Update Dita training configuration and bundled runtime to 2.1.0

Enable full-backbone fine-tuning and inference checkpoint bypass; add native LoRA support. Align optimizer groups and the warmup/cosine schedule with the upstream LIBERO recipe, including LIBERO-Long's higher learning rate. Preserve all model weights and processors. Bundle the tested 2.1.0 wheel, source hashes, validation reports, and archived 2.0.1 metadata.

.gitattributes

@@ -36,3 +36,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 lerobot_policy_dita-1.0.5-py3-none-any.whl filter=lfs diff=lfs merge=lfs -text
 lerobot_policy_dita-2.0.0-py3-none-any.whl filter=lfs diff=lfs merge=lfs -text
 lerobot_policy_dita-2.0.1-py3-none-any.whl filter=lfs diff=lfs merge=lfs -text
+lerobot_policy_dita-2.1.0-py3-none-any.whl filter=lfs diff=lfs merge=lfs -text

GRIPPER_MODES.md

@@ -1,6 +1,6 @@
 # DITA gripper modes
 
-Runtime **2.0.1** supports `policy.gripper_mode=binary` and `continuous`. The mode is
+Runtime **2.1.0** supports `policy.gripper_mode=binary` and `continuous`. The mode is
 saved in `config.json` and both processor JSON files. The public LeRobot processor
 factory applies the active policy mode when loading pretrained templates, including
 CLI overrides. Conflicting explicit processor modes fail. New fine-tuning binds the
@@ -8,7 +8,7 @@ target dataset statistics; resume keeps the saved statistics.
 
 The config schema replaces the serialized `action_normalization_mask` and
 `gripper_sign_flip` fields with `gripper_mode`. Old policy and processor configs are
-not supported. Use the updated Hub revisions and their 2.0.1 wheel together.
+not supported. Use the updated Hub revisions and their 2.1.0 wheel together.
 
 | Contract | `binary` | `continuous` |
 | --- | --- | --- |
@@ -40,7 +40,7 @@ all converted LIBERO inference outputs unchanged.
 Install the runtime bundled with the updated initialization:
 
 ```bash
-pip install 'lerobot_policy_dita[training] @ https://huggingface.co/albus2024/dita_droid_pretrain/resolve/main/lerobot_policy_dita-2.0.1-py3-none-any.whl'
+pip install 'lerobot_policy_dita[training] @ https://huggingface.co/albus2024/dita_droid_pretrain/resolve/main/lerobot_policy_dita-2.1.0-py3-none-any.whl'
 ```
 
 LIBERO:

README.md

@@ -38,7 +38,7 @@ This repository is a strict LeRobot conversion of `libero_object.pth` from the o
 ## Install
 
 ```bash
-pip install "lerobot_policy_dita[libero] @ https://huggingface.co/albus2024/dita_object/resolve/main/lerobot_policy_dita-2.0.1-py3-none-any.whl"
+pip install "lerobot_policy_dita[libero] @ https://huggingface.co/albus2024/dita_object/resolve/main/lerobot_policy_dita-2.1.0-py3-none-any.whl"
 ```
 
 ## Evaluate
@@ -65,34 +65,68 @@ The converter checks all parameter names and shapes, loads with `strict=True`, a
 saved artifact files. Run the project's parity and converted-checkpoint verification scripts
 separately before using the policy; conversion alone does not establish rollout parity.
 
-## Runtime 2.0.0 gripper contract
-
-`gripper_mode` replaces the serialized action mask and gripper sign fields. Old
-policy/processor configs are not supported. Install the 2.0.1 wheel with this
-revision. See [gripper modes](GRIPPER_MODES.md) for training and SO101 EEF use.
-
-The four LIBERO models default to `binary`; the DROID initialization defaults to
-`continuous` and has no target-dataset statistics. For LIBERO fine-tuning from
-DROID, pass `--policy.gripper_mode=binary`. Both saved processors follow that
-override. Binary labels are -1=open/+1=closed; continuous gripper uses dataset
-min/max while pose uses quantiles.
-
-The archived [2.0.0 validation report](history/runtime-2.0.0/runtime_validation_report.json) covers real LIBERO and SO101 training (10 steps plus
-one resumed update), offline inference, wheel verification, four LIBERO checkpoint
-action parity, and three LIBERO Spatial simulator steps. These checks do not
-establish convergence, task success, or a physical SO101 rollout. Model weights
-and existing LIBERO statistics are unchanged in this update.
-
-## Runtime 2.0.1 precision configuration
-
-`policy.dtype` is the only precision setting: `float16` (default), `bfloat16`,
-or `float32`. `null` is invalid. `use_amp` is omitted from model `config.json`;
-the inherited LeRobot runtime flag is derived from dtype and does not select
-precision. Use `--policy.dtype=float32` to disable AMP, or
-`--policy.dtype=bfloat16` for BF16 mixed precision. Parameters remain FP32.
-
-This refresh removes only `use_amp` from the model configuration and packages
-the updated runtime. Model weights, processor files, action statistics and
-existing binary/continuous gripper modes are unchanged. See
-[runtime validation](runtime_validation_report.json) for the checks run with
-2.0.1; previous validation reports are historical.
+## Runtime 2.1.0 training configuration
+
+Install the **2.1.0 wheel above** for this revision. The new training fields
+require this runtime. The policy weights, processor files and action statistics
+are preserved; this is a configuration/runtime update, not a newly trained model.
+
+| Setting | Value |
+| --- | --- |
+| `freeze_backbone` | `false` |
+| `gradient_checkpointing` | `true`, only during gradient-enabled training |
+| `optimizer_lr` | `0.0001` |
+| `optimizer_backbone_lr_scale` | `0.1` |
+| AdamW betas / epsilon | `[0.9, 0.95]` / `1e-8` |
+| Weight decay | `0.05`; bias and one-dimensional parameters use `0` |
+| Gradient clipping | Disabled (`optimizer_grad_clip_norm=0`) |
+| LR schedule | 1000-step warmup, cosine to 100000 steps, minimum scale `0.01` |
+| `lora_dropout` | `0.0` |
+
+Inference automatically skips ViT and QFormer activation-checkpoint wrappers.
+`policy.dtype` remains `float16` by default; BF16 and FP32 are also supported.
+CLIP text encoding remains frozen. The gripper mode remains
+`binary`; see [gripper modes](GRIPPER_MODES.md).
+
+### Fine-tuning and LoRA
+
+Use the following for full fine-tuning (choose an appropriate target dataset):
+
+```bash
+lerobot-train \
+  --policy.path=albus2024/dita_object \
+  --policy.device=cuda \
+  --policy.dtype=bfloat16 \
+  --policy.input_features=null \
+  --policy.push_to_hub=false \
+  --dataset.repo_id=YOUR_USERNAME/YOUR_DATASET \
+  --output_dir=outputs/train/dita_object_full \
+  --steps=100000
+```
+
+For LoRA, install the optional dependency and add the adapter arguments:
+
+```bash
+pip install "lerobot_policy_dita[lora] @ https://huggingface.co/albus2024/dita_object/resolve/main/lerobot_policy_dita-2.1.0-py3-none-any.whl"
+# Add to lerobot-train and choose a fresh output directory:
+# --peft.method_type=LORA --peft.r=32
+```
+
+LoRA targets every Dita linear layer, including the ViT, QFormer/FiLM and action
+head. Only adapter weights are trained. The default alpha is twice the rank;
+initialization is Gaussian for A and zero for B. Use LeRobot's native policy
+factory/evaluation entry points to load saved adapters together with their base
+model. Do not set `policy.use_peft=true` when starting from a base checkpoint.
+
+The LR schedule keeps its configured 100000-step period even in shorter runs;
+the first update has zero LR. Set `policy.scheduler_decay_steps` explicitly for
+a different period. These defaults follow the upstream LIBERO recipe, including
+the higher `5e-4` peak LR for LIBERO-Long; DROID uses `1e-4` as a fine-tuning
+starting point and can be overridden for the target task. When adapting DROID
+to LIBERO, pass `--policy.gripper_mode=binary`.
+
+See [runtime validation](runtime_validation_report.json) for the checks performed
+with this release and [source hashes](runtime_source_manifest.json) for the exact
+packaged code. The previous [runtime 2.0.1 report](history/runtime-2.0.1/runtime_validation_report.json)
+is archived. Short training/inference checks do not establish convergence or
+robot task success.

