Add files using upload-large-folder tool

InternW0-RoboTwin-clean.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b718314f57c2febb159deb03d92cf9ebba830f7c75920c0b6e34c4903b91f5a2
+size 14705453767

LICENSE

@@ -0,0 +1,24 @@
+MIT License
+
+Copyright (c) 2026 The W0 Authors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+Except for third-party code and files that retain their original license
+headers, this repository is released under the MIT License.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

PROVENANCE.md

@@ -0,0 +1,68 @@
+# InternW0 RoboTwin-clean checkpoint provenance
+
+## Selected artifact
+
+`InternW0-RoboTwin-clean.pt` is the model-only checkpoint selected for the
+RoboTwin 2.0 Clean2Random release. It is derived from the historical run
+`w0-cln-cj02-s2-64g-dh-0919r5` at step 10,545.
+
+The historical source implementation was `ahawam`; the public runtime is `w0`.
+The resolved numerical training contract is preserved in
+`source_training_config.yaml` with private filesystem paths replaced by portable
+environment variables.
+
+## Lineage
+
+1. Heterogeneous InternW0 pretraining at global step 30,000.
+2. RoboTwin target-domain initialization with `new_domain` action/proprio
+   interfaces, seed 42.
+3. Clean-demonstration posttraining run
+   `w0-cln-p30k-cj02-s2-64g-0918r3`, resumed at step 5,000.
+4. Continued best Clean2Random recipe
+   `w0-cln-cj02-s2-64g-dh-0919r5`, selected at step 10,545.
+
+The checkpoint embeds `behavior_config`, `training_recipe`,
+`posttrain_statistics`, and `posttrain_provenance`. These fields are also
+exported verbatim to `checkpoint_metadata.json`.
+
+## Training contract
+
+- Clean demonstrations only.
+- RoboTwin native joint-space representation.
+- 25 Hz sampling (`sampling_stride=2` from 50 Hz).
+- Every valid native window.
+- ColorJitter: brightness 0.2, contrast 0.2, saturation 0.2, hue 0.
+- Offset 0.
+- Batch size 12 per process.
+- 15 epochs.
+- Learning rate 3e-5, cosine schedule.
+- Seed 42.
+- 37-dimensional canonical action/proprio interface.
+- 14 valid ALOHA joint/gripper slots:
+  `[0, 1, 2, 3, 4, 5, 16, 17, 18, 19, 20, 21, 22, 33]`.
+- Action horizon 64, chunk size 16, video plan horizon 96, video/action stride 8.
+- K=8 target soft prompts.
+- Native absolute joint commands at the interval end.
+
+The matching normalization file is `normalization.json`.
+
+## Evaluation reference
+
+The source resolved evaluation reference is preserved in
+`source_evaluation_config.yaml`. It uses the fixed-period RoboTwin adapter at
+25 Hz, 10 inference steps, a 64-action horizon, four action chunks per video
+prefill, and 15 episodes per task.
+
+The paper-reported RoboTwin 2.0 Clean2Random results are:
+
+| Split | Success rate |
+|---|---:|
+| Clean | 83.20% |
+| Randomized (unseen during training) | 68.00% |
+| Average | 75.60% |
+
+## Excluded from this repository
+
+The checkpoint package does not include optimizer state, distributed training
+state, logs, RoboTwin data, Wan2.2 files, DINOv3 weights, text caches, or
+simulation checkouts.

