GnanaPraveen commited on
Commit
12bfd47
·
verified ·
1 Parent(s): e5311b5

Rename checkpoint references to Bagel-Robotics-1 and document private evaluation access

Browse files
README.md CHANGED
@@ -7,9 +7,9 @@ tags:
7
  - robodojo
8
  ---
9
 
10
- # AdaPrune — Bagel-Robotics-1 checkpoint 1000
11
 
12
- Prepared public inference release of the **exact checkpoint 1000** used in the Bagel-Robotics-1 evaluation campaign. **Repository visibility is currently private, awaiting a bageldotcom organization administrator change.** Once public, downloads require no account, token or access request. This is distinct from Bagel-Robotics-1-Preview.
13
 
14
  The video expert has **12,468,409,664 parameters**, 23.95% fewer than its recorded parent. Learned masks were converted into smaller FFN and cross-attention layers. Self-attention heads and the action expert are retained. No teacher or learned-mask module is required for inference. Parameter reduction is not a measured inference-speedup claim.
15
 
@@ -24,8 +24,10 @@ The release adds documentation and inference packaging without changing any chec
24
 
25
  ## Download the exact checkpoint
26
 
 
 
27
  ```bash
28
- hf download bageldotcom/adaprune \
29
  --revision 2651631c5d08819088475e8778676f5cf6d9afcf \
30
  --include video.pt action.pt compact_architecture.json dataset_stats.json COMPLETE.json validation.json \
31
  --local-dir /absolute/path/to/checkpoints/cp1000
@@ -35,17 +37,17 @@ There are approximately 29.0 GB of checkpoint tensors. Keep all six files togeth
35
 
36
  ## Inference adapter and reevaluation
37
 
38
- The standalone XPolicyLab adapter is provided under [`inference/`](inference/), at release tag `adaprune-cp1000-inference-v1`. No private research repository is required.
39
 
40
  ```bash
41
- hf download bageldotcom/adaprune \
42
- --revision adaprune-cp1000-inference-v1 \
43
  --include 'inference/**' \
44
- --local-dir /absolute/path/to/adaprune-release
45
 
46
  git clone https://github.com/XPolicyLab/XPolicyLab.git /absolute/path/to/workspace/XPolicyLab
47
  git -C /absolute/path/to/workspace/XPolicyLab checkout 6baccd57ff958570d0db1588370289ee36c8bb70
48
- python /absolute/path/to/adaprune-release/inference/install_adapter.py /absolute/path/to/workspace/XPolicyLab
49
  ```
50
 
51
  Follow the included [policy README](inference/policy/bagel_robotics_1/README.md) to install the Python 3.10 environment, download/verify assets and run both debug checks. Follow [REEVALUATION.md](inference/policy/bagel_robotics_1/REEVALUATION.md) for task, suite and three-seed commands. Install RoboDojo and simulator assets separately using its [official guide](https://robodojo-benchmark.com/doc/usage/install-and-download/).
 
7
  - robodojo
8
  ---
9
 
10
+ # Bagel-Robotics-1 — checkpoint 1000
11
 
12
+ Inference release of the **exact checkpoint 1000** used in the Bagel-Robotics-1 evaluation campaign. **The repository is currently private.** Evaluator access is being arranged using a dedicated fine-grained token with read access only to this repository, delivered through an agreed private channel. Credential provisioning and evaluator assignment remain pending. Public release is prepared for the later verified-leaderboard milestone and requires a bageldotcom administrator. This is distinct from Bagel-Robotics-1-Preview.
13
 
14
  The video expert has **12,468,409,664 parameters**, 23.95% fewer than its recorded parent. Learned masks were converted into smaller FFN and cross-attention layers. Self-attention heads and the action expert are retained. No teacher or learned-mask module is required for inference. Parameter reduction is not a measured inference-speedup claim.
15
 
 
24
 
25
  ## Download the exact checkpoint
26
 
27
+ While the repository is private, supply your authorized evaluator credential securely through `HF_TOKEN` or your Hugging Face login. Do not put credentials in commands, public documents, or application text. The commands below require that access.
28
+
29
  ```bash
30
+ hf download bageldotcom/Bagel-Robotics-1 \
31
  --revision 2651631c5d08819088475e8778676f5cf6d9afcf \
32
  --include video.pt action.pt compact_architecture.json dataset_stats.json COMPLETE.json validation.json \
33
  --local-dir /absolute/path/to/checkpoints/cp1000
 