artifact_sha256.json

@@ -1,21 +1,28 @@
 {
-  "GRIPPER_MODES.md": "8abbf9d35fcaef99b2ec536fb48dd304655332d4009faf89086b59c5dafd5461",
-  "README.md": "b36cb5100a3459339bb25ac556b779f0c1f6b5326bc90b7bb8e438992d9a86aa",
-  "config.json": "a7a815e98874d262a98d506a6b32d99d3b206cca377627ec7ae611f184358dea",
+  "GRIPPER_MODES.md": "a5c18991f3540c5ae338742d3650c6eace179d6ffb8a283ee352bdc95c246c06",
+  "README.md": "2b05bed04b404b7b2152659f6f64bc3c27025c2a8a1cf1446efd8c64c24cf3a6",
+  "config.json": "f62e186ccca501f429443ae002594919b3b933c7274a78f2cbedbb62e4c03905",
   "conversion_metadata.json": "b71fe2f657baf6cce196475da00a16f8938d4ad4d12d9e7a2f1368fa6822f5d5",
   "history/runtime-2.0.0/runtime_metadata.json": "bfd305faf2fd9a108cbbf2d6030bab3034e6177f5fbf32cc85fb1161746df324",
   "history/runtime-2.0.0/runtime_source_manifest.json": "9da1a65a4edc8a1fe56a580fe4ace28358b42393fb8a6d0170e6bb6a2df9af04",
   "history/runtime-2.0.0/runtime_validation_report.json": "1872e9a2560ff916960b55aa70b26f98ac328b2c1c16c7724b191bff69b81792",
+  "history/runtime-2.0.1/README.md": "b36cb5100a3459339bb25ac556b779f0c1f6b5326bc90b7bb8e438992d9a86aa",
+  "history/runtime-2.0.1/artifact_sha256.json": "7fa61b8bcf3e5c91dd81f5ee9054cca146a1c90b6a4c4851e16a7b6e027d3601",
+  "history/runtime-2.0.1/config.json": "a7a815e98874d262a98d506a6b32d99d3b206cca377627ec7ae611f184358dea",
+  "history/runtime-2.0.1/runtime_metadata.json": "321ad1e01b2b7af799257b9731f57b553afd44cedae336c8b0f21696ce9ff306",
+  "history/runtime-2.0.1/runtime_source_manifest.json": "0960d128de0ee20e9379e83f67be8242124cf57d189633f17c426786d9c8d32e",
+  "history/runtime-2.0.1/runtime_validation_report.json": "2abb4b4a61c465a27ffecad8401cca1abed459aa6c35acaf7ddd1bcce2a3a5af",
   "lerobot_policy_dita-1.0.5-py3-none-any.whl": "8e4a6c5bf9d52a636751b97237305c13454d8091a9eb900fabcac875a7ae809e",
   "lerobot_policy_dita-2.0.0-py3-none-any.whl": "b948a1e1b2f2fb22ab1f952195ccae3006e08ad3f2695825533262296395ce01",
   "lerobot_policy_dita-2.0.1-py3-none-any.whl": "85b9d092f9b2ad072b0d53aa2b197d9694116ad5a9677ee7620dbbc1fab53854",
+  "lerobot_policy_dita-2.1.0-py3-none-any.whl": "745162bb849fbbcd85507627df813757a33b0664b8d517d0befe6706f47e9f0f",
   "local_verification.json": "7421ef531e16b3d15d3d4a0f9d57f29d52b5b646790163090f14fec15efbd044",
   "model.safetensors": "941a7189946392390540e36cd8c4e9bd8b0813cc93af8acb8b6a5a0ecb4808d5",
   "policy_postprocessor.json": "2668a59651f8fc638914065f71d7efebc48addca535891e38d75d4c2ba5e924c",
   "policy_postprocessor_step_0_dita_masked_unnormalizer_processor.safetensors": "5a696aedec9ccfbcb3aa5977f827023ccefd1c85dc8e176e3e3dc82f9c2bd90b",
   "policy_preprocessor.json": "ead05cc569fedb9f4ad5bc1338af09cd2174a1b83ac324be56f5c48049286a01",
   "policy_preprocessor_step_4_dita_masked_normalizer_processor.safetensors": "5a696aedec9ccfbcb3aa5977f827023ccefd1c85dc8e176e3e3dc82f9c2bd90b",
-  "runtime_metadata.json": "321ad1e01b2b7af799257b9731f57b553afd44cedae336c8b0f21696ce9ff306",
-  "runtime_source_manifest.json": "0960d128de0ee20e9379e83f67be8242124cf57d189633f17c426786d9c8d32e",
-  "runtime_validation_report.json": "2abb4b4a61c465a27ffecad8401cca1abed459aa6c35acaf7ddd1bcce2a3a5af"
+  "runtime_metadata.json": "44886d6c9ba0d362edcfde5236085a7f35d6215ebfb7c8c8a212bcbe295a3453",
+  "runtime_source_manifest.json": "4b670906b4013a2e8dc17fb8c019b65df59d01fef53c8434193aeb38c78a9421",
+  "runtime_validation_report.json": "24da1087a545f1cad803c1e72e22e226cc39a08930b1bf3528def9fe498e22a8"
 }