README.md

@@ -0,0 +1,232 @@
+---
+license: mit
+library_name: pytorch
+pipeline_tag: robotics
+base_model: InternRobotics/InternW0-Base
+tags:
+- robotics
+- world-model
+- robot-learning
+- manipulation
+- flow-matching
+- video-action-model
+- robotwin
+- InternW0
+language:
+- en
+---
+
+# InternW0-RoboTwin-clean
+
+[![Paper](https://img.shields.io/badge/Paper-arXiv%202609.27656-b31b1b)](https://arxiv.org/abs/2609.27656)
+[![Project](https://img.shields.io/badge/Project-InternW0-2f80ed)](https://internrobotics.github.io/internw0)
+[![Code](https://img.shields.io/badge/Code-GitHub-181717?logo=github)](https://github.com/InternRobotics/InternW0)
+[![Base model](https://img.shields.io/badge/Base-InternW0--Base-6b7280)](https://huggingface.co/InternRobotics/InternW0-Base)
+[![Downloads last month](https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fhuggingface.co%2Fapi%2Fmodels%2FInternRobotics%2FInternW0-RoboTwin-clean&query=%24.downloads&label=downloads%20last%20month&logo=huggingface)](https://huggingface.co/InternRobotics/InternW0-RoboTwin-clean)
+
+## Model description
+
+**InternW0-RoboTwin-clean** is the task-posttrained InternW0 checkpoint for the
+RoboTwin 2.0 **Clean2Random** setting. It is initialized from the heterogeneous
+[InternW0-Base](https://huggingface.co/InternRobotics/InternW0-Base) pretraining
+checkpoint and posttrained only on clean demonstrations. The selected checkpoint
+is then evaluated on clean scenes and unseen randomized scenes.
+
+InternW0 couples a slow Wan2.2-TI2V-5B VideoDiT planner with a fast ActionDiT
+executor. Observation-conditioned chunk K/V routing adapts cached future-video
+context to each newly observed state, allowing the model to reuse long-horizon
+predictions while generating responsive action chunks.
+
+This is a **native `w0` checkpoint**, not a standalone Transformers model. It is
+intended for RoboTwin evaluation or compatible continued posttraining.
+
+## Files and integrity
+
+| File | Purpose | Size | SHA-256 |
+|---|---|---:|---|
+| `InternW0-RoboTwin-clean.pt` | Canonical model-only native checkpoint | 14,705,453,767 bytes | `b718314f57c2febb159deb03d92cf9ebba830f7c75920c0b6e34c4903b91f5a2` |
+| `pytorch_model.bin` | Zero-copy LFS alias used for PyTorch download statistics | Same LFS object | `b718314f57c2febb159deb03d92cf9ebba830f7c75920c0b6e34c4903b91f5a2` |
+| `config.json` | Machine-readable model/download-counting manifest | 1,702 bytes | `82d8a7a6f25fe14720dc7d6e02136fa02fff0f0b2650a29866eacc4edf275347` |
+| `checkpoint_metadata.json` | Embedded behavior/training/statistics/provenance export | 16,176 bytes | `6e4b67d29ca9a369c870f510e9304ffb3e4aec974f6e38a09c35d598c4fbaf54` |
+| `normalization.json` | Matching RoboTwin action/proprio statistics | 4,014 bytes | `31362a24460803c5080ef925b807acbc7af3e08d847c9db222bc2a33c0d96653` |
+| `posttrain_config.yaml` | Portable public-code posttraining config | 1,661 bytes | `109ca7e501c65e4072060ec7477da2a5486f45aed095ece661714ef92f477f88` |
+| `evaluation_config.yaml` | Portable public-code evaluation config | 885 bytes | `89545dbed7a002c3a1e09edf48a3540d30d977e96b89e6fad64ab0b92020fd38` |
+| `source_training_config.yaml` | Historical resolved config with portable paths | 6,175 bytes | `6455e1dda379528fe76d5c85e4161bce651286897567bd55ebe2c45c7e9e4312` |
+| `source_evaluation_config.yaml` | Historical evaluation reference with portable paths | 7,312 bytes | `6b16dc8583a79d69a253f1b3a367aab072c8dcf24737ece784c9a926d807aff3` |
+| `paper_results.json` | Paper-reported Clean2Random result record | 621 bytes | `9396065aded7413f6d4ce0228ac512bb3c97fa0b32157fb8b15f19c1e82ef2cf` |
+| `PROVENANCE.md` | Selection, lineage, training, and evaluation provenance | 2,446 bytes | `0cca6d854f5bf6ee3dccd84128584bdae398f9f02e053c016d9d097bd944680a` |
+| `LICENSE` | MIT license | 1,205 bytes | `b1d3d0160bdd9748bfc4e8acb9ab8fbce7ab725c827fd6fc673261f57aae0306` |
+
+`pytorch_model.bin` is a second Git-LFS path to the exact same object as
+`InternW0-RoboTwin-clean.pt`. It does not duplicate the 14.7 GB payload.
+
+## Checkpoint summary
+
+- Selected step: `10545`
+- Serialization dtype: `bfloat16`
+- Model-only checkpoint: yes
+- Optimizer state: excluded
+- Tensors: 1,977
+- Parameter elements: 7,352,528,643
+- Source heterogeneous pretraining step: `30000`
+- Target domain: `robotwin_wam/aloha`
+- Target soft prompts: K=8
+- Canonical action/proprio dimensionality: 37
+- Valid ALOHA joint/gripper slots: 14
+- Action horizon: 64
+- Action chunk size: 16
+- Video plan horizon: 96 action ticks
+- Video/action stride: 8
+- History frames: 0
+- Action semantics: native absolute joint command at interval end
+
+The checkpoint embeds a schema-v4 `behavior_config`, `training_recipe`,
+`posttrain_statistics`, and `posttrain_provenance`. The public loader performs a
+semantic contract check before loading weights.
+
+## RoboTwin Clean2Random training
+
+The retained recipe uses:
+
+- clean demonstrations only;
+- native RoboTwin joint-space data;
+- 25 Hz sampling (`sampling_stride=2` from 50 Hz);
+- every valid native window;
+- ColorJitter with brightness 0.2, contrast 0.2, saturation 0.2, and hue 0;
+- action offset 0;
+- batch size 12 per process;
+- 15 epochs;
+- learning rate 3e-5 with cosine decay;
+- seed 42.
+
+The matching normalization statistics are in `normalization.json`. The
+historical run resumed from step 5,000 of the preceding clean run and selected
+step 10,545 from the continued best Clean2Random run. See `PROVENANCE.md` for
+the full lineage.
+
+## Evaluation result
+
+The paper-reported RoboTwin 2.0 Clean2Random results are:
+
+| Split | Success rate |
+|---|---:|
+| Clean | **83.20%** |
+| Randomized, unseen during training | **68.00%** |
+| Average | **75.60%** |
+
+The source evaluation reference uses 50 tasks, 15 episodes per task, 10
+inference steps, a 64-action horizon, four action chunks per video prefill, and
+a fixed-period 25 Hz controller. The historical resolved protocol is preserved
+in `source_evaluation_config.yaml`.
+
+## Required external assets
+
+This checkpoint is not self-contained. Users must separately obtain and comply
+with the licenses for:
+
+```bash
+hf download Wan-AI/Wan2.2-TI2V-5B --local-dir assets/wan22
+hf download facebook/dinov3-vitb16-pretrain-lvd1689m \
+  --local-dir assets/dinov3-vitb16-pretrain-lvd1689m
+```
+
+A RoboTwin checkout, the matching RoboTwin data export, and its UMT5 text cache
+are also required. The repository does not bundle datasets, simulation code,
+Wan2.2 files, or DINOv3 weights.
+
+## Quick start
+
+### 1. Install the public codebase
+
+```bash
+git clone https://github.com/InternRobotics/InternW0.git
+cd InternW0
+
+conda create -n w0 python=3.10 -y
+conda activate w0
+pip install -U pip setuptools wheel
+pip install torch==2.7.1 torchvision==0.22.1 \
+  --extra-index-url https://download.pytorch.org/whl/cu128
+pip install -e .
+```
+
+### 2. Download this checkpoint
+
+```bash
+hf download InternRobotics/InternW0-RoboTwin-clean \
+  config.json pytorch_model.bin checkpoint_metadata.json \
+  normalization.json posttrain_config.yaml evaluation_config.yaml \
+  source_training_config.yaml source_evaluation_config.yaml \
+  paper_results.json PROVENANCE.md LICENSE \
+  --local-dir assets/robotwin
+
+mv assets/robotwin/pytorch_model.bin \
+   assets/robotwin/InternW0-RoboTwin-clean.pt
+```
+
+### 3. Run RoboTwin evaluation
+
+```bash
+export W0_WAN_MODEL_DIR="$PWD/assets/wan22"
+export W0_DINO_WEIGHTS="$PWD/assets/dinov3-vitb16-pretrain-lvd1689m"
+export W0_ROBOTWIN_DATA="$PWD/data/robotwin"
+export W0_ROBOTWIN_STATS="$PWD/assets/robotwin/normalization.json"
+export W0_ROBOTWIN_CHECKPOINT="$PWD/assets/robotwin/InternW0-RoboTwin-clean.pt"
+export ROBOTWIN_ROOT="$PWD/external/RoboTwin"
+
+python experiments/robotwin/eval_robotwin_single.py \
+  EVALUATION.robotwin_root="$ROBOTWIN_ROOT" \
+  EVALUATION.task_name=<task> \
+  EVALUATION.ckpt="$W0_ROBOTWIN_CHECKPOINT"
+```
+
+For compatible continued posttraining:
+
+```bash
+bash scripts/train_zero1.sh 8 --config-name posttrain_robotwin
+```
+
+## Download statistics
+
+Hugging Face uses library-specific query files for model download statistics.
+For the `pytorch` library, the Hub counts HTTP `GET` and `HEAD` requests to
+`pytorch_model.bin`. This repository provides that path as a zero-copy alias of
+the canonical checkpoint. The badge above reads the public Hub API `downloads`
+field, corresponding to downloads in the last 30 days. Hub analytics are
+aggregated and may lag briefly.
+
+## Intended uses
+
+- RoboTwin 2.0 Clean2Random reproduction and extension.
+- Research on world-action models, context routing, and cross-embodiment robot
+  learning.
+- Compatible continued posttraining with the public `w0` codebase.
+
+## Limitations
+
+- This is a simulation-domain research checkpoint, not a deployed robot
+  controller.
+- It requires the public InternW0 codebase, external Wan2.2 and DINOv3 assets,
+  RoboTwin data, text caches, and substantial compute.
+- It is not directly compatible with `transformers.AutoModel`.
+- Real-robot use requires independent safety analysis, constrained workspaces,
+  and human oversight.
+
+## License
+
+The weights and metadata in this repository are released under the
+[MIT License](./LICENSE). Third-party assets obtained separately remain subject
+to their original licenses.
+
+## Citation
+
+If you use InternW0, please cite:
+
+```bibtex
+@article{internw0,
+  title = {InternW0: A Foundational Physical World Model for Efficient Real-World Interactions},
+  author = {Jisong Cai and Yao Mu and Ganlin Yang and Zhe Cao and Zhangzheng Tu and Xing Gao and Kailin Li and Xinyu Zhan and Lixin Yang and Yangkun Zhu and Haoxiang Ma and Ming Zhou and Qiaojun Yu and Yufei Xue and Liqun He and Yifei Yao and Yifan Zhu and Long Ling and Bingqi Jiang and Haoyu Guo and Xueyue Zhu and Bowen Zhou and Bin Zhao and Tianfan Xue and Chunhua Shen and Weinan Zhang},
+  journal = {arXiv preprint arXiv:2609.27656},
+  year = {2026}
+}
+```

