repro: locked Ego-Exo and robot environment snapshots, installation and training recipes

CHECKSUMS.sha256

@@ -0,0 +1,46 @@
+4ef5bb645114b5c72c1f794a6d5e418b9bcf10a08eb2f06f28c1ea52ae8338d0  INSTALL_COMMANDS.sh
+41abcf86b8bd9c17bbe4c0b23428fafde870a11a9ecbd90512cafdb5f5e2241e  README.md
+bffaadfbc5c59c5c9761e3608d33c20b27b2cdbfd39a98c49d379657ccde197f  SOURCE_LOCK.json
+38170c0681f48407253459b44868c2e13bf21bccd84395372ee0df7c1736f98a  SOURCE_OVERLAY_MANIFEST.json
+2569cdcdb30128ffb7d1f178a2669580ead36e532f9f8c65a67f0f25c543caaf  TRAINING_DATA_INDEX.json
+30cff49b711b43bae2ac4d3d09a04bd592a40103d811d16b7ec9c622b65d3f46  VALIDATION_REPORT.json
+82e482d5724921d65633d0410837af359dc1f08936159f7f09f964ca5977af39  egoexo/captioner_and_phiseg/conda-explicit-linux-64.txt
+4edfa90306638af9d33fba0c5dd5c5d3030c32fffd7067880ffc2d21744b00ba  egoexo/captioner_and_phiseg/environment.snapshot.json
+4a4c601e83800b5787ef1c4276e26010d1a04a24a3b59713357d216d1c34ea51  egoexo/captioner_and_phiseg/import_validation.json
+21553365198912c8a61484a9e0a33bf7fd8f9e16cbfd8e97fa1f1771051ed4e5  egoexo/captioner_and_phiseg/requirements.core.in
+718cb5833bb3287fda8fbf11414767f85dd480a10617de4198badfeeb605d2b0  egoexo/captioner_and_phiseg/requirements.observed-full.lock.txt
+18e81e63b39ce813973c6fba2c5a9cf9034e11be4800e8ae4bec2cbccc0d6856  egoexo/generator/conda-explicit-linux-64.txt
+9e2942c07ed91ac90c6d6cb2e37f166946c238d36169de84f41bd7cf254fcd52  egoexo/generator/environment.snapshot.json
+ace563e6b723b63384b0be0657ac2c776d2ff28fbbf713dffcfd7c82eea8547a  egoexo/generator/import_validation.json
+64f62d8b6a2dfe25a84e75d7f2ff437140335e40288784d970baa140fd1777be  egoexo/generator/requirements.core.in
+408b4d508b243e38efb766583915d4f330984c1fedc9d9342bd686568e8cc6a5  egoexo/generator/requirements.observed-full.lock.txt
+2a26954fb61da336c4f686e83d99feea5ee3459c1649bbe6e3d4b4e5fd3f0f51  robot/data_cpu/environment.snapshot.json
+cbb6a0e2fdb294f9e01f2a2e6d275d9fbc1c273aa8d74a99400744b7a1b2072e  robot/data_cpu/requirements.core.in
+095f462f553c9fc40c564a16deacde429cd9daf5273968879a3b260c913923b1  robot/data_cpu/requirements.install.lock.txt
+bda16cd32423754f9896e1549a6716015f2b238989a5b3bc8eba90e1e4a770e4  robot/data_cpu/requirements.observed-full.lock.txt
+2e9bed0ddcf6459e1ff705761d9e2db07641d5f94bef2551acc98e9f02359618  robot/legacy_sft_grpo/environment.snapshot.json
+909e571070b5aeb0ac239e45ef292acf4b9173041ae3d11808d78d85314c55fd  robot/legacy_sft_grpo/import_validation.json
+73b0475808496e52c2fe1f336519899240e394c3417352f039587a0060811387  robot/legacy_sft_grpo/requirements.core.in
+7adcc53292416d576501febbfac0f5bbf1b03d1f576e6b637b23eca97934237b  robot/legacy_sft_grpo/requirements.observed-full.lock.txt
+22e06189577a0c6347807f1e6209b63cf1b10778332d05266bfa5a68ca085f5c  source-overlays/egoexo/code/core/data/build_testset_from_masks.py
+92cd54185a0d3878cbed558ee412bbd7a4d31f74141e538aa538ae3f252d3aa9  source-overlays/egoexo/code/core/diffusion/caption_clip_probe.py
+e2245db96e1da73a54f390bc943c17c4d5d3d047b1da5996c2f9f73e695ec1b8  source-overlays/egoexo/code/core/diffusion/eval_masked_grpo3800_3wan.py
+52a4b152561e3458424ad3e133d72a63d270f27f43941cab66e6b1eabd127ce5  source-overlays/egoexo/code/core/diffusion/eval_masked_region_metrics.py