config.json

@@ -31,7 +31,7 @@
   "num_layers": 12,
   "traj_length": 11,
   "dropout_rate": 0.0,
-  "freeze_backbone": true,
+  "freeze_backbone": false,
   "use_qformer": true,
   "prediction_type": "epsilon",
   "attn_implementation": "eager",
@@ -50,14 +50,20 @@
     "ACTION": "QUANTILES",
     "LANGUAGE": "IDENTITY"
   },
-  "optimizer_lr": 0.0002,
+  "optimizer_lr": 0.0001,
   "optimizer_betas": [
     0.9,
     0.95
   ],
   "optimizer_eps": 1e-08,
-  "optimizer_weight_decay": 0.01,
-  "optimizer_grad_clip_norm": 1.0,
+  "optimizer_weight_decay": 0.05,
+  "optimizer_grad_clip_norm": 0.0,
   "dtype": "float16",
-  "gripper_mode": "binary"
+  "gripper_mode": "binary",
+  "gradient_checkpointing": true,
+  "optimizer_backbone_lr_scale": 0.1,
+  "scheduler_warmup_steps": 1000,
+  "scheduler_decay_steps": 100000,
+  "scheduler_min_lr_scale": 0.01,
+  "lora_dropout": 0.0
 }

history/runtime-2.0.1/README.md

@@ -0,0 +1,98 @@
+---
+library_name: lerobot
+pipeline_tag: robotics
+tags:
+- lerobot
+- dita
+- libero
+---
+
+# Dita — libero_object_no_noops
+
+This repository is a strict LeRobot conversion of `libero_object.pth` from the original
+[Dita](https://github.com/RoboDita/Dita) implementation. It is configured for the
+`libero_object_no_noops` action distribution and is not interchangeable with another LIBERO suite.
+
+## Provenance and compatibility contract
+
+- Source checkpoint: `model_ckpt/libero_object.pth`
+- Source checkpoint SHA256: `96572a9468599abdc3c064cd64d41ad215aa16a85121866d90fa3a48155736d3`
+- Conversion code commit: `f506ceccb9b2816ffebb024b14aeaa2b34fbd46b`
+- Dataset statistics: `/home/shared/lyanam/huggingface/hub/datasets--openvla--modified_libero_rlds/snapshots/6ce6aaaaabdbe590b1eef5cd29c0d33f14a08551/libero_object_no_noops/1.0.0/dataset_statistics_42700def7a889694fac2f7b5feaab0e8752c9352c8a25b034d5353eb4e572459.json`
+- Dataset statistics SHA256: `718875b12d371e6d76be4efef946ef9436d69b84adc8a4105a9b34b354059d28`
+- Image processing: one RGB primary/agentview camera, resized to 224×224, followed once by
+  ImageNet mean `(0.485, 0.456, 0.406)` and std `(0.229, 0.224, 0.225)`.
+- Action processing: dimensions 0–5 use this suite's `q01`/`q99`; dimension 6 is not normalized.
+- Action mask: `[true, true, true, true, true, true, false]`.
+- q01: `[-0.5383928418159485, -0.8758928775787354, -0.9375, -0.06964285671710968, -0.11678571254014969, -0.15964286029338837, 0.0]`
+- q99: `[0.8464285731315613, 0.84375, 0.9375, 0.08142857253551483, 0.14892856776714325, 0.0867857113480568, 1.0]`
+- Gripper: normalized output is thresholded with `> 0.5`, then the Cartesian dimensions are
+  unnormalized, then the original Dita gripper sign flip is applied.
+- Rollout: `traj_length=11`, two-frame history, `chunk_size=10`, `n_action_steps=1`,
+  `num_inference_steps=10`, DDIM `prediction_type=epsilon`, and CUDA AMP by default.
+- `policy.dtype` controls compute precision (`float32`, `float16`, `bfloat16`);
+  parameters remain FP32. `dtype` is the sole precision setting; `use_amp` is derived internally.
+- LIBERO's LeRobot environment processor performs the 180° image rotation. The policy does not
+  rotate the image a second time.
+
+## Install
+
+```bash
+pip install "lerobot_policy_dita[libero] @ https://huggingface.co/albus2024/dita_object/resolve/main/lerobot_policy_dita-2.0.1-py3-none-any.whl"
+```
+
+## Evaluate
+
+```bash
+lerobot-eval \
+  --policy.path=albus2024/dita_object \
+  --policy.gripper_mode=binary \
+  --policy.device=cuda \
+  --env.type=libero \
+  --env.task=libero_object \
+  --env.max_parallel_tasks=1 \
+  --eval.batch_size=1 \
+  --eval.n_episodes=10 \
+  --output_dir=./outputs/dita_object
+```
+
+The processor state in this repository contains the required statistics; no source `.pth` file or
+local dataset metadata is needed at evaluation time.
+
+## Validation
+
+The converter checks all parameter names and shapes, loads with `strict=True`, and checks the
+saved artifact files. Run the project's parity and converted-checkpoint verification scripts
+separately before using the policy; conversion alone does not establish rollout parity.
+
+## Runtime 2.0.0 gripper contract
+
+`gripper_mode` replaces the serialized action mask and gripper sign fields. Old
+policy/processor configs are not supported. Install the 2.0.1 wheel with this
+revision. See [gripper modes](GRIPPER_MODES.md) for training and SO101 EEF use.
+
+The four LIBERO models default to `binary`; the DROID initialization defaults to
+`continuous` and has no target-dataset statistics. For LIBERO fine-tuning from
+DROID, pass `--policy.gripper_mode=binary`. Both saved processors follow that
+override. Binary labels are -1=open/+1=closed; continuous gripper uses dataset
+min/max while pose uses quantiles.
+
+The archived [2.0.0 validation report](history/runtime-2.0.0/runtime_validation_report.json) covers real LIBERO and SO101 training (10 steps plus
+one resumed update), offline inference, wheel verification, four LIBERO checkpoint
+action parity, and three LIBERO Spatial simulator steps. These checks do not
+establish convergence, task success, or a physical SO101 rollout. Model weights
+and existing LIBERO statistics are unchanged in this update.
+
+## Runtime 2.0.1 precision configuration
+
+`policy.dtype` is the only precision setting: `float16` (default), `bfloat16`,
+or `float32`. `null` is invalid. `use_amp` is omitted from model `config.json`;
+the inherited LeRobot runtime flag is derived from dtype and does not select
+precision. Use `--policy.dtype=float32` to disable AMP, or
+`--policy.dtype=bfloat16` for BF16 mixed precision. Parameters remain FP32.
+
+This refresh removes only `use_amp` from the model configuration and packages
+the updated runtime. Model weights, processor files, action statistics and
+existing binary/continuous gripper modes are unchanged. See
+[runtime validation](runtime_validation_report.json) for the checks run with
+2.0.1; previous validation reports are historical.