checkpoint_metadata.json

@@ -0,0 +1,616 @@
+{
+  "file": "InternW0-RoboTwin-clean.pt",
+  "source_checkpoint": "/data/shared/wamrsi/jisong/w0/release/robotwin/c2r/step_10545/step_010545.pt",
+  "size_bytes": 14705453767,
+  "sha256": "b718314f57c2febb159deb03d92cf9ebba830f7c75920c0b6e34c4903b91f5a2",
+  "model_only": true,
+  "optimizer_state_included": false,
+  "step": 10545,
+  "torch_dtype": "torch.bfloat16",
+  "tensor_count": 1977,
+  "parameter_elements": 7352528643,
+  "tensor_bytes": 14705057286,
+  "behavior_config": {
+    "schema_version": 4,
+    "num_soft_prompts": 8,
+    "domain_configs": [
+      {
+        "name": "robotwin_wam/aloha",
+        "dataset": "robotwin_wam",
+        "embodiment": "aloha",
+        "action_dim": 37,
+        "proprio_dim": 37,
+        "action_valid_indices": [
+          0,
+          1,
+          2,
+          3,
+          4,
+          5,
+          16,
+          17,
+          18,
+          19,
+          20,
+          21,
+          22,
+          33
+        ],
+        "proprio_valid_indices": [
+          0,
+          1,
+          2,
+          3,
+          4,
+          5,
+          16,
+          17,
+          18,
+          19,
+          20,
+          21,
+          22,
+          33
+        ],
+        "raw_joint_names": [
+          "left_waist",
+          "left_shoulder",
+          "left_elbow",
+          "left_forearm_roll",
+          "left_wrist_angle",
+          "left_wrist_rotate",
+          "left_gripper",
+          "right_waist",
+          "right_shoulder",
+          "right_elbow",
+          "right_forearm_roll",
+          "right_wrist_angle",
+          "right_wrist_rotate",
+          "right_gripper"
+        ],
+        "action_semantics": "native_absolute_joint_command_interval_end"
+      }
+    ],
+    "soft_prompt_domains": [
+      "robotwin_wam/aloha"
+    ],
+    "soft_prompt_layout": "per_chunk_prefix",
+    "soft_prompt_position_mode": "identity_rope",
+    "soft_prompt_timestep_mode": "follow_chunk",
+    "observation_encoder": "dino",
+    "dino_encoder": {
+      "model_name": "dinov3-vitb16-hf",
+      "weights_sha256": "1bde6e2cbdb2973afa32f8b000da32c7686c5be6d8e82399ac581f9bb35a1e50",
+      "input_resolution": [
+        384,
+        320
+      ],
+      "feature_dim": 768,
+      "patch_size": 16,
+      "use_cls_token": false,
+      "normalize_features": false,
+      "preprocessing": "uint8_or_minus_one_one_to_imagenet_float32_no_resize",
+      "weights_hash_format": "hf_config_and_safetensors_v1"
+    },
+    "architecture": "ahawam_one_way_ovcr",
+    "action_video_read_mode": "future",
+    "video_cache_mode": "generated_future_last_denoise",
+    "action_position_unit": "action_step",
+    "video_rope_position_unit": "latent_frame",
+    "action_horizon": 64,
+    "action_chunk_size": 16,
+    "num_history_frames": 0,
+    "video_rope_frame_stride": null,
+    "video_attention_mask_mode": "first_frame_causal",
+    "video_action_conditioned": false,
+    "autoregressive_teacher_forcing": true,
+    "action_teacher_forcing_mask_mode": "stage1_chunkwise",
+    "action_obs_downsample_factor": 1,
+    "video_plan_horizon": 96,
+    "video_action_stride": 8,
+    "vae_temporal_downsample_factor": 4
+  },
+  "training_recipe": {
+    "version": 2,
+    "future_training_mode": "idm",
+    "max_action_offset": 0,
+    "data_offset_config": {
+      "train": {
+        "max_action_offset": 0,
+        "action_horizon": 64,
+        "action_chunk_size": 16,
+        "action_video_freq_ratio": 8,
+        "action_offset_enabled": false,
+        "domains": {
+          "version": 1,
+          "source_fps": 50,
+          "target_fps": 25.0,
+          "stride": 2,
+          "phases": [
+            0,
+            1
+          ],
+          "action_horizon": 64,
+          "action_chunk_size": 16,
+          "max_action_offset": 0,
+          "video_plan_horizon": 96,
+          "num_history_frames": 0,
+          "short_episodes": "pad_mask",
+          "slots": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            16,
+            17,
+            18,
+            19,
+            20,
+            21,
+            22,
+            33
+          ],
+          "domain": {
+            "name": "robotwin_wam/aloha",
+            "dataset": "robotwin_wam",
+            "embodiment": "aloha",
+            "action_dim": 37,
+            "proprio_dim": 37,
+            "action_valid_indices": [
+              0,
+              1,
+              2,
+              3,
+              4,
+              5,
+              16,
+              17,
+              18,
+              19,
+              20,
+              21,
+              22,
+              33
+            ],
+            "proprio_valid_indices": [
+              0,
+              1,
+              2,
+              3,
+              4,
+              5,
+              16,
+              17,
+              18,
+              19,
+              20,
+              21,
+              22,
+              33
+            ],
+            "raw_joint_names": [
+              "left_waist",
+              "left_shoulder",
+              "left_elbow",
+              "left_forearm_roll",
+              "left_wrist_angle",
+              "left_wrist_rotate",
+              "left_gripper",
+              "right_waist",
+              "right_shoulder",
+              "right_elbow",
+              "right_forearm_roll",
+              "right_wrist_angle",
+              "right_wrist_rotate",
+              "right_gripper"
+            ],
+            "action_semantics": "native_absolute_joint_command_interval_end"
+          },
+          "normalization": "bounds_q99_minmax_degenerate_clip5",
+          "sigma": 1e-08,
+          "camera_layout": "cover_center_crop_main_top_wrists_bottom_384x320",
+          "text": "episode_first_instruction_default_prompt_128_wan22ti2v5b",
+          "metadata_sha256": "b79ffeee79313dabad593d571cddb0bcf9acd4ee88da2030a39b487afb60e5d8",
+          "split_seed": 42,
+          "val_set_proportion": 0.01,
+          "stats_sha256": "31362a24460803c5080ef925b807acbc7af3e08d847c9db222bc2a33c0d96653",
+          "anchor_sampling": "all_valid_native_anchors_two_phases_v1",
+          "samples_per_epoch": 539287,
+          "color_jitter": {
+            "enabled": true,
+            "brightness": 0.2,
+            "contrast": 0.2,
+            "saturation": 0.2,
+            "hue": 0.0,
+            "version": 1,
+            "implementation": "torchvision_float32_rgb_v1",
+            "order": [
+              "brightness",
+              "contrast",
+              "saturation",
+              "hue"
+            ],
+            "sharing": "sample_all_frames_and_camera_slots",
+            "rng": "numpy_seed_epoch_index_cjit_v1",
+            "exclusion": "explicit_missing_camera_mask"
+          }
+        }
+      },
+      "val": {
+        "max_action_offset": 0,
+        "action_horizon": 64,
+        "action_chunk_size": 16,
+        "action_video_freq_ratio": 8,
+        "action_offset_enabled": false,
+        "domains": {
+          "version": 1,
+          "source_fps": 50,
+          "target_fps": 25.0,
+          "stride": 2,
+          "phases": [
+            0,
+            1
+          ],
+          "action_horizon": 64,
+          "action_chunk_size": 16,
+          "max_action_offset": 0,
+          "video_plan_horizon": 96,
+          "num_history_frames": 0,
+          "short_episodes": "pad_mask",
+          "slots": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            16,
+            17,
+            18,
+            19,
+            20,
+            21,
+            22,
+            33
+          ],
+          "domain": {
+            "name": "robotwin_wam/aloha",
+            "dataset": "robotwin_wam",
+            "embodiment": "aloha",
+            "action_dim": 37,
+            "proprio_dim": 37,
+            "action_valid_indices": [
+              0,
+              1,
+              2,
+              3,
+              4,
+              5,
+              16,
+              17,
+              18,
+              19,