37
 
38
  ## Inference adapter and reevaluation
39
 
40
+ The standalone XPolicyLab adapter is provided under [`inference/`](inference/), at release tag `bagel-robotics-1-cp1000-inference-v1`. No private research repository is required.
41
 
42
  ```bash
43
+ hf download bageldotcom/Bagel-Robotics-1 \
44
+ --revision bagel-robotics-1-cp1000-inference-v1 \
45
  --include 'inference/**' \
46
+ --local-dir /absolute/path/to/Bagel-Robotics-1-release
47
 
48
  git clone https://github.com/XPolicyLab/XPolicyLab.git /absolute/path/to/workspace/XPolicyLab
49
  git -C /absolute/path/to/workspace/XPolicyLab checkout 6baccd57ff958570d0db1588370289ee36c8bb70
50
+ python /absolute/path/to/Bagel-Robotics-1-release/inference/install_adapter.py /absolute/path/to/workspace/XPolicyLab
51
  ```
52
 
53
  Follow the included [policy README](inference/policy/bagel_robotics_1/README.md) to install the Python 3.10 environment, download/verify assets and run both debug checks. Follow [REEVALUATION.md](inference/policy/bagel_robotics_1/REEVALUATION.md) for task, suite and three-seed commands. Install RoboDojo and simulator assets separately using its [official guide](https://robodojo-benchmark.com/doc/usage/install-and-download/).
RELEASE.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
  "model_name": "Bagel-Robotics-1",
3
- "release_name": "AdaPrune checkpoint 1000",
4
- "repository": "bageldotcom/adaprune",
5
  "checkpoint_revision": "2651631c5d08819088475e8778676f5cf6d9afcf",
6
  "checkpoint_identity": "790bffc9464fa7a026aa43b217aec2033d186895cad643166daa0bcc01914d5c",
7
  "files": {
@@ -30,7 +30,7 @@
30
  "size": 24937214042
31
  }
32
  },
33
- "scope": "Public inference checkpoint and supporting adapter; training/pruning/export source planned November 20, 2026",
34
  "historical_provenance": "Original checkpoint revision and checksums preserved. Older commits retain original provenance and export records; the current release view contains inference materials.",
35
  "official_leaderboard_result": false
36
  }
 
1
  {
2
  "model_name": "Bagel-Robotics-1",
3
+ "release_name": "Bagel-Robotics-1 checkpoint 1000",
4
+ "repository": "bageldotcom/Bagel-Robotics-1",
5
  "checkpoint_revision": "2651631c5d08819088475e8778676f5cf6d9afcf",
6
  "checkpoint_identity": "790bffc9464fa7a026aa43b217aec2033d186895cad643166daa0bcc01914d5c",
7
  "files": {
 
30
  "size": 24937214042
31
  }
32
  },
33
+ "scope": "Private inference checkpoint and supporting adapter for evaluator access; public release prepared; training/pruning/export source planned November 20, 2026",
34
  "historical_provenance": "Original checkpoint revision and checksums preserved. Older commits retain original provenance and export records; the current release view contains inference materials.",
35
  "official_leaderboard_result": false
36
  }
artifact-manifest.json CHANGED
@@ -4,9 +4,9 @@
4
  "checkpoint_identity": "790bffc9464fa7a026aa43b217aec2033d186895cad643166daa0bcc01914d5c",
5
  "update": 1000,
6
  "video_parameters": 12468409664,
7
- "repo_id": "bageldotcom/adaprune",
8
  "revision": "2651631c5d08819088475e8778676f5cf6d9afcf",
9
- "private": false,
10
  "files": {
11
  "COMPLETE.json": {
12
  "sha256": "eee062fbcb8c92a881117aa7f06b5cd039dfb9912d85571b59a20aa30837b0e1",
 
4
  "checkpoint_identity": "790bffc9464fa7a026aa43b217aec2033d186895cad643166daa0bcc01914d5c",
5
  "update": 1000,
6
  "video_parameters": 12468409664,
7
+ "repo_id": "bageldotcom/Bagel-Robotics-1",
8
  "revision": "2651631c5d08819088475e8778676f5cf6d9afcf",
9
+ "private": true,
10
  "files": {
11
  "COMPLETE.json": {
12
  "sha256": "eee062fbcb8c92a881117aa7f06b5cd039dfb9912d85571b59a20aa30837b0e1",
inference/README.md CHANGED
@@ -1,6 +1,6 @@
1
- # Bagel-Robotics-1 public inference adapter
2
 