history/runtime-2.0.1/artifact_sha256.json

@@ -0,0 +1,21 @@
+{
+  "GRIPPER_MODES.md": "8abbf9d35fcaef99b2ec536fb48dd304655332d4009faf89086b59c5dafd5461",
+  "README.md": "b36cb5100a3459339bb25ac556b779f0c1f6b5326bc90b7bb8e438992d9a86aa",
+  "config.json": "a7a815e98874d262a98d506a6b32d99d3b206cca377627ec7ae611f184358dea",
+  "conversion_metadata.json": "b71fe2f657baf6cce196475da00a16f8938d4ad4d12d9e7a2f1368fa6822f5d5",
+  "history/runtime-2.0.0/runtime_metadata.json": "bfd305faf2fd9a108cbbf2d6030bab3034e6177f5fbf32cc85fb1161746df324",
+  "history/runtime-2.0.0/runtime_source_manifest.json": "9da1a65a4edc8a1fe56a580fe4ace28358b42393fb8a6d0170e6bb6a2df9af04",
+  "history/runtime-2.0.0/runtime_validation_report.json": "1872e9a2560ff916960b55aa70b26f98ac328b2c1c16c7724b191bff69b81792",
+  "lerobot_policy_dita-1.0.5-py3-none-any.whl": "8e4a6c5bf9d52a636751b97237305c13454d8091a9eb900fabcac875a7ae809e",
+  "lerobot_policy_dita-2.0.0-py3-none-any.whl": "b948a1e1b2f2fb22ab1f952195ccae3006e08ad3f2695825533262296395ce01",
+  "lerobot_policy_dita-2.0.1-py3-none-any.whl": "85b9d092f9b2ad072b0d53aa2b197d9694116ad5a9677ee7620dbbc1fab53854",
+  "local_verification.json": "7421ef531e16b3d15d3d4a0f9d57f29d52b5b646790163090f14fec15efbd044",
+  "model.safetensors": "941a7189946392390540e36cd8c4e9bd8b0813cc93af8acb8b6a5a0ecb4808d5",
+  "policy_postprocessor.json": "2668a59651f8fc638914065f71d7efebc48addca535891e38d75d4c2ba5e924c",
+  "policy_postprocessor_step_0_dita_masked_unnormalizer_processor.safetensors": "5a696aedec9ccfbcb3aa5977f827023ccefd1c85dc8e176e3e3dc82f9c2bd90b",
+  "policy_preprocessor.json": "ead05cc569fedb9f4ad5bc1338af09cd2174a1b83ac324be56f5c48049286a01",
+  "policy_preprocessor_step_4_dita_masked_normalizer_processor.safetensors": "5a696aedec9ccfbcb3aa5977f827023ccefd1c85dc8e176e3e3dc82f9c2bd90b",
+  "runtime_metadata.json": "321ad1e01b2b7af799257b9731f57b553afd44cedae336c8b0f21696ce9ff306",
+  "runtime_source_manifest.json": "0960d128de0ee20e9379e83f67be8242124cf57d189633f17c426786d9c8d32e",
+  "runtime_validation_report.json": "2abb4b4a61c465a27ffecad8401cca1abed459aa6c35acaf7ddd1bcce2a3a5af"
+}

history/runtime-2.0.1/config.json

@@ -0,0 +1,63 @@
+{
+  "type": "dita",
+  "n_obs_steps": 1,
+  "input_features": {
+    "observation.images.image": {
+      "type": "VISUAL",
+      "shape": [
+        3,
+        256,
+        256
+      ]
+    }
+  },
+  "output_features": {
+    "action": {
+      "type": "ACTION",
+      "shape": [
+        7
+      ]
+    }
+  },
+  "device": "cuda",
+  "use_peft": false,
+  "push_to_hub": false,
+  "repo_id": null,
+  "private": null,
+  "tags": null,
+  "license": null,
+  "pretrained_path": null,
+  "dataset_suite": "libero_object_no_noops",
+  "num_layers": 12,
+  "traj_length": 11,
+  "dropout_rate": 0.0,
+  "freeze_backbone": true,
+  "use_qformer": true,
+  "prediction_type": "epsilon",
+  "attn_implementation": "eager",
+  "num_inference_steps": 10,
+  "tokenizer_max_length": 77,
+  "image_resolution": [
+    224,
+    224
+  ],
+  "chunk_size": 10,
+  "n_action_steps": 1,
+  "history_frames": 2,
+  "gripper_threshold": 0.5,
+  "normalization_mapping": {
+    "VISUAL": "IDENTITY",
+    "ACTION": "QUANTILES",
+    "LANGUAGE": "IDENTITY"
+  },
+  "optimizer_lr": 0.0002,
+  "optimizer_betas": [
+    0.9,
+    0.95
+  ],
+  "optimizer_eps": 1e-08,
+  "optimizer_weight_decay": 0.01,
+  "optimizer_grad_clip_norm": 1.0,
+  "dtype": "float16",
+  "gripper_mode": "binary"
+}

history/runtime-2.0.1/runtime_metadata.json

@@ -0,0 +1,28 @@
+{
+  "updated_at_utc": "2026-09-14T11:02:40.266967+00:00",
+  "runtime_version": "2.0.1",
+  "runtime_wheel": "lerobot_policy_dita-2.0.1-py3-none-any.whl",
+  "runtime_wheel_sha256": "85b9d092f9b2ad072b0d53aa2b197d9694116ad5a9677ee7620dbbc1fab53854",
+  "source_git_head": "fcca4f20c77503155e82518262a347176f737192",
+  "source_git_dirty": true,
+  "source_manifest": "runtime_source_manifest.json",
+  "previous_hub_revision": "108698b2b03333642fb838515a233c68ee2d146d",
+  "model_sha256": "941a7189946392390540e36cd8c4e9bd8b0813cc93af8acb8b6a5a0ecb4808d5",
+  "model_weights_changed": false,
+  "processor_configs_changed": false,
+  "processor_statistics_changed": false,
+  "processor_sha256": {
+    "policy_postprocessor_step_0_dita_masked_unnormalizer_processor.safetensors": "5a696aedec9ccfbcb3aa5977f827023ccefd1c85dc8e176e3e3dc82f9c2bd90b",
+    "policy_preprocessor_step_4_dita_masked_normalizer_processor.safetensors": "5a696aedec9ccfbcb3aa5977f827023ccefd1c85dc8e176e3e3dc82f9c2bd90b",
+    "policy_preprocessor.json": "ead05cc569fedb9f4ad5bc1338af09cd2174a1b83ac324be56f5c48049286a01",
+    "policy_postprocessor.json": "2668a59651f8fc638914065f71d7efebc48addca535891e38d75d4c2ba5e924c"
+  },
+  "configuration_change": {
+    "removed": [
+      "use_amp"
+    ],
+    "dtype": "float16"
+  },
+  "gripper_mode": "binary",
+  "validation_report": "runtime_validation_report.json"
+}