+              20,
+              21,
+              22,
+              33
+            ],
+            "proprio_valid_indices": [
+              0,
+              1,
+              2,
+              3,
+              4,
+              5,
+              16,
+              17,
+              18,
+              19,
+              20,
+              21,
+              22,
+              33
+            ],
+            "raw_joint_names": [
+              "left_waist",
+              "left_shoulder",
+              "left_elbow",
+              "left_forearm_roll",
+              "left_wrist_angle",
+              "left_wrist_rotate",
+              "left_gripper",
+              "right_waist",
+              "right_shoulder",
+              "right_elbow",
+              "right_forearm_roll",
+              "right_wrist_angle",
+              "right_wrist_rotate",
+              "right_gripper"
+            ],
+            "action_semantics": "native_absolute_joint_command_interval_end"
+          },
+          "normalization": "bounds_q99_minmax_degenerate_clip5",
+          "sigma": 1e-08,
+          "camera_layout": "cover_center_crop_main_top_wrists_bottom_384x320",
+          "text": "episode_first_instruction_default_prompt_128_wan22ti2v5b",
+          "metadata_sha256": "b79ffeee79313dabad593d571cddb0bcf9acd4ee88da2030a39b487afb60e5d8",
+          "split_seed": 42,
+          "val_set_proportion": 0.01,
+          "stats_sha256": "31362a24460803c5080ef925b807acbc7af3e08d847c9db222bc2a33c0d96653",
+          "anchor_sampling": "prefer_full_horizon_v1",
+          "samples_per_epoch": 128
+        }
+      }
+    },
+    "idm_video_cond_noise_prob": 0.5,
+    "idm_checkpoint_full_layer": true,
+    "idm_checkpoint_layer_stride": 1,
+    "video_scheduler": {
+      "type": "ahawam.models.wan22.schedulers.scheduler_continuous.WanContinuousFlowMatchScheduler",
+      "shift": 5.0,
+      "num_train_timesteps": 1000
+    },
+    "action_scheduler": {
+      "type": "ahawam.models.wan22.schedulers.scheduler_continuous.WanContinuousFlowMatchScheduler",
+      "shift": 5.0,
+      "num_train_timesteps": 1000
+    },
+    "loss_lambda_video": 1.0,
+    "loss_lambda_action_prior": 1.0,
+    "action_timestep_mode": "per_chunk",
+    "soft_prompt_training": {
+      "lr_scale": 1.0,
+      "schedule": "per_group_relative_v1",
+      "init_std": 0.02
+    }
+  },
+  "posttrain_statistics": {
+    "recipe": {
+      "version": 1,
+      "source_fps": 50,
+      "target_fps": 12.5,
+      "stride": 4,
+      "phases": [
+        0,
+        1,
+        2,
+        3
+      ],
+      "action_horizon": 64,
+      "action_chunk_size": 16,
+      "max_action_offset": 15,
+      "video_plan_horizon": 96,
+      "num_history_frames": 0,
+      "short_episodes": "pad_mask",
+      "slots": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        16,
+        17,
+        18,
+        19,
+        20,
+        21,
+        22,
+        33
+      ],
+      "domain": {
+        "name": "robotwin_wam/aloha",
+        "dataset": "robotwin_wam",
+        "embodiment": "aloha",
+        "action_dim": 37,
+        "proprio_dim": 37,
+        "action_valid_indices": [
+          0,
+          1,
+          2,
+          3,
+          4,
+          5,
+          16,
+          17,
+          18,
+          19,
+          20,
+          21,
+          22,
+          33
+        ],
+        "proprio_valid_indices": [
+          0,
+          1,
+          2,
+          3,
+          4,
+          5,
+          16,
+          17,
+          18,
+          19,
+          20,
+          21,
+          22,
+          33
+        ],
+        "raw_joint_names": [
+          "left_waist",
+          "left_shoulder",
+          "left_elbow",
+          "left_forearm_roll",
+          "left_wrist_angle",
+          "left_wrist_rotate",
+          "left_gripper",
+          "right_waist",
+          "right_shoulder",
+          "right_elbow",
+          "right_forearm_roll",
+          "right_wrist_angle",
+          "right_wrist_rotate",
+          "right_gripper"
+        ],
+        "action_semantics": "native_absolute_joint_command_interval_end"
+      },
+      "normalization": "bounds_q99_minmax_degenerate_clip5",
+      "sigma": 1e-08,
+      "camera_layout": "cover_center_crop_main_top_wrists_bottom_384x320",
+      "text": "episode_first_instruction_default_prompt_128_wan22ti2v5b"
+    },
+    "metadata_sha256": "b79ffeee79313dabad593d571cddb0bcf9acd4ee88da2030a39b487afb60e5d8",
+    "split_seed": 42,
+    "val_set_proportion": 0.01,
+    "train_episodes": 2472,
+    "validation_episodes": 28,
+    "position_selection": "union_all_reachable_phase_tick_no_padding",
+    "quantile_method": "exact_numpy_linear",
+    "action": {
+      "q01": [
+        -0.9637558460235596,
+        -5.257390398583084e-07,
+        -2.296771708643064e-05,
+        -1.5930664610862733,
+        -0.44254571497440337,
+        -2.2050026798248292,
+        0.0,
+        -0.1566295325756073,
+        0.0,
+        -2.81171942333458e-05,
+        -1.7002750635147095,
+        -1.0270036816596986,
+        -1.6663912868499755,
+        0.0
+      ],
+      "q99": [
+        0.17338556051254272,
+        2.582751655578613,
+        2.4753708839416504,
+        1.325472354888916,
+        1.2205402231216373,
+        1.4509482383728027,
+        1.0,
+        1.0792511796951278,
+        2.602708811759947,
+        2.4975505542754988,
+        1.3120307922363281,
+        1.0698820352554321,
+        2.101373972892743,
+        1.0
+      ],
+      "minmax_bounds": {},
+      "valid_count": [
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343
+      ]
+    },
+    "proprio": {
+      "q01": [
+        -0.9626896333694458,
+        -5.257390398583084e-07,
+        -2.296771708643064e-05,
+        -1.587372796535492,
+        -0.43772934675216674,
+        -2.146702766418457,
+        0.0,
+        -0.1528666913509369,
+        0.0,
+        -2.81171942333458e-05,
+        -1.6975934505462646,
+        -1.0255959033966064,
+        -1.6527643203735352,
+        0.0
+      ],
+      "q99": [
+        0.17338168621063232,
+        2.581955127716058,
+        2.469698667526245,
+        1.3165214133262622,
+        1.2082784199714538,
+        1.4458667039871216,
+        1.0,
+        1.0600751638412476,
+        2.594983887672424,
+        2.489859433174133,
+        1.3023490905761719,
+        1.0461957573890615,
+        2.079715509414667,
+        1.0
+      ],
+      "minmax_bounds": {},
+      "valid_count": [
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343,
+        534343
+      ]
+    }
+  },
+  "posttrain_provenance": {
+    "source": "/data/shared/caijisong/ckpt/robotwin_posttrain/assets/rlh_step30000",
+    "global_step": 30000,
+    "domain_names": [
+      "interndata/franka_panda",
+      "interndata/franka_robotiq85",
+      "interndata/genie1",
+      "interndata/lift2",
+      "interndata/split_aloha",
+      "interndata/acone",
+      "agibotworld/g1_gripper",
+      "galaxea/r1lite",
+      "ebench/lift2_long_horizon",
+      "robodojo/dual_arx_x5",
+      "robomind/agilex",
+      "robomind/franka",
+      "robomind/franka_fr3_dual",
+      "robomind/ur",
+      "robocoin/agibot_g1_profile5",
+      "robocoin/agibot_g1_profile6",
+      "robocoin/cobot_magic_profile7",
+      "robocoin/galbot_g1",
+      "robocoin/r1lite",
+      "robocoin/rmc_aida_l",
+      "robocoin/split_aloha_profile20",
+      "robocoin/tianqin_a2",
+      "agibotworld/g1_tactile",
+      "ebench/lift2_simple_pnp",
+      "ebench/lift2_teleop_tasks",
+      "robocoin/cobot_magic_profile8",
+      "robocoin/cobot_magic_profile9",
+      "robocoin/cobot_magic_profile10",
+      "robocoin/cobot_magic_profile11",
+      "robocoin/split_aloha_profile21"
+    ],
+    "source_format": "rlh-fastwam-canonical-2",
+    "source_checkpoint": "/nvme/datahub/tmp/caozhe/aha-wam-pretrain/w0-37d-sixsrc-b16-s43-384gpu-bf16-a40-20260917-r5-proprio-sharded/checkpoints/dcp/30000",
+    "strategy": "new_domain",
+    "seed": 42,
+    "proprio_init": "new_domain",
+    "proprio_source_domain": null
+  }
+}