3
- This folder is the standalone XPolicyLab adapter for AdaPrune checkpoint 1000. It contains inference code and evaluation instructions, with no training repository dependency. Model tensors download separately from `bageldotcom/adaprune`, pinned to `2651631c5d08819088475e8778676f5cf6d9afcf` and verified against `policy/bagel_robotics_1/artifact-manifest.json`.
4
 
5
  Install this adapter into an existing XPolicyLab checkout (the installer refuses to overwrite an existing policy):
6
 
@@ -10,4 +10,6 @@ python install_adapter.py /absolute/path/to/XPolicyLab
10
 
11
  Then follow [Installation](policy/bagel_robotics_1/README.md#installation) for the Python 3.10 environment, exact checkpoint and encoder download, and debug checks. [REEVALUATION.md](policy/bagel_robotics_1/REEVALUATION.md) gives native simulator task/suite/three-seed commands. RoboDojo and its assets are installed separately.
12
 
13
- The adapter is prepared against XPolicyLab `6baccd57ff958570d0db1588370289ee36c8bb70`; source and dependency pins are preserved in the policy folder. This public release is not an XPolicyLab PR or an official leaderboard result. Training/pruning/export source and the technical report are planned for November 20, 2026.
 
 
 
1
+ # Bagel-Robotics-1 inference adapter
2
 
3
+ This folder is the standalone XPolicyLab adapter for Bagel-Robotics-1 checkpoint 1000. It contains inference code and evaluation instructions, with no training repository dependency. Model tensors download separately from `bageldotcom/Bagel-Robotics-1`, pinned to `2651631c5d08819088475e8778676f5cf6d9afcf` and verified against `policy/bagel_robotics_1/artifact-manifest.json`.
4
 
5
  Install this adapter into an existing XPolicyLab checkout (the installer refuses to overwrite an existing policy):
6
 
 
10
 
11
  Then follow [Installation](policy/bagel_robotics_1/README.md#installation) for the Python 3.10 environment, exact checkpoint and encoder download, and debug checks. [REEVALUATION.md](policy/bagel_robotics_1/REEVALUATION.md) gives native simulator task/suite/three-seed commands. RoboDojo and its assets are installed separately.
12
 
13
+ The adapter is prepared against XPolicyLab `6baccd57ff958570d0db1588370289ee36c8bb70`; source and dependency pins are preserved in the policy folder. This inference package is not an XPolicyLab PR or an official leaderboard result. Training/pruning/export source and the technical report are planned for November 20, 2026.
14
+
15
+ The repository is currently private. Evaluators need separately arranged repository-only read access; public release is a later leaderboard milestone.
inference/SHA256SUMS CHANGED
@@ -1,14 +1,14 @@
1
- 9ee395e4c4603776300a8cf2b8e1b986c47df4f2aa00bccaef7d32fe1a302724 README.md
2
  56f26e5b6ba87080c82703e102240fdfbd955c731fb2a044609ffff099795b3f install_adapter.py
3
  6c9437916a0dc5b6ed5fd56b361a4a1894be910e6965c9570e8683cc74a40813 policy/bagel_robotics_1/.gitignore
4
  24cb153ee8bababbfc20fe7acd36768923fbcf2a07d0aa653f460282761ccf8f policy/bagel_robotics_1/EVALUATION_RESULTS.md
5
  8379dddfbbb1154ad9cc57f008f1f8d4a8d26c3de99bd6b36aa4d8067888aa92 policy/bagel_robotics_1/LICENSE
6
- 2b2e02ed1e08922744249a51678dea3b765dd0d10d9d908e93b3b9b53531775a policy/bagel_robotics_1/README.md
7
- 4fa25718ab410d9926541d5c682693a0772f2d7df2f1fd9c1c4a205d1e0b2521 policy/bagel_robotics_1/REEVALUATION.md
8
  2f051521079a9de8815faafcb20098f1fe7fa167425e8c8088e7a42c2d50bb8e policy/bagel_robotics_1/THIRD_PARTY_NOTICES.md
9
  1d22bc8164fcca87f47b003406878112087118c4a5e23e6019ae1998e2b53418 policy/bagel_robotics_1/VALIDATION.md
10
  e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 policy/bagel_robotics_1/__init__.py
11
- 1d89d4f59bf0d9c5d7cdf29de28baec2280c99c6a28b9176365781f4f506fd30 policy/bagel_robotics_1/artifact-manifest.json
12
  76fec0d61c9344d8d2c5f61853a93269c8b19ae16d08de673f2ced5bdfe08b79 policy/bagel_robotics_1/assets.py
13
  2a16b03892ede7aa0a93cfd65e01a239498aa992bc4f1affb9cdde54253dae08 policy/bagel_robotics_1/backend.py
14
  68f337b21b1e517e1df27550e64fbc542b644528adc8f3f47fb8e44849185af6 policy/bagel_robotics_1/deploy.py
 