history/runtime-2.0.1/runtime_source_manifest.json

@@ -0,0 +1,87 @@
+{
+  "wheel": "lerobot_policy_dita-2.0.1-py3-none-any.whl",
+  "sha256": "85b9d092f9b2ad072b0d53aa2b197d9694116ad5a9677ee7620dbbc1fab53854",
+  "source_python_files": 79,
+  "all_source_files_match": true,
+  "source_sha256": {
+    "lerobot_policy_dita/__init__.py": "1d66ff4033566f53629f7e6c5f9f38f79d9790fd5e46bd46cb3e8566bc3caf7d",
+    "lerobot_policy_dita/_loading.py": "325e7c57a422f2cc5827c6d95ad19df4b51acae4af0a9b70b0c98a37852e6689",
+    "lerobot_policy_dita/approximation/__init__.py": "b0646b441737d19c58ce395dd402fa0c1dcb0e8531e4df8d8be251de20b11d5e",
+    "lerobot_policy_dita/approximation/adapters/__init__.py": "e0b5bf1a2a01a82af6b1ab03954aa039f396096b8b3a67d17483a7f054fd98c5",
+    "lerobot_policy_dita/approximation/adapters/base.py": "4acb6cea075925f00ba2bff356e0cafc798458c18b19790ae17fd3c53fe86811",
+    "lerobot_policy_dita/approximation/backends/__init__.py": "c398254ffabafa0b73c50d7f59e2c838aeb5a6c99d195ef22701f3c7bc89d6c1",
+    "lerobot_policy_dita/approximation/backends/torch.py": "fc61e6bcaa7d571c13048a0f8b329e5118b99a870ed6e4600eedadd9e1cfb62f",
+    "lerobot_policy_dita/approximation/operators/__init__.py": "fe875d6ddc6f5160428dcef4ba253b12f445906070aa57dddff6e3dd71932bdd",
+    "lerobot_policy_dita/approximation/operators/ddim.py": "d453de05452c9a7297e971800416cfbe8dbded0482c2ff49c05e9e6a213a038d",
+    "lerobot_policy_dita/approximation/operators/precision.py": "a19c5045814278929aaa181e637b32e8a578146801c4a9d29b822ea99635c7ee",
+    "lerobot_policy_dita/approximation/operators/rmsnorm.py": "89edadb7872af468f0a8b56a8dc1d48fb51f3ffda1b76c4ec5108d4c71f66bb5",
+    "lerobot_policy_dita/approximation/operators/silu.py": "06ac4a8e519eadb01695026d2a2149195e1d7e1136df65e7d7e0e2b3eb17f287",
+    "lerobot_policy_dita/approximation/operators/softmax.py": "407e4adebaa4d94ce4360915d6e3df66d85c74c27fc4b030cd234286a2c0d22f",
+    "lerobot_policy_dita/approximation/runtime/__init__.py": "aa53dcaadd7d11e2427d902fb6c70cca7985ce87cc91bfec0232b04fbe2e3307",
+    "lerobot_policy_dita/approximation/runtime/configuration.py": "4ba580d67c078ed4ff3b5a3f71cf9eaa87f29f1fc08bc6fd6ff831f4dc2c18de",
+    "lerobot_policy_dita/approximation/spec.py": "0ed50dc74f1abeb2d19819e3bec1ae852ff2ebfe179da5832e36590ee4c4560a",
+    "lerobot_policy_dita/benchmarking/__init__.py": "87e434eb6b29bff4b14b701fc85c9293bfe12e93726c72edaa893be2d47cafc4",
+    "lerobot_policy_dita/benchmarking/dataset_sampler.py": "db3d1f49d604bc882db38e48f32d0d48af0f1699d3bc95ae2359d67a1b70dbe5",
+    "lerobot_policy_dita/evaluation/__init__.py": "3adb7eff2a3bea441a95726cca157b9502555afab6bf4062e97b62bb1bb53419",
+    "lerobot_policy_dita/evaluation/cli/__init__.py": "5a7fb6106f531d483e0795d3d52f95ab33ccef955ea4e54ef344da7e6806d503",
+    "lerobot_policy_dita/evaluation/cli/_workload.py": "c62648d67766d8831c08729551532b6aa962a2b513afd17dee9534119b53531a",
+    "lerobot_policy_dita/evaluation/cli/compare_dynamic.py": "44b467e1955eda66541391abb6733cae9a78c86d060ef41a5a2e6156287ccb04",
+    "lerobot_policy_dita/evaluation/cli/evaluate_dynamic.py": "e92fd7f8a6b724ff80a8b61905a4498cbb472faa2b6f419d8ef0381a59755db6",
+    "lerobot_policy_dita/evaluation/cli/evaluate_static.py": "0fb864387dad74ec09fc7fe9d6cb57bd9cc9014094de904179c74bf54af9d11c",
+    "lerobot_policy_dita/evaluation/cli/trace.py": "b78981d14701f11d6bed3bdad84cf2ae444e08d1aa32e0aa3dfd16b6e2700365",
+    "lerobot_policy_dita/evaluation/dynamic.py": "3379df73f478ba5b5959ef340e1d9ce88e96418b131f804a24315b55b93f7215",
+    "lerobot_policy_dita/evaluation/static.py": "758ea97ef6b68c0a32197d962d9ff9af62b2aeb337c248bc72d8d960cdeba712",
+    "lerobot_policy_dita/evaluation/trace.py": "a4399277a49e412589df69529d952e5931d87742f0a6dd16832d39cba113e1d3",
+    "lerobot_policy_dita/policy/__init__.py": "9785f8e7279dc6760b4b7dbbe587ed6c6f467daa1dbbdf2ea095a645c1c52a57",
+    "lerobot_policy_dita/policy/_lerobot_compat.py": "d3434f59edf9d1dbb5616b874d0d6babf67a254f0b48196559e06d0dae1c6214",
+    "lerobot_policy_dita/policy/approximation/__init__.py": "0632173bd9a1595a72c9741e6bb5538f644c0f448c8caca606542ae3d2cb3b5a",
+    "lerobot_policy_dita/policy/approximation/adapter.py": "d603f444c2237a76a23f44dd8b2df734b66525d3522695f200478b3c055152a4",
+    "lerobot_policy_dita/policy/approximation/configuration.py": "8fa1f0f24624352b691d73c71f9fc75c0196c8f81e888e6b77a9f11eba7926af",
+    "lerobot_policy_dita/policy/approximation/ddim.py": "d32a16f85aa670bd07375ce2691498c47b9f9e700160fde12eea33e6a56f09fa",
+    "lerobot_policy_dita/policy/attention.py": "3253866b5ea0bdf24a85b2047d813573e24588a4294adc9b72d01488d236909e",
+    "lerobot_policy_dita/policy/configuration_dita.py": "edc19d5e348270b9bebf22a0bc48d6dd1a56b084099544513e02c4d724f954ac",
+    "lerobot_policy_dita/policy/evaluation/__init__.py": "8c0ca0bbd994637ce95307daec7647082fdb3e8e982073ab6c60ba7f9a16745f",
+    "lerobot_policy_dita/policy/evaluation/adapter.py": "fdbfadfbfa13772eef20a513c74a06cda0e6cea456a4b2197cb314af42c2dbec",