config.json

@@ -0,0 +1,40 @@
+{
+  "model_type": "internw0",
+  "model_name": "InternW0-RoboTwin-clean",
+  "task": "RoboTwin 2.0 Clean2Random posttraining",
+  "base_model": "InternRobotics/InternW0-Base",
+  "architecture": {
+    "name": "ahawam_one_way_ovcr",
+    "public_runtime": "w0",
+    "video_expert": "Wan2.2-TI2V-5B VideoDiT",
+    "action_expert": "ActionDiT",
+    "context_routing": "observation-conditioned chunk K/V editor",
+    "objective": "video/action continuous flow matching"
+  },
+  "weight_file": "InternW0-RoboTwin-clean.pt",
+  "weight_alias_file": "pytorch_model.bin",
+  "weight_format": "native PyTorch checkpoint with embedded InternW0 behavior contract",
+  "weight_size_bytes": 14705453767,
+  "weight_sha256": "b718314f57c2febb159deb03d92cf9ebba830f7c75920c0b6e34c4903b91f5a2",
+  "step": 10545,
+  "torch_dtype": "bfloat16",
+  "runtime": {
+    "code": "https://github.com/InternRobotics/InternW0",
+    "loader": "native w0 checkpoint loader",
+    "transformers_compatible": false
+  },
+  "target_domain": "robotwin_wam/aloha",
+  "source_pretraining_step": 30000,
+  "normalization_file": "normalization.json",
+  "metadata_file": "checkpoint_metadata.json",
+  "posttrain_config_file": "posttrain_config.yaml",
+  "source_training_config_file": "source_training_config.yaml",
+  "source_evaluation_config_file": "source_evaluation_config.yaml",
+  "download_counting": {
+    "library": "pytorch",
+    "query_file": "pytorch_model.bin",
+    "query_file_alias_of": "InternW0-RoboTwin-clean.pt",
+    "policy": "Hugging Face PyTorch model download statistics",
+    "note": "HTTP GET and HEAD requests to pytorch_model.bin are counted server-side. The alias points to the same LFS object as the canonical checkpoint."
+  }
+}

evaluation_config.yaml

@@ -0,0 +1,38 @@
+# @package _global_
+defaults:
+  - posttrain_robotwin
+  - _self_
+
+ckpt: ${oc.env:W0_ROBOTWIN_CHECKPOINT,null}
+gpu_id: 0
+model:
+  load_text_encoder: true
+
+EVALUATION:
+  robotwin_root: ${oc.env:ROBOTWIN_ROOT,external/RoboTwin}
+  policy_name: w0_policy
+  task_name: null
+  task_config: demo_randomized
+  instruction_type: unseen
+  eval_num_episodes: 15
+  output_dir: ${oc.env:W0_OUTPUT_DIR,outputs}/robotwin/evaluation/${now:%Y%m%d_%H%M%S}
+  action_horizon: 64
+  chunks_per_video_prefill: 4
+  num_inference_steps: 10
+  sigma_shift: null
+  text_cfg_scale: 1.0
+  negative_prompt: ""
+  rand_device: cpu
+  tiled: false
+  timing_enabled: false
+  training_config: ${oc.env:W0_ROBOTWIN_TRAINING_CONFIG,null}
+  control_mode: fixed_period
+  target_fps: 25.0
+  step_limit_reference_fps: 12.5
+  dataset_stats_path: null
+  device: cuda
+
+MULTIRUN:
+  enabled: false
+  num_gpus: 8
+  max_tasks_per_gpu: 1