1
+ 9b146bf4e4b90a6bc16f5e904ee7b6a195e2ab99b7a1b2b6169dcbc614714182 README.md
2
  56f26e5b6ba87080c82703e102240fdfbd955c731fb2a044609ffff099795b3f install_adapter.py
3
  6c9437916a0dc5b6ed5fd56b361a4a1894be910e6965c9570e8683cc74a40813 policy/bagel_robotics_1/.gitignore
4
  24cb153ee8bababbfc20fe7acd36768923fbcf2a07d0aa653f460282761ccf8f policy/bagel_robotics_1/EVALUATION_RESULTS.md
5
  8379dddfbbb1154ad9cc57f008f1f8d4a8d26c3de99bd6b36aa4d8067888aa92 policy/bagel_robotics_1/LICENSE
6
+ 430474f9f20a388569d7d848a03c97b1fde55fcdaec8efe4c95db663f895789d policy/bagel_robotics_1/README.md
7
+ 3efd04c891dc86f4742e5a70a6ac19beb71c0d1a4767549afdf7f2dc87d791fd policy/bagel_robotics_1/REEVALUATION.md
8
  2f051521079a9de8815faafcb20098f1fe7fa167425e8c8088e7a42c2d50bb8e policy/bagel_robotics_1/THIRD_PARTY_NOTICES.md
9
  1d22bc8164fcca87f47b003406878112087118c4a5e23e6019ae1998e2b53418 policy/bagel_robotics_1/VALIDATION.md
10
  e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 policy/bagel_robotics_1/__init__.py
11
+ e7667cec0939b219dc1396dcf34c8fadcd95f509e6606df9757b6c909967b5f2 policy/bagel_robotics_1/artifact-manifest.json
12
  76fec0d61c9344d8d2c5f61853a93269c8b19ae16d08de673f2ced5bdfe08b79 policy/bagel_robotics_1/assets.py
13
  2a16b03892ede7aa0a93cfd65e01a239498aa992bc4f1affb9cdde54253dae08 policy/bagel_robotics_1/backend.py
14
  68f337b21b1e517e1df27550e64fbc542b644528adc8f3f47fb8e44849185af6 policy/bagel_robotics_1/deploy.py
inference/policy/bagel_robotics_1/README.md CHANGED
@@ -38,9 +38,15 @@ establish numerical equivalence. The recorded GPU has 96 GB VRAM; a lower
38
  minimum has not been qualified. The isolated validation jobs reserved 128 GiB
39
  host RAM and eight CPUs; these are tested allocations, not measured minima.
40
 
41
- **2. Download and verify before evaluation.** The checkpoint is public at
42
- [bageldotcom/adaprune](https://huggingface.co/bageldotcom/adaprune). No login, token
43
- or access request is required.
 
 
 
 
 
 
44
 
45
  ```bash
46
  PYTHON_BIN="$BAGEL_POLICY_ENV/bin/python" bash download_checkpoint.sh
@@ -153,8 +159,8 @@ The Installation quickstart authenticates and downloads before evaluation.
153
  The script downloads the six exact bundle files into `checkpoints/cp1000` and
154
  the pinned public VAE/CLIP/UMT5/tokenizer assets into
155
  `checkpoints/Wan2.1-I2V-14B-480P`. Bundle repository:
156
- [bageldotcom/adaprune](https://huggingface.co/bageldotcom/adaprune),
157
- revision `2651631c5d08819088475e8778676f5cf6d9afcf` (**public; no login required**).
158
  Access must be arranged for evaluators. No token is included.
159
 
160
  For existing artifacts, set `BAGEL_CP1000_BUNDLE` and `BAGEL_CP1000_WAN_ROOT`.
 
38
  minimum has not been qualified. The isolated validation jobs reserved 128 GiB
39
  host RAM and eight CPUs; these are tested allocations, not measured minima.
40
 
41
+ **2. Download and verify before evaluation.** The public inference release is
42
+ prepared at [bageldotcom/Bagel-Robotics-1](https://huggingface.co/bageldotcom/Bagel-Robotics-1).
43
+ The repository is currently private. Evaluators need a dedicated fine-grained
44
+ token granting read access only to this repository, supplied separately through
45
+ an agreed private channel. Credential creation, validation and handoff are pending.
46
+ Set `HF_TOKEN` securely before downloading (for example, `read -r -s HF_TOKEN`
47
+ followed by `export HF_TOKEN`; use `set +x` first and `unset HF_TOKEN` afterward).
48
+ Do not place the token in scripts, the PR, logs, or the application text.
49
+ Public release is a separate leaderboard milestone; once public, no token is needed.
50
 
51
  ```bash
52
  PYTHON_BIN="$BAGEL_POLICY_ENV/bin/python" bash download_checkpoint.sh
 