+bc208dab05bfc0310fa39be211bb6872995eb9eba22502ef65eff1f9d44be18f  source-overlays/egoexo/code/core/diffusion/eval_opensource_baselines.py
+24a5260ca673f92cee2b2da2598ba2883c07e1bc4c58d23dfef8c3290dbeffd4  source-overlays/egoexo/code/core/diffusion/gen_ego_caption.py
+46701eac89b1d2efbe03efc030e25b7e5d2ee6231ee097f5b5182d452152eaeb  source-overlays/egoexo/code/core/diffusion/infer_caption.py
+0d81d38afc9952fa181b2e77f6761725248649d5c1cf0eefc8761707e0d378e1  source-overlays/egoexo/code/core/diffusion/infer_vace_hoi.py
+c0ec3689edb6a83e60e50fb942b37e5f2cb0afb90e0e02536c23b51a76313204  source-overlays/egoexo/code/core/diffusion/judge_media_vllm.py
+31bd4d6398fcdeb9edfb20cb120a85932e9bbe4e092733af3f570bd64fea68b8  source-overlays/egoexo/code/core/diffusion/per_clip_3metric_3variant.py
+e05cd230252d1ffc64f4381d02e977b6d400e5c8a4adcb09075dcc659c8e0941  source-overlays/egoexo/code/core/diffusion/train_caption.py
+0c4fe88f770d8f4efbd06c076dad262f6edab6d2a31be79d21d666f9cd23f622  source-overlays/egoexo/code/core/diffusion/train_hoi.py
+48195fb1515b26d0bc6440db02d3181bd1ab536143bc44b5c251c7247b2cdc4a  source-overlays/egoexo/code/core/generator_defaults.py
+d7449d7c527adbe7645d1ffe1942e22cc411ad7afe59a40f652ebf5da3f1daf1  source-overlays/egoexo/code/core/model_resources.py
+b8e14989be327292e0a224527dd116ba6ae9b0ff856838ff4c196b6a546cbef2  tools/apply_source_overlay.py
+ef8eb641f58f6808adcaf53f76c139f30f166c6ebd189ba93af7f5fe82929566  tools/capture_environment.py
+67acb0f58a052ce7fd309e50cdaf473a9880bc31a8d1c7f68212d528088d275c  tools/fetch_training_data.py
+2345c6ca6c3cf06a52c57a00e14062f85fbe311b090a4c4f229ec8cca13e42fe  tools/install_environment.py
+4cd8ba59a93ab631419f29b61153ab8024cc7392f81fca1e0470c400144c8826  tools/restore_legacy_mix6.py
+7b4bcca87edce3f261dca556dce23d6c2019ea9a4ba7ff3f470481284c004b40  tools/run_training_recipe.py
+dc71bd874a80e1b41a3c74dcb2029a7cb45f8d14ead191fb42cbb40e1884d3ce  tools/verify_environment.py
+ef9c838c099c7496255bb0c676d4c397ecbec80cee763c7b01b1e46ba461aad5  REPRO_MANIFEST.json

INSTALL_COMMANDS.sh

@@ -0,0 +1,7 @@
+#!/usr/bin/env bash
+set -euo pipefail
+# Plans only: installer refuses existing prefixes and requires --execute for writes.
+python tools/install_environment.py --profile egoexo-generator --python /path/to/python3.11.15 --prefix /path/to/ego-generator
+python tools/install_environment.py --profile egoexo-captioner-phiseg --python /path/to/python3.12.14 --prefix /path/to/ego-captioner
+python tools/install_environment.py --profile robot-sft-grpo --python /path/to/python3.12.13 --prefix /path/to/robot-train
+python tools/install_environment.py --profile robot-data-cpu --python /path/to/python3.12.13 --prefix /path/to/robot-cpu

README.md

@@ -0,0 +1,96 @@
+---
+license: other
+license_name: observed-training-environments-and-reproduction-recipes
+license_link: https://huggingface.co/datasets/diffusion-gif/training-environments/blob/main/README.md
+---
+# Ego-Exo / robot training reproduction environments
+
+This package contains **actual retained environment snapshots**, exact observed
+package-version constraints, role-specific installation requirements, source
+commit locks, portable training command generators and preflight checks. It is
+not a claim that a clean GPU reinstall or an eight-GPU training/metric reproduction
+was completed on this machine.