normalization.json

@@ -0,0 +1,205 @@
+{
+  "recipe": {
+    "version": 1,
+    "source_fps": 50,
+    "target_fps": 12.5,
+    "stride": 4,
+    "phases": [
+      0,
+      1,
+      2,
+      3
+    ],
+    "action_horizon": 64,
+    "action_chunk_size": 16,
+    "max_action_offset": 15,
+    "video_plan_horizon": 96,
+    "num_history_frames": 0,
+    "short_episodes": "pad_mask",
+    "slots": [
+      0,
+      1,
+      2,
+      3,
+      4,
+      5,
+      16,
+      17,
+      18,
+      19,
+      20,
+      21,
+      22,
+      33
+    ],
+    "domain": {
+      "name": "robotwin_wam/aloha",
+      "dataset": "robotwin_wam",
+      "embodiment": "aloha",
+      "action_dim": 37,
+      "proprio_dim": 37,
+      "action_valid_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        16,
+        17,
+        18,
+        19,
+        20,
+        21,
+        22,
+        33
+      ],
+      "proprio_valid_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        16,
+        17,
+        18,
+        19,
+        20,
+        21,
+        22,
+        33
+      ],
+      "raw_joint_names": [
+        "left_waist",
+        "left_shoulder",
+        "left_elbow",
+        "left_forearm_roll",
+        "left_wrist_angle",
+        "left_wrist_rotate",
+        "left_gripper",
+        "right_waist",
+        "right_shoulder",
+        "right_elbow",
+        "right_forearm_roll",
+        "right_wrist_angle",
+        "right_wrist_rotate",
+        "right_gripper"
+      ],
+      "action_semantics": "native_absolute_joint_command_interval_end"
+    },
+    "normalization": "bounds_q99_minmax_degenerate_clip5",
+    "sigma": 1e-08,
+    "camera_layout": "cover_center_crop_main_top_wrists_bottom_384x320",
+    "text": "episode_first_instruction_default_prompt_128_wan22ti2v5b"
+  },
+  "metadata_sha256": "b79ffeee79313dabad593d571cddb0bcf9acd4ee88da2030a39b487afb60e5d8",
+  "split_seed": 42,
+  "val_set_proportion": 0.01,
+  "train_episodes": 2472,
+  "validation_episodes": 28,
+  "position_selection": "union_all_reachable_phase_tick_no_padding",
+  "quantile_method": "exact_numpy_linear",
+  "action": {
+    "q01": [
+      -0.9637558460235596,
+      -5.257390398583084e-07,
+      -2.296771708643064e-05,
+      -1.5930664610862733,
+      -0.44254571497440337,
+      -2.2050026798248292,
+      0.0,
+      -0.1566295325756073,
+      0.0,
+      -2.81171942333458e-05,
+      -1.7002750635147095,
+      -1.0270036816596986,
+      -1.6663912868499755,
+      0.0
+    ],
+    "q99": [
+      0.17338556051254272,
+      2.582751655578613,
+      2.4753708839416504,
+      1.325472354888916,
+      1.2205402231216373,
+      1.4509482383728027,
+      1.0,
+      1.0792511796951278,
+      2.602708811759947,
+      2.4975505542754988,
+      1.3120307922363281,
+      1.0698820352554321,
+      2.101373972892743,
+      1.0
+    ],
+    "minmax_bounds": {},
+    "valid_count": [
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343
+    ]
+  },
+  "proprio": {
+    "q01": [
+      -0.9626896333694458,
+      -5.257390398583084e-07,
+      -2.296771708643064e-05,
+      -1.587372796535492,
+      -0.43772934675216674,
+      -2.146702766418457,
+      0.0,
+      -0.1528666913509369,
+      0.0,
+      -2.81171942333458e-05,
+      -1.6975934505462646,
+      -1.0255959033966064,
+      -1.6527643203735352,
+      0.0
+    ],
+    "q99": [
+      0.17338168621063232,
+      2.581955127716058,
+      2.469698667526245,
+      1.3165214133262622,
+      1.2082784199714538,
+      1.4458667039871216,
+      1.0,
+      1.0600751638412476,
+      2.594983887672424,
+      2.489859433174133,
+      1.3023490905761719,
+      1.0461957573890615,
+      2.079715509414667,
+      1.0
+    ],
+    "minmax_bounds": {},
+    "valid_count": [
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343,
+      534343
+    ]
+  }
+}

paper_results.json

@@ -0,0 +1,20 @@
+{
+  "benchmark": "RoboTwin 2.0 Clean2Random",
+  "paper_reported": true,
+  "source": "InternW0 paper, Table 4 (arXiv:2609.27656)",
+  "metrics": {
+    "clean_success_rate_percent": 83.20,
+    "randomized_success_rate_percent": 68.00,
+    "average_success_rate_percent": 75.60
+  },
+  "training_setting": "posttrained only on clean demonstrations; evaluated on clean and unseen randomized scenes",
+  "reference_evaluation": {
+    "tasks": 50,
+    "episodes_per_task": 15,
+    "inference_steps": 10,
+    "action_horizon": 64,
+    "chunks_per_video_prefill": 4,
+    "control_mode": "fixed_period",
+    "target_fps": 25.0
+  }
+}

posttrain_config.yaml

@@ -0,0 +1,52 @@
+# @package _global_
+defaults:
+  - train
+  - override /data: robotwin
+  - override /model: w0
+  - _self_
+
+posttrain_required: true
+posttrain_stats: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+init_checkpoint: ${oc.env:W0_ROBOTWIN_CHECKPOINT,checkpoints/robotwin-clean.pt}
+resume: null
+seed: 42
+model:
+  proprio_dim: 37
+  action_horizon: 64
+  action_chunk_size: 16
+  max_action_offset: 0
+  num_history_frames: 0
+  video_plan_horizon: 96
+  video_action_stride: 8
+  action_dit_config:
+    action_dim: 37
+    num_soft_prompts: 8
+    soft_prompt_domains: [robotwin_wam/aloha]
+    domain_configs:
+      - name: robotwin_wam/aloha
+        dataset: robotwin_wam
+        embodiment: aloha
+        action_dim: 37
+        proprio_dim: 37
+        action_valid_indices: [0, 1, 2, 3, 4, 5, 16, 17, 18, 19, 20, 21, 22, 33]
+        proprio_valid_indices: [0, 1, 2, 3, 4, 5, 16, 17, 18, 19, 20, 21, 22, 33]
+        raw_joint_names: [left_waist, left_shoulder, left_elbow, left_forearm_roll, left_wrist_angle, left_wrist_rotate, left_gripper, right_waist, right_shoulder, right_elbow, right_forearm_roll, right_wrist_angle, right_wrist_rotate, right_gripper]
+        action_semantics: native_absolute_joint_command_interval_end
+
+batch_size: 12
+gradient_accumulation_steps: 1
+learning_rate: 3.0e-5
+soft_prompt_lr_scale: 1.0
+num_epochs: 15
+num_workers: 4
+history_aware_batching: false
+optimizer_group_max_numel: 200000000
+save_every: 2500
+eval_every: 500
+log_every: 10
+keep_training_states: 1
+checkpoint_min_free_gib: 128
+freeze_video_dit: false
+output_dir: ${oc.env:W0_OUTPUT_DIR,outputs}/robotwin/${now:%Y%m%d_%H%M%S}
+wandb:
+  name: internw0_robotwin_clean_c2r