+    "lerobot_policy_dita/policy/evaluation/static.py": "d5d2cb5c86b774eb7215b0b36c565888a6fc620da430a212a79b278a6c2d9f31",
+    "lerobot_policy_dita/policy/modeling_dita.py": "1fb98b081deb3c47f51180e31874abdf342b2fb9e5b8d5d8d7bef5e60a0c7a1d",
+    "lerobot_policy_dita/policy/processor_dita.py": "3173c16c7be81bdb5f86cce305eee8f96a16235acfb45586f874e80324f8257b",
+    "lerobot_policy_dita/policy/quantization/__init__.py": "8b130b24c8c2a6e5b585c1e5c80e37ba317578ded5bfc624b68a9548d82fcd67",
+    "lerobot_policy_dita/policy/quantization/adapter.py": "1633d4eb613972afe14d0de3138ec45391764702449e1b7f6312cfeb209d2068",
+    "lerobot_policy_dita/policy/quantization/attention.py": "e54d5105cb5930050b268d53e098bfcde6303908884f3d909e83bb50e740ec5e",
+    "lerobot_policy_dita/policy/quantization/configuration.py": "f567c8cc1c12a6d11792e11d9b514234ad45f668ee3c602aa57778fae8d485a8",
+    "lerobot_policy_dita/policy/quantization/export.py": "ee6d4a9d97fd2b865c9db192cf844d16a5db4134a2e6ac1810d7574c2f1836f1",
+    "lerobot_policy_dita/policy/quantization/factory.py": "59cd054211f2654e0f1d25dc5ad97f317f85e16088e529f2e6ab46ada4102949",
+    "lerobot_policy_dita/policy/quantization/spec.py": "1261fc33bf61aede0794066ad521b05df7ed3b39be5001bdaafc724a2f19663f",
+    "lerobot_policy_dita/policy/runtime.py": "87c21f4d58c9345acb8d4e54408987b00d2cc0e105fb14d233b75fbd8b331ff1",
+    "lerobot_policy_dita/policy/workflow.py": "c64442c65115237a5c67d96c6bd8b425e1b1c2094f4470bd3b9c053d1afdce7d",
+    "lerobot_policy_dita/quantization/__init__.py": "ebb9b9fd422b62ce670ddb74e4b257d32754297b9d8e20bc57bfbad6de16544d",
+    "lerobot_policy_dita/quantization/adapters/__init__.py": "da73ba087d9c60cfa06f067f7cf6f2b345683dd93a15916515e3b2cdcd57268b",
+    "lerobot_policy_dita/quantization/adapters/base.py": "9afbad7b3d27498384d8a9f39a04da88351799e243757dd4bf19ca41190c7fdc",
+    "lerobot_policy_dita/quantization/artifacts.py": "2b1da7433310adef900fbec56b45ea777ec67044e5cc121db0dc04bec3819904",
+    "lerobot_policy_dita/quantization/attention.py": "d46ed81671de87d338c6092a8f1dc0037de1d1c5d2ecb79da10cfe66aa592dce",
+    "lerobot_policy_dita/quantization/backends/__init__.py": "5fdd47d2e09eef5aac8f91724e6c6c69fe46fd02c3b4c4820ac5e61e3d073f25",
+    "lerobot_policy_dita/quantization/backends/fake_quant.py": "3c3cdebcd256bf5d65d0a56bf0fa34ad1c2142f068f1762d752754ec21c9f42a",
+    "lerobot_policy_dita/quantization/calibration.py": "242df70e46ccb59070a23da1b8617805e86027859caae18f05bc6501ae44498d",
+    "lerobot_policy_dita/quantization/cli/__init__.py": "0dac1165b80f8eb27606f2786549f17527787ad5383433dc8bfd1f1e64d191cd",
+    "lerobot_policy_dita/quantization/cli/_workload.py": "7324ec6b3818cf869d6cd28b24bc38c028717f59373b95b0be9c86f50985a900",
+    "lerobot_policy_dita/quantization/cli/calibrate.py": "34d709f49e0fe7fe4155aba842528b212eaa10be61b4f62d025d373bbde834b4",
+    "lerobot_policy_dita/quantization/cli/export_onnx_qdq.py": "b7e46aa84490f167906c8ee6e1babae0dcef166d229c3aa2bedc3bcb84400e35",
+    "lerobot_policy_dita/quantization/cli/inspect_artifact.py": "e3a815325e1d475efd082bf16e0f6f5f2f3f6ef093d81a2719352fda346d344c",
+    "lerobot_policy_dita/quantization/evaluation/__init__.py": "51870904f114404daf5e5a68fe260ebcb07c22c4b80f979be4ff33f0eefcd46c",
+    "lerobot_policy_dita/quantization/evaluation/saturation.py": "1caf177485695c9b797b798d41c4ec3de025372b854b18c867aa15fda3851d2e",
+    "lerobot_policy_dita/quantization/export/__init__.py": "a1d612b846eb72502d49fefb826316a107f87fbd0176021bfb22afe8f11b9aae",
+    "lerobot_policy_dita/quantization/export/onnx_qdq.py": "fecd53764a6cd0c375565951adb4265d13dbd2e889ec27e380f06dbc15bdaa68",
+    "lerobot_policy_dita/quantization/math.py": "fc6f78a03eca538b27074dd6872ee0977f898c79f2f1acc4e6f9dc3710855eb8",
+    "lerobot_policy_dita/quantization/observers.py": "13ec75c6e09eadb86f621fb14a309e6e80414ba6439c42ceb65ac8426f1d116b",
+    "lerobot_policy_dita/quantization/runtime/__init__.py": "4b2d994863700e986860834d9b328b72eb0a8fecdc6a07cad2b9a2c101e3371b",
+    "lerobot_policy_dita/quantization/runtime/configuration.py": "c09e263598dc058bc1b4a0d9c27846f20c737a37dcaae67063ef96c3de5d2652",
+    "lerobot_policy_dita/quantization/spec.py": "4cd8766b6ad9bb8c4f64e4735dce32c089802acb0cf95a05f402bd2a0b397071",
+    "lerobot_policy_dita/scripts/__init__.py": "08089d9d38c1f9a32032e18f218b5d517705c34c67aa1f9d34839c7a12e85389",
+    "lerobot_policy_dita/scripts/calibrate_libero_dataset.py": "731e960fce5c9eefe48ec878cf8e86d1ccc7a1931ea4083d84ed6f2b464151b9",
+    "lerobot_policy_dita/scripts/evaluate_workflow.py": "38da1221a985fbbcabd82cf86c3179509cf2732f0d96a6e12861f9bd0073be83",
+    "lerobot_policy_dita/scripts/trace_libero_dataset.py": "cbfaec59e385d23bf669cd675eabfb59ae246ef28f86e9c738023accdf3f1de0",
+    "lerobot_policy_dita/workflows/__init__.py": "1d34c52aa0e51861c390f50cb70b0f5feda49ef87ad23b919ffeace5279e2ca8",
+    "lerobot_policy_dita/workflows/configuration.py": "ccb63b5dad8da24b56359ea605bb088fda516df4465f6de93e603ca4ebd7de26",
+    "lerobot_policy_dita/workflows/runtime.py": "b83f86a5df4ef4f8852ee375a41b48037d7f68a1d32cc19dd1130e17b7861b89"
+  }
+}