+
+## Four separate profiles — do not mix them
+
+| Profile | Python | Torch / CUDA | Transformers | Purpose |
+|---|---|---|---|---|
+| egoexo-generator | 3.11.15 | 2.9.1+cu130 / 13.0 | 4.54.0 | Wan/VACE HOI, caption Global/Spatial generation training |
+| egoexo-captioner-phiseg | 3.12.14 | 2.10.0+cu128 / 12.8 | 4.57.6 | Qwen3-VL-8B/ms-swift4.4.2 captioner + PhiSeg |
+| robot-sft-grpo | 3.12.13 | 2.11.0+cu130 / 13.0 | 5.9.0 | Original 27B mix6 SFT / verl0.9.1 + vLLM0.24.0 GRPO |
+| robot-data-cpu | 3.12.13 | no Torch required | not required | data restoration, validation and CPU tests |
+
+The 27B checkpoint nickname is Qwen3.8; its actual architecture is `qwen3_5`.
+It requires the 5.x Transformers stack. Do not use the generator's Transformers
+4.54 environment for that model or mix vLLM0.19/Torch2.10 with vLLM0.24/Torch2.11.
+
+## Create a NEW environment
+
+Choose/create the exact listed Python interpreter first (Conda explicit package
+URLs with MD5 are supplied for the observed Conda profiles). Never install into
+an existing training environment by mistake.
+
+```bash
+python tools/install_environment.py --profile robot-sft-grpo --python /path/to/python3.12.13 --prefix /path/to/new_env
+# inspect the plan; then add --execute
+/path/to/new_env/bin/python tools/verify_environment.py --profile robot-sft-grpo --gpu-basic-check
+```
+
+`requirements.observed-full.lock.txt` captures the retained inventory;
+`requirements.core.in` selects training-role dependencies while using that inventory
+as constraints. The CPU profile additionally provides a resolved hash-locked
+install file. Full GPU clean installation has not been revalidated here. FlashAttention
+is a post-Torch build step (`--flash-attn --cuda-home /path/to/cuda13`), not an
+uncontrolled build-isolation install before Torch exists.
+
+## CUDA and platform pitfalls actually found
+
+- This host has driver580.126.20 / one SM12.0 GPU. Three GPU profiles passed a
+  tiny BF16 CUDA tensor operation; that is not eight-GPU model training.
+- The system toolkit here is nvcc12.8, even when Torch's wheel runtime is13.0.
+  Use a **matching CUDA13 devel toolkit** for torch-cu130 source builds and
+  FlashAttention/GDN kernels. A driver reporting CUDA13 is not an nvcc installation.
+- Historical robot GRPO uses TP4 and eight GPUs; one-GPU runs are not equivalent.
+- The retained Decord0.6.0 wheel has an incorrect internal cp36 tag although its
+  ctypes library imports on Python3.11/3.12. Its original pip-check platform warning
+  is recorded, not hidden. Use the official py3 Linux wheel; validate import/video
+  decoding for your host before training. Do not classify that warning as a
+  proven numerical-model failure or claim all fresh wheel installations passed.
+- The uv-created robot env has no pip module. `uv pip check --python ENV/bin/python`
+  passed all256 installed packages; `python -m pip check` alone is not a valid
+  dependency-health test of that original no-pip environment.
+
+## Get exact source / models / old training data
+
+Use `SOURCE_LOCK.json` and checkout those commits, not moving main branches.
+Some source repositories remain private: owner authorization is required.
+For Ego-Exo use `code/core`, the supplied source overlay where needed, and the
+public weight bundle's sharded-Wan compatibility/attachment tool. DiffSynth Wan
+code is pinned; unrelated local Flux experiments/images are excluded.
+
+Download the already published `TRAINING_DATA_INDEX.json`: it distinguishes
+legacy mix6 training (16,174 train / 3,228 heldout), later joint data, and Any2Ego
+original versus common filtered CSVs. Original training files/NPZ are protected
+in the pinned private vault. Large Ego-Exo original RGB/mask media are NOT fully
+backed up remotely and require original-source permission. No tokens are bundled.