159
  The script downloads the six exact bundle files into `checkpoints/cp1000` and
160
  the pinned public VAE/CLIP/UMT5/tokenizer assets into
161
  `checkpoints/Wan2.1-I2V-14B-480P`. Bundle repository:
162
+ [bageldotcom/Bagel-Robotics-1](https://huggingface.co/bageldotcom/Bagel-Robotics-1),
163
+ revision `2651631c5d08819088475e8778676f5cf6d9afcf` (**public release prepared; visibility change pending**).
164
  Access must be arranged for evaluators. No token is included.
165
 
166
  For existing artifacts, set `BAGEL_CP1000_BUNDLE` and `BAGEL_CP1000_WAN_ROOT`.
inference/policy/bagel_robotics_1/REEVALUATION.md CHANGED
@@ -6,11 +6,11 @@ debug checks. The standalone inference runtime is included; the private research
6
  training repository is not needed. Install RoboDojo and its assets separately.
7
 
8
  The only policy checkpoint is `cp1000`, from
9
- [bageldotcom/adaprune](https://huggingface.co/bageldotcom/adaprune),
10
  revision `2651631c5d08819088475e8778676f5cf6d9afcf`. The downloader and loader reject
11
  files that do not match `artifact-manifest.json`. The VAE, image/text encoders and
12
  tokenizer are downloaded automatically at their separately pinned public revision.
13
- The checkpoint and all required public encoder assets download without authentication.
14
 
15
  ## Set the local paths
16
 
 
6
  training repository is not needed. Install RoboDojo and its assets separately.
7
 
8
  The only policy checkpoint is `cp1000`, from
9
+ [bageldotcom/Bagel-Robotics-1](https://huggingface.co/bageldotcom/Bagel-Robotics-1),
10
  revision `2651631c5d08819088475e8778676f5cf6d9afcf`. The downloader and loader reject
11
  files that do not match `artifact-manifest.json`. The VAE, image/text encoders and
12
  tokenizer are downloaded automatically at their separately pinned public revision.
13
+ The checkpoint is prepared for public access, pending an organization administrator visibility change. Once public, it and the required encoder assets download without authentication.
14
 
15
  ## Set the local paths
16
 
inference/policy/bagel_robotics_1/artifact-manifest.json CHANGED
@@ -4,9 +4,9 @@
4
  "checkpoint_identity": "790bffc9464fa7a026aa43b217aec2033d186895cad643166daa0bcc01914d5c",
5
  "update": 1000,
6
  "video_parameters": 12468409664,
7
- "repo_id": "bageldotcom/adaprune",
8
  "revision": "2651631c5d08819088475e8778676f5cf6d9afcf",
9
- "private": false,
10
  "files": {
11
  "COMPLETE.json": {
12
  "sha256": "eee062fbcb8c92a881117aa7f06b5cd039dfb9912d85571b59a20aa30837b0e1",
 
4
  "checkpoint_identity": "790bffc9464fa7a026aa43b217aec2033d186895cad643166daa0bcc01914d5c",
5
  "update": 1000,
6
  "video_parameters": 12468409664,
7
+ "repo_id": "bageldotcom/Bagel-Robotics-1",
8
  "revision": "2651631c5d08819088475e8778676f5cf6d9afcf",
9
+ "private": true,
10
  "files": {
11
  "COMPLETE.json": {
12
  "sha256": "eee062fbcb8c92a881117aa7f06b5cd039dfb9912d85571b59a20aa30837b0e1",