history/runtime-2.0.1/runtime_validation_report.json

@@ -0,0 +1,36 @@
+{
+  "source": "albus2024/dita_object",
+  "dataset_suite": "libero_object_no_noops",
+  "strict_from_pretrained": true,
+  "processor_stats_roundtrip": true,
+  "model_roundtrip_max_abs_diff": 0.0,
+  "inference_steps": 3,
+  "actions_finite": true,
+  "action_shape": [
+    1,
+    7
+  ],
+  "gripper_values": [
+    -1.0,
+    -1.0,
+    -1.0
+  ],
+  "history_lengths": [
+    1,
+    2,
+    2
+  ],
+  "result": "passed",
+  "runtime_version": "2.0.1",
+  "runtime_from_bundled_wheel": true,
+  "dtype_cli_overrides": [
+    "float32",
+    "float16",
+    "bfloat16"
+  ],
+  "serialized_use_amp_absent": true,
+  "dtype_and_derived_amp_roundtrip": true,
+  "weights_and_processor_files_unchanged": true,
+  "pytest_passed": 309,
+  "full_task_benchmark_run": false
+}

lerobot_policy_dita-2.1.0-py3-none-any.whl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:745162bb849fbbcd85507627df813757a33b0664b8d517d0befe6706f47e9f0f
+size 118699

runtime_metadata.json

@@ -1,27 +1,63 @@
 {
-  "updated_at_utc": "2026-09-14T11:02:40.266967+00:00",
-  "runtime_version": "2.0.1",
-  "runtime_wheel": "lerobot_policy_dita-2.0.1-py3-none-any.whl",
-  "runtime_wheel_sha256": "85b9d092f9b2ad072b0d53aa2b197d9694116ad5a9677ee7620dbbc1fab53854",
-  "source_git_head": "fcca4f20c77503155e82518262a347176f737192",
+  "updated_at_utc": "2026-09-15T02:22:48.008083+00:00",
+  "runtime_version": "2.1.0",
+  "runtime_wheel": "lerobot_policy_dita-2.1.0-py3-none-any.whl",
+  "runtime_wheel_sha256": "745162bb849fbbcd85507627df813757a33b0664b8d517d0befe6706f47e9f0f",
+  "source_git_head": "67e68060fa3c51923dc1ab1332604b266c1e2b1d",
   "source_git_dirty": true,
   "source_manifest": "runtime_source_manifest.json",
-  "previous_hub_revision": "108698b2b03333642fb838515a233c68ee2d146d",
+  "previous_hub_revision": "81d19798426d81a857b983831051454a3a134b12",
   "model_sha256": "941a7189946392390540e36cd8c4e9bd8b0813cc93af8acb8b6a5a0ecb4808d5",
   "model_weights_changed": false,
   "processor_configs_changed": false,
   "processor_statistics_changed": false,
   "processor_sha256": {
+    "policy_postprocessor.json": "2668a59651f8fc638914065f71d7efebc48addca535891e38d75d4c2ba5e924c",
     "policy_postprocessor_step_0_dita_masked_unnormalizer_processor.safetensors": "5a696aedec9ccfbcb3aa5977f827023ccefd1c85dc8e176e3e3dc82f9c2bd90b",
-    "policy_preprocessor_step_4_dita_masked_normalizer_processor.safetensors": "5a696aedec9ccfbcb3aa5977f827023ccefd1c85dc8e176e3e3dc82f9c2bd90b",
     "policy_preprocessor.json": "ead05cc569fedb9f4ad5bc1338af09cd2174a1b83ac324be56f5c48049286a01",
-    "policy_postprocessor.json": "2668a59651f8fc638914065f71d7efebc48addca535891e38d75d4c2ba5e924c"
+    "policy_preprocessor_step_4_dita_masked_normalizer_processor.safetensors": "5a696aedec9ccfbcb3aa5977f827023ccefd1c85dc8e176e3e3dc82f9c2bd90b"
   },
   "configuration_change": {
-    "removed": [
-      "use_amp"
-    ],
-    "dtype": "float16"
+    "freeze_backbone": {
+      "previous": true,
+      "current": false
+    },
+    "optimizer_lr": {
+      "previous": 0.0002,
+      "current": 0.0001
+    },
+    "optimizer_weight_decay": {
+      "previous": 0.01,
+      "current": 0.05
+    },
+    "optimizer_grad_clip_norm": {
+      "previous": 1.0,
+      "current": 0.0
+    },
+    "gradient_checkpointing": {
+      "previous": null,
+      "current": true
+    },
+    "optimizer_backbone_lr_scale": {
+      "previous": null,
+      "current": 0.1
+    },
+    "scheduler_warmup_steps": {
+      "previous": null,
+      "current": 1000
+    },
+    "scheduler_decay_steps": {
+      "previous": null,
+      "current": 100000
+    },
+    "scheduler_min_lr_scale": {
+      "previous": null,
+      "current": 0.01
+    },
+    "lora_dropout": {
+      "previous": null,
+      "current": 0.0
+    }
   },
   "gripper_mode": "binary",
   "validation_report": "runtime_validation_report.json"