+
+## Portable training recipes (dry-run by default)
+
+```bash
+python tools/run_training_recipe.py --recipe robot-sft --code-root /your/trick_prediction --python /new_env/bin/python --data-root /your/legacy_data --model-root /your/base27b --output /your/new_run --gpus 8
+python tools/run_training_recipe.py --recipe egoexo-spatial --code-root /your/exo-ego/code/core --python /new_env/bin/python --data-root /your/ego_data --model-root /your/any2ego_bundle --diffsynth-root /your/DiffSynth-Studio --output /your/new_run --gpus 8
+# Only add --execute after data/model/version preflight. Outputs must be new.
+```
+
+These are safe, explicit entrypoint recipes rather than evidence of convergence.
+The historical Spatial snapshot used teacher captions; switching to newly generated
+GRPO captions is a different experiment. Old 27B checkpoints were not trained on
+new fixed-EEF joint labels. SFT warm-start does not restore absent optimizer/RNG state.
+
+## Verification boundaries
+
+Recorded: package versions/ABI, CPU dependency hash-lock resolution, source revisions,
+core imports, tiny GPU operations, standalone command dry-runs and syntax checks.
+Not completed: fresh full GPU installation, eight-GPU distributed training, all
+optional native source builds, model metrics, or the numerical identity of an
+unrecorded original training-launch environment. See VALIDATION_REPORT.json.

REPRO_MANIFEST.json

@@ -0,0 +1,231 @@
+{
+  "files": [
+    {
+      "path": "INSTALL_COMMANDS.sh",
+      "bytes": 628,
+      "sha256": "4ef5bb645114b5c72c1f794a6d5e418b9bcf10a08eb2f06f28c1ea52ae8338d0"
+    },
+    {
+      "path": "README.md",
+      "bytes": 5758,
+      "sha256": "41abcf86b8bd9c17bbe4c0b23428fafde870a11a9ecbd90512cafdb5f5e2241e"
+    },
+    {
+      "path": "SOURCE_LOCK.json",
+      "bytes": 1319,
+      "sha256": "bffaadfbc5c59c5c9761e3608d33c20b27b2cdbfd39a98c49d379657ccde197f"
+    },
+    {
+      "path": "SOURCE_OVERLAY_MANIFEST.json",
+      "bytes": 6352,
+      "sha256": "38170c0681f48407253459b44868c2e13bf21bccd84395372ee0df7c1736f98a"
+    },
+    {
+      "path": "TRAINING_DATA_INDEX.json",
+      "bytes": 146175,
+      "sha256": "2569cdcdb30128ffb7d1f178a2669580ead36e532f9f8c65a67f0f25c543caaf"
+    },
+    {
+      "path": "VALIDATION_REPORT.json",
+      "bytes": 697,
+      "sha256": "30cff49b711b43bae2ac4d3d09a04bd592a40103d811d16b7ec9c622b65d3f46"
+    },
+    {
+      "path": "egoexo/captioner_and_phiseg/conda-explicit-linux-64.txt",
+      "bytes": 17604,
+      "sha256": "82e482d5724921d65633d0410837af359dc1f08936159f7f09f964ca5977af39"
+    },
+    {
+      "path": "egoexo/captioner_and_phiseg/environment.snapshot.json",
+      "bytes": 69541,
+      "sha256": "4edfa90306638af9d33fba0c5dd5c5d3030c32fffd7067880ffc2d21744b00ba"
+    },
+    {
+      "path": "egoexo/captioner_and_phiseg/import_validation.json",
+      "bytes": 914,
+      "sha256": "4a4c601e83800b5787ef1c4276e26010d1a04a24a3b59713357d216d1c34ea51"
+    },
+    {
+      "path": "egoexo/captioner_and_phiseg/requirements.core.in",
+      "bytes": 513,
+      "sha256": "21553365198912c8a61484a9e0a33bf7fd8f9e16cbfd8e97fa1f1771051ed4e5"
+    },
+    {
+      "path": "egoexo/captioner_and_phiseg/requirements.observed-full.lock.txt",
+      "bytes": 5259,
+      "sha256": "718cb5833bb3287fda8fbf11414767f85dd480a10617de4198badfeeb605d2b0"
+    },
+    {
+      "path": "egoexo/generator/conda-explicit-linux-64.txt",
+      "bytes": 3547,
+      "sha256": "18e81e63b39ce813973c6fba2c5a9cf9034e11be4800e8ae4bec2cbccc0d6856"
+    },