source_evaluation_config.yaml

@@ -0,0 +1,270 @@
+# Historical resolved RoboTwin evaluation reference for step 10545.
+# Private filesystem paths are normalized to portable environment variables;
+# protocol values are unchanged.
+
+output_dir: /data/shared/caijisong/ckpt/robotwin_posttrain/runs/${now:%Y-%m-%d}_${now:%H-%M-%S}
+batch_size: 2
+num_workers: 4
+gc_collect_every: 1
+lr_scheduler_type: cosine
+learning_rate: 1.0e-05
+soft_prompt_lr_scale: 1.0
+num_epochs: 10
+max_steps: null
+log_every: 10
+save_every: 1000
+eval_every: 500
+eval_num_inference_steps: 10
+gradient_accumulation_steps: 1
+mixed_precision: bf16
+seed: 42
+max_grad_norm: 1.0
+weight_decay: 0.0
+freeze_video_dit: false
+resume: null
+init_checkpoint: null
+init_soft_prompts: false
+resume_global_batch_size: null
+wandb:
+  enabled: false
+  workspace: null
+  project: ahawam
+  name: robotwin_posttrain_new_domain_12p5hz
+  group: null
+  mode: online
+data:
+  train:
+    _target_: w0.datasets.robotwin_posttrain.RobotwinPosttrainDataset
+    dataset_root: ${oc.env:W0_ROBOTWIN_DATA,data/robotwin}
+    is_training_set: true
+    samples_per_epoch: null
+    seed: 42
+    val_set_proportion: 0.01
+    pretrained_norm_stats: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+    max_action_offset: 0
+    action_horizon: 64
+    action_chunk_size: 16
+    video_plan_horizon: 96
+    action_video_freq_ratio: 8
+    sampling_stride: 2
+    color_jitter:
+      enabled: true
+      brightness: 0.2
+      contrast: 0.2
+      saturation: 0.2
+      hue: 0.0
+    processor:
+      _target_: w0.datasets.robotwin_posttrain.RobotwinPosttrainProcessor
+      stats_path: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+    epoch_sampling: all_valid_windows
+  val:
+    _target_: w0.datasets.robotwin_posttrain.RobotwinPosttrainDataset
+    dataset_root: ${oc.env:W0_ROBOTWIN_DATA,data/robotwin}
+    is_training_set: false
+    samples_per_epoch: 128
+    seed: 42
+    val_set_proportion: 0.01
+    pretrained_norm_stats: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+    max_action_offset: 0
+    action_horizon: 64
+    action_chunk_size: 16
+    video_plan_horizon: 96
+    action_video_freq_ratio: 8
+    sampling_stride: 2
+    processor:
+      _target_: w0.datasets.robotwin_posttrain.RobotwinPosttrainProcessor
+      stats_path: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+    color_jitter:
+      enabled: true
+      brightness: 0.2
+      contrast: 0.2
+      saturation: 0.2
+      hue: 0.0
+model:
+  _target_: w0.runtime.create_w0
+  model_id: Wan-AI/Wan2.2-TI2V-5B
+  tokenizer_model_id: Wan-AI/Wan2.1-T2V-1.3B
+  tokenizer_max_len: 128
+  load_text_encoder: false
+  proprio_dim: 37
+  redirect_common_files: true
+  mot_checkpoint_mixed_attn: true
+  action_dit_pretrained_path: null
+  skip_dit_load_from_pretrain: true
+  legacy_checkpoint_config: null
+  observation_encoder: dino
+  dino_config:
+    model_name: dinov3-vitb16-hf
+    weights_path: ${oc.env:W0_DINO_WEIGHTS,assets/dinov3-vitb16-pretrain-lvd1689m}
+    input_resolution:
+    - 384
+    - 320
+    feature_dim: 768
+    patch_size: 16
+    use_cls_token: false
+    normalize_features: false
+    encode_microbatch_size: 16
+  checkpoint_observation_encoder_adapt: false
+  action_horizon: 64
+  action_chunk_size: 16
+  max_action_offset: 0
+  action_train_timestep_mode: per_chunk
+  future_training_mode: idm
+  idm_video_cond_noise_prob: 0.5
+  idm_checkpoint_full_layer: true
+  idm_checkpoint_layer_stride: 1
+  action_obs_downsample_factor: 1
+  chunk_kv_editor_num_queries: 32
+  num_history_frames: 0
+  action_video_read_mode: future
+  video_rope_frame_stride: 1
+  video_dit_config:
+    compact_timestep_embedding: true
+    has_image_input: false
+    patch_size:
+    - 1
+    - 2
+    - 2
+    in_dim: 48
+    hidden_dim: 3072
+    ffn_dim: 14336
+    freq_dim: 256
+    text_dim: 4096
+    out_dim: 48
+    num_heads: 24
+    attn_head_dim: 128
+    num_layers: 30
+    eps: 1.0e-06
+    seperated_timestep: true
+    require_clip_embedding: false
+    require_vae_embedding: false
+    fuse_vae_embedding_in_latents: true
+    use_gradient_checkpointing: true
+    video_attention_mask_mode: first_frame_causal
+    action_conditioned: false
+    action_dim: 37
+    action_group_causal_mask_mode: group_diagonal
+  action_dit_config:
+    num_soft_prompts: 8
+    soft_prompt_domains:
+    - robotwin_wam/aloha
+    soft_prompt_init_std: 0.02
+    domain_configs:
+    - name: robotwin_wam/aloha
+      dataset: robotwin_wam
+      embodiment: aloha
+      action_dim: 37
+      proprio_dim: 37
+      action_valid_indices:
+      - 0
+      - 1
+      - 2
+      - 3
+      - 4
+      - 5
+      - 16
+      - 17
+      - 18
+      - 19
+      - 20
+      - 21
+      - 22
+      - 33
+      proprio_valid_indices:
+      - 0
+      - 1
+      - 2
+      - 3
+      - 4
+      - 5
+      - 16
+      - 17
+      - 18
+      - 19
+      - 20
+      - 21
+      - 22
+      - 33
+      raw_joint_names:
+      - left_waist
+      - left_shoulder
+      - left_elbow
+      - left_forearm_roll
+      - left_wrist_angle
+      - left_wrist_rotate
+      - left_gripper
+      - right_waist
+      - right_shoulder
+      - right_elbow
+      - right_forearm_roll
+      - right_wrist_angle
+      - right_wrist_rotate
+      - right_gripper
+      action_semantics: native_absolute_joint_command_interval_end
+    action_dim: 37
+    hidden_dim: 1024
+    ffn_dim: 4096
+    num_heads: 24
+    attn_head_dim: 128
+    num_layers: 30
+    text_dim: 4096
+    freq_dim: 256
+    eps: 1.0e-06
+    autoregressive_teacher_forcing: true
+    action_chunk_size: 16
+    action_teacher_forcing_mask_mode: stage1_chunkwise
+    use_gradient_checkpointing: true
+  video_scheduler:
+    train_shift: 5.0
+    infer_shift: 5.0
+    num_train_timesteps: 1000
+  action_scheduler:
+    train_shift: 5.0
+    infer_shift: 5.0
+    num_train_timesteps: 1000
+  loss:
+    lambda_action: 1.0
+    lambda_action_prior: 1.0
+  video_plan_horizon: 96
+  video_action_stride: 8
+  video_plan_steps: 10
+  domain_registry_path: null
+posttrain_required: true
+posttrain_stats: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+ckpt: ${oc.env:W0_ROBOTWIN_CHECKPOINT,InternW0-RoboTwin-clean.pt}
+gpu_id: 0
+EVALUATION:
+  robotwin_root: ${oc.env:ROBOTWIN_ROOT,external/RoboTwin}
+  policy_name: ahawam_policy
+  task_name: null
+  task_config: demo_randomized
+  instruction_type: unseen
+  eval_num_episodes: 15
+  output_dir: ${oc.env:W0_OUTPUT_DIR,outputs}/robotwin/evaluation/source_step_10545
+  action_horizon: 64
+  chunks_per_video_prefill: 4
+  num_inference_steps: 10
+  sigma_shift: null
+  text_cfg_scale: 1.0
+  negative_prompt: ''
+  rand_device: cpu
+  tiled: false
+  timing_enabled: false
+  training_config: training_config.yaml
+  control_mode: fixed_period
+  target_fps: 25.0
+  step_limit_reference_fps: 12.5
+  dataset_stats_path: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+  prompt_template: 'A video recorded from a robot''s point of view executing the following
+    instruction: {task}'
+  dino_weights_path: ${oc.env:W0_DINO_WEIGHTS,assets/dinov3-vitb16-pretrain-lvd1689m}
+  model_id: Wan-AI/Wan2.2-TI2V-5B
+  tokenizer_model_id: Wan-AI/Wan2.1-T2V-1.3B
+  device: cuda
+  training_config_sha256: 6b8c8129cfea185d9b2938f629b4384829b27d22fb6ba7026c6e1b0fed36529d
+  sampling_stride: 2
+  text_cache_dir: ${oc.env:W0_ROBOTWIN_DATA,data/robotwin}/text_embeds_cache
+MULTIRUN:
+  enabled: false
+  num_gpus: 4
+  max_tasks_per_gpu: 1