runtime_source_manifest.json

@@ -1,10 +1,10 @@
 {
-  "wheel": "lerobot_policy_dita-2.0.1-py3-none-any.whl",
-  "sha256": "85b9d092f9b2ad072b0d53aa2b197d9694116ad5a9677ee7620dbbc1fab53854",
-  "source_python_files": 79,
+  "wheel": "lerobot_policy_dita-2.1.0-py3-none-any.whl",
+  "sha256": "745162bb849fbbcd85507627df813757a33b0664b8d517d0befe6706f47e9f0f",
+  "source_python_files": 80,
   "all_source_files_match": true,
   "source_sha256": {
-    "lerobot_policy_dita/__init__.py": "1d66ff4033566f53629f7e6c5f9f38f79d9790fd5e46bd46cb3e8566bc3caf7d",
+    "lerobot_policy_dita/__init__.py": "54728aea71d40c0ba779ac4d0cadaa5d00a6c0aab686f38129774d786d93756e",
     "lerobot_policy_dita/_loading.py": "325e7c57a422f2cc5827c6d95ad19df4b51acae4af0a9b70b0c98a37852e6689",
     "lerobot_policy_dita/approximation/__init__.py": "b0646b441737d19c58ce395dd402fa0c1dcb0e8531e4df8d8be251de20b11d5e",
     "lerobot_policy_dita/approximation/adapters/__init__.py": "e0b5bf1a2a01a82af6b1ab03954aa039f396096b8b3a67d17483a7f054fd98c5",
@@ -39,11 +39,12 @@
     "lerobot_policy_dita/policy/approximation/configuration.py": "8fa1f0f24624352b691d73c71f9fc75c0196c8f81e888e6b77a9f11eba7926af",
     "lerobot_policy_dita/policy/approximation/ddim.py": "d32a16f85aa670bd07375ce2691498c47b9f9e700160fde12eea33e6a56f09fa",
     "lerobot_policy_dita/policy/attention.py": "3253866b5ea0bdf24a85b2047d813573e24588a4294adc9b72d01488d236909e",
-    "lerobot_policy_dita/policy/configuration_dita.py": "edc19d5e348270b9bebf22a0bc48d6dd1a56b084099544513e02c4d724f954ac",
+    "lerobot_policy_dita/policy/configuration_dita.py": "c202b51d2dabea0b111d342c3ab62e3ecec8ef5bfa75c9a6995da43aedb2ec9c",
     "lerobot_policy_dita/policy/evaluation/__init__.py": "8c0ca0bbd994637ce95307daec7647082fdb3e8e982073ab6c60ba7f9a16745f",
     "lerobot_policy_dita/policy/evaluation/adapter.py": "fdbfadfbfa13772eef20a513c74a06cda0e6cea456a4b2197cb314af42c2dbec",
     "lerobot_policy_dita/policy/evaluation/static.py": "d5d2cb5c86b774eb7215b0b36c565888a6fc620da430a212a79b278a6c2d9f31",
-    "lerobot_policy_dita/policy/modeling_dita.py": "1fb98b081deb3c47f51180e31874abdf342b2fb9e5b8d5d8d7bef5e60a0c7a1d",
+    "lerobot_policy_dita/policy/modeling_dita.py": "84ca6d1e4e7aefddb2758385c61be7ff11fa836a835807bcba944a9b7f924f20",
+    "lerobot_policy_dita/policy/optimization.py": "26af1c104eb8f247cd13918b4d27b367803abd96425539b1dbeb362a7a653fe2",
     "lerobot_policy_dita/policy/processor_dita.py": "3173c16c7be81bdb5f86cce305eee8f96a16235acfb45586f874e80324f8257b",
     "lerobot_policy_dita/policy/quantization/__init__.py": "8b130b24c8c2a6e5b585c1e5c80e37ba317578ded5bfc624b68a9548d82fcd67",
     "lerobot_policy_dita/policy/quantization/adapter.py": "1633d4eb613972afe14d0de3138ec45391764702449e1b7f6312cfeb209d2068",

runtime_validation_report.json

@@ -1,36 +1,40 @@
 {
-  "source": "albus2024/dita_object",
-  "dataset_suite": "libero_object_no_noops",
-  "strict_from_pretrained": true,
-  "processor_stats_roundtrip": true,
-  "model_roundtrip_max_abs_diff": 0.0,
-  "inference_steps": 3,
-  "actions_finite": true,
-  "action_shape": [
-    1,
-    7
-  ],
-  "gripper_values": [
-    -1.0,
-    -1.0,
-    -1.0
-  ],
-  "history_lengths": [
-    1,
-    2,
-    2
-  ],
-  "result": "passed",
-  "runtime_version": "2.0.1",
+  "runtime_version": "2.1.0",
   "runtime_from_bundled_wheel": true,
-  "dtype_cli_overrides": [
-    "float32",
-    "float16",
-    "bfloat16"
-  ],
-  "serialized_use_amp_absent": true,
-  "dtype_and_derived_amp_roundtrip": true,
+  "torch_version": "2.9.1+cu128",
+  "lerobot_version": "0.6.1",
+  "source": "albus2024/dita_object",
+  "previous_hub_revision": "81d19798426d81a857b983831051454a3a134b12",
+  "pytest_passed": 321,
+  "inference": {
+    "strict_reload": true,
+    "parameter_tensors": 373,
+    "checkpoint_calls_during_inference": 0,
+    "normalized_actions": [
+      [
+        -0.17592374980449677,
+        -0.18163743615150452,
+        -0.35942649841308594,
+        -0.06041604280471802,
+        -0.09618508070707321,
+        0.26031768321990967,
+        1.0
+      ]
+    ],
+    "gripper_mode": "binary",
+    "actions": [
+      [
+        0.03220635652542114,
+        -0.17224717140197754,
+        -0.33696234226226807,
+        0.0013292878866195679,
+        0.003292553126811981,
+        -0.004353716969490051,
+        -1.0
+      ]
+    ],
+    "previous_runtime_action_parity": "exact"
+  },
   "weights_and_processor_files_unchanged": true,
-  "pytest_passed": 309,
   "full_task_benchmark_run": false
 }