgevla/             # BridgeVLA (original)
 │   └── rlbench/ colosseum/ gembench/
-└── bridgevla_plus/        # BridgeVLA++ 权重
+└── bridgevla_plus/        # BridgeVLA++ weights
     ├── rlbench/ colosseum/ gembench/ memorybench/
-    └── rmbench/<task>/    # per-task, 9 tasks 每任务一个目录,共 9 个
+    └── rmbench/<task>/    # per-task, 9 tasks
 pretrain_data/
-├── coco.tar.gz            # COCO images 图像
-└── detection_data.json    # RoboPoint grounding annotations 标注
+├── coco.tar.gz            # COCO images
+└── detection_data.json    # RoboPoint grounding annotations
+datasets/                  # benchmark keyframe data (see below)
+├── rlbench/keyframe_cache/size128_v2/<task>/episode<N>.npz + .npz.meta
+├── memorybench/keyframe_cache/size128_v3/<task>/episode<N>.npz + .npz.meta
+└── rmbench/
+    ├── keyframe_data/<task>/keyframe_depth/   # keyframe-only HDF5 training data
+    └── keyframes/<task>.json                  # keyframe metadata -> memory labels
 ```
 
+### Benchmark keyframe data
+
+`datasets/` ships the precomputed keyframe artifacts training depends on — do not
+rearrange them by hand; the code repo's `scripts/download_checkpoints_hf.sh` (or
+`scripts/download_checkpoints_ms.sh` for the identical ModelScope mirror) with a
+target of `rlbench_cache` / `memorybench_cache` / `rmbench_data` (or per-task
+`rmbench_data:<task>`) places each into the exact layout the trainers expect:
+
+- **rlbench / memorybench `keyframe_cache`** — pre-built episode caches
+  (`.npz` decoded observations + `.meta` canonical keyframe indices). The
+  `.meta` files pin RLBench's per-(task, variation) majority-vote canonical
+  keyframes: shipping them makes training runs reproduce ours exactly, and
+  skips a multi-hour local build. RLBench training *requires* this cache.
+- **rmbench `keyframe_data` + `keyframes`** — the actual RMBench training set
+  (keyframe-only re-rendered HDF5, 50 episodes x 10 tasks) plus per-keyframe
+  metadata whose `language_annotation`/`subtask_idx` provide the memory
+  supervision labels. With these, the raw 37 GiB demo_clean demos are NOT
+  needed for training or evaluation. The two directories must stay siblings.
+
+If you download this repo manually instead (e.g. `huggingface-cli download` /
+`modelscope download` of the whole repo), place each tree as below. The local
+names differ from the repo paths **on purpose** (`_keyframe_cache` has a leading
+underscore; capitalization and nesting differ too), so copying `datasets/` into
+the code repo as-is will NOT be found by the trainers:
+
+| in this repo | local path in the code repo |
+|---|---|
+| `checkpoints/` | `data/bridgevla_ckpt/` (inner layout unchanged) |
+| `pretrain_data/` | `data/bridgevla_data/pretrain_data/` |
+| `datasets/rlbench/keyframe_cache/` | `data/bridgevla_data/RLBench/_keyframe_cache/` |
+| `datasets/memorybench/keyframe_cache/` | `data/bridgevla_data/memorybench/data/train/_keyframe_cache/` |
+| `datasets/rmbench/keyframe_data/` | `data/bridgevla_data/RMBench/data/keyframe_data/` |
+| `datasets/rmbench/keyframes/` | `data/bridgevla_data/RMBench/data/keyframes/` |
+
+A misplaced path never degrades silently: training fails fast with an error
+naming the expected location and the download command that fills it.
+
 Every checkpoint directory holds `model_<epoch>.pth` together with `exp_cfg.yaml` and
 `mvt_cfg.yaml`; the two configs define the network architecture and must stay next to
 the weights.
 
-���个权重目录都是 `model_<epoch>.pth` + `exp_cfg.yaml` + `mvt_cfg.yaml`;后两者决定网络
-结构,必须与权重放在一起。
-
 | epoch | RLBench | COLOSSEUM | GemBench | memoryBench | RMBench |
 |---|---|---|---|---|---|
-| BridgeVLA++ | 130 | 200 | 200 | 160 | per-task 每任务不同 |
+| BridgeVLA++ | 130 | 200 | 200 | 160 | per-task |
 | BridgeVLA | 80 | 80 | 40 | – | – |
 
 The `bridgevla/` checkpoints belong to the original BridgeVLA codebase and are **not**
 loadable by BridgeVLA++; run them with that codebase. Only `pretrain/` is shared by both.
 
-`bridgevla/` 下的权重属于原版 BridgeVLA 代码库,**无法**被 BridgeVLA++ 加载,
-请配合原版代码库使用。两代共用的只有 `pretrain/`。
-
-## Papers 论文
+## Papers
 
 - **BridgeVLA++: A Data-Efficient, Generalizable, and Memory-Augmented Vision-Language-Action Framework for 3D Manipulation** — arXiv coming soon.
 - **BridgeVLA: Input-Output Alignment for Efficient 3D Manipulation Learning with Vision-Language Models** — [arXiv:2506.07961](https://arxiv.org/abs/2506.07961)
 
 Usage instructions live in the code repository.
-使用方法见代码仓库。
 
-## Provenance 出处
+## Provenance
 
 `pretrain/`, `bridgevla/` and `pretrain_data/` are the artifacts released with
 BridgeVLA. `pretrain_data/` builds on COCO images and RoboPoint-style grounding
 annotations, which remain subject to their original terms; the Apache-2.0 license above
 applies to the model weights.
 
-`pretrain/`、`bridgevla/` 与 `pretrain_data/` 来自 BridgeVLA 的发布内容。
-`pretrain_data/` 基于 COCO 图像与 RoboPoint 形式的 grounding 标注,仍受其原始条款约束;
-上方的 Apache-2.0 许可适用于模型权重。
+`datasets/` is derived data: the rlbench cache from the PerAct RLBench demos
+(hqfang/rlbench-18-tasks), the memorybench cache from SAM2Act's MemoryBench data
+(hqfang/memorybench), and the rmbench trees re-rendered from RoboTwin 2.0 / RMBench
+(TianxingChen/RMBench). Each remains subject to its upstream benchmark's terms.