source_training_config.yaml

@@ -0,0 +1,237 @@
+# Historical resolved RoboTwin Clean2Random training config.
+# Private filesystem paths are normalized to portable environment variables;
+# numerical and training values are unchanged.
+
+output_dir: ${oc.env:W0_OUTPUT_DIR,outputs}/robotwin/source_0919r5
+batch_size: 12
+num_workers: 4
+gc_collect_every: 1
+lr_scheduler_type: cosine
+learning_rate: 3.0e-05
+soft_prompt_lr_scale: 1.0
+num_epochs: 15
+max_steps: null
+log_every: 10
+save_every: 2500
+keep_training_states: 1
+checkpoint_min_free_gib: 128
+eval_every: 500
+eval_num_inference_steps: 10
+gradient_accumulation_steps: 1
+mixed_precision: bf16
+seed: 42
+max_grad_norm: 1.0
+weight_decay: 0.0
+freeze_video_dit: false
+resume: ${oc.env:W0_ROBOTWIN_SOURCE_RESUME_STATE,checkpoints/source/step_005000}
+init_checkpoint: null
+init_soft_prompts: false
+resume_global_batch_size: null
+wandb:
+  enabled: false
+  workspace: null
+  project: ahawam
+  name: w0-cln-cj02-s2-64g-dh-0919r5
+  group: null
+  mode: online
+data:
+  train:
+    _target_: w0.datasets.robotwin_posttrain.RobotwinPosttrainDataset
+    dataset_root: ${oc.env:W0_ROBOTWIN_DATA,data/robotwin}
+    is_training_set: true
+    samples_per_epoch: null
+    seed: 42
+    val_set_proportion: 0.01
+    pretrained_norm_stats: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+    max_action_offset: 0
+    action_horizon: 64
+    action_chunk_size: 16
+    video_plan_horizon: 96
+    action_video_freq_ratio: 8
+    sampling_stride: 2
+    color_jitter:
+      enabled: true
+      brightness: 0.2
+      contrast: 0.2
+      saturation: 0.2
+      hue: 0.0
+    processor:
+      _target_: w0.datasets.robotwin_posttrain.RobotwinPosttrainProcessor
+      stats_path: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+    epoch_sampling: all_valid_windows
+  val:
+    _target_: w0.datasets.robotwin_posttrain.RobotwinPosttrainDataset
+    dataset_root: ${oc.env:W0_ROBOTWIN_DATA,data/robotwin}
+    is_training_set: false
+    samples_per_epoch: 128
+    seed: 42
+    val_set_proportion: 0.01
+    pretrained_norm_stats: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+    max_action_offset: 0
+    action_horizon: 64
+    action_chunk_size: 16
+    video_plan_horizon: 96
+    action_video_freq_ratio: 8
+    sampling_stride: 2
+    processor:
+      _target_: w0.datasets.robotwin_posttrain.RobotwinPosttrainProcessor
+      stats_path: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+    color_jitter:
+      enabled: true
+      brightness: 0.2
+      contrast: 0.2
+      saturation: 0.2
+      hue: 0.0
+model:
+  _target_: w0.runtime.create_w0
+  model_id: ${oc.env:W0_WAN_MODEL_DIR,assets/wan22}
+  tokenizer_model_id: Wan-AI/Wan2.1-T2V-1.3B
+  tokenizer_max_len: 128
+  load_text_encoder: false
+  proprio_dim: 37
+  redirect_common_files: true
+  mot_checkpoint_mixed_attn: true
+  action_dit_pretrained_path: null
+  skip_dit_load_from_pretrain: true
+  legacy_checkpoint_config: null
+  observation_encoder: dino
+  dino_config:
+    model_name: dinov3-vitb16-hf
+    weights_path: ${oc.env:W0_DINO_WEIGHTS,assets/dinov3-vitb16-pretrain-lvd1689m}
+    input_resolution:
+    - 384
+    - 320
+    feature_dim: 768
+    patch_size: 16
+    use_cls_token: false
+    normalize_features: false
+    encode_microbatch_size: 16
+  checkpoint_observation_encoder_adapt: false
+  action_horizon: 64
+  action_chunk_size: 16
+  max_action_offset: 0
+  action_train_timestep_mode: per_chunk
+  future_training_mode: idm
+  idm_video_cond_noise_prob: 0.5
+  idm_checkpoint_full_layer: true
+  idm_checkpoint_layer_stride: 1
+  action_obs_downsample_factor: 1
+  chunk_kv_editor_num_queries: 32
+  num_history_frames: 0
+  action_video_read_mode: future
+  video_rope_frame_stride: 1
+  video_dit_config:
+    compact_timestep_embedding: true
+    has_image_input: false
+    patch_size:
+    - 1
+    - 2
+    - 2
+    in_dim: 48
+    hidden_dim: 3072
+    ffn_dim: 14336
+    freq_dim: 256
+    text_dim: 4096
+    out_dim: 48
+    num_heads: 24
+    attn_head_dim: 128
+    num_layers: 30
+    eps: 1.0e-06
+    seperated_timestep: true
+    require_clip_embedding: false
+    require_vae_embedding: false
+    fuse_vae_embedding_in_latents: true
+    use_gradient_checkpointing: true
+    video_attention_mask_mode: first_frame_causal
+    action_conditioned: false
+    action_dim: 37
+    action_group_causal_mask_mode: group_diagonal
+  action_dit_config:
+    num_soft_prompts: 8
+    soft_prompt_domains:
+    - robotwin_wam/aloha
+    soft_prompt_init_std: 0.02
+    domain_configs:
+    - name: robotwin_wam/aloha
+      dataset: robotwin_wam
+      embodiment: aloha
+      action_dim: 37
+      proprio_dim: 37
+      action_valid_indices:
+      - 0
+      - 1
+      - 2
+      - 3
+      - 4
+      - 5
+      - 16
+      - 17
+      - 18
+      - 19
+      - 20
+      - 21
+      - 22
+      - 33
+      proprio_valid_indices:
+      - 0
+      - 1
+      - 2
+      - 3
+      - 4
+      - 5
+      - 16
+      - 17
+      - 18
+      - 19
+      - 20
+      - 21
+      - 22
+      - 33
+      raw_joint_names:
+      - left_waist
+      - left_shoulder
+      - left_elbow
+      - left_forearm_roll
+      - left_wrist_angle
+      - left_wrist_rotate
+      - left_gripper
+      - right_waist
+      - right_shoulder
+      - right_elbow
+      - right_forearm_roll
+      - right_wrist_angle
+      - right_wrist_rotate
+      - right_gripper
+      action_semantics: native_absolute_joint_command_interval_end
+    action_dim: 37
+    hidden_dim: 1024
+    ffn_dim: 4096
+    num_heads: 24
+    attn_head_dim: 128
+    num_layers: 30
+    text_dim: 4096
+    freq_dim: 256
+    eps: 1.0e-06
+    autoregressive_teacher_forcing: true
+    action_chunk_size: 16
+    action_teacher_forcing_mask_mode: stage1_chunkwise
+    use_gradient_checkpointing: true
+  video_scheduler:
+    train_shift: 5.0
+    infer_shift: 5.0
+    num_train_timesteps: 1000
+  action_scheduler:
+    train_shift: 5.0
+    infer_shift: 5.0
+    num_train_timesteps: 1000
+  loss:
+    lambda_action: 1.0
+    lambda_action_prior: 1.0
+  video_plan_horizon: 96
+  video_action_stride: 8
+  video_plan_steps: 10
+  domain_registry_path: null
+posttrain_required: true
+posttrain_stats: ${oc.env:W0_ROBOTWIN_STATS,assets/robotwin/normalization.json}
+history_aware_batching: false
+optimizer_group_max_numel: 200000000