+    {
+      "path": "egoexo/generator/environment.snapshot.json",
+      "bytes": 54748,
+      "sha256": "9e2942c07ed91ac90c6d6cb2e37f166946c238d36169de84f41bd7cf254fcd52"
+    },
+    {
+      "path": "egoexo/generator/import_validation.json",
+      "bytes": 968,
+      "sha256": "ace563e6b723b63384b0be0657ac2c776d2ff28fbbf713dffcfd7c82eea8547a"
+    },
+    {
+      "path": "egoexo/generator/requirements.core.in",
+      "bytes": 568,
+      "sha256": "64f62d8b6a2dfe25a84e75d7f2ff437140335e40288784d970baa140fd1777be"
+    },
+    {
+      "path": "egoexo/generator/requirements.observed-full.lock.txt",
+      "bytes": 2389,
+      "sha256": "408b4d508b243e38efb766583915d4f330984c1fedc9d9342bd686568e8cc6a5"
+    },
+    {
+      "path": "robot/data_cpu/environment.snapshot.json",
+      "bytes": 14140,
+      "sha256": "2a26954fb61da336c4f686e83d99feea5ee3459c1649bbe6e3d4b4e5fd3f0f51"
+    },
+    {
+      "path": "robot/data_cpu/requirements.core.in",
+      "bytes": 274,
+      "sha256": "cbb6a0e2fdb294f9e01f2a2e6d275d9fbc1c273aa8d74a99400744b7a1b2072e"
+    },
+    {
+      "path": "robot/data_cpu/requirements.install.lock.txt",
+      "bytes": 57372,
+      "sha256": "095f462f553c9fc40c564a16deacde429cd9daf5273968879a3b260c913923b1"
+    },
+    {
+      "path": "robot/data_cpu/requirements.observed-full.lock.txt",
+      "bytes": 535,
+      "sha256": "bda16cd32423754f9896e1549a6716015f2b238989a5b3bc8eba90e1e4a770e4"
+    },
+    {
+      "path": "robot/legacy_sft_grpo/environment.snapshot.json",
+      "bytes": 70170,
+      "sha256": "2e9bed0ddcf6459e1ff705761d9e2db07641d5f94bef2551acc98e9f02359618"
+    },
+    {
+      "path": "robot/legacy_sft_grpo/import_validation.json",
+      "bytes": 798,
+      "sha256": "909e571070b5aeb0ac239e45ef292acf4b9173041ae3d11808d78d85314c55fd"
+    },
+    {
+      "path": "robot/legacy_sft_grpo/requirements.core.in",
+      "bytes": 571,
+      "sha256": "73b0475808496e52c2fe1f336519899240e394c3417352f039587a0060811387"
+    },
+    {
+      "path": "robot/legacy_sft_grpo/requirements.observed-full.lock.txt",
+      "bytes": 5235,
+      "sha256": "7adcc53292416d576501febbfac0f5bbf1b03d1f576e6b637b23eca97934237b"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/data/build_testset_from_masks.py",
+      "bytes": 7585,
+      "sha256": "22e06189577a0c6347807f1e6209b63cf1b10778332d05266bfa5a68ca085f5c"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/caption_clip_probe.py",
+      "bytes": 8804,
+      "sha256": "92cd54185a0d3878cbed558ee412bbd7a4d31f74141e538aa538ae3f252d3aa9"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/eval_masked_grpo3800_3wan.py",
+      "bytes": 4549,
+      "sha256": "e2245db96e1da73a54f390bc943c17c4d5d3d047b1da5996c2f9f73e695ec1b8"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/eval_masked_region_metrics.py",
+      "bytes": 7216,
+      "sha256": "52a4b152561e3458424ad3e133d72a63d270f27f43941cab66e6b1eabd127ce5"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/eval_opensource_baselines.py",
+      "bytes": 4963,
+      "sha256": "bc208dab05bfc0310fa39be211bb6872995eb9eba22502ef65eff1f9d44be18f"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/gen_ego_caption.py",
+      "bytes": 10918,
+      "sha256": "24a5260ca673f92cee2b2da2598ba2883c07e1bc4c58d23dfef8c3290dbeffd4"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/infer_caption.py",
+      "bytes": 9544,
+      "sha256": "46701eac89b1d2efbe03efc030e25b7e5d2ee6231ee097f5b5182d452152eaeb"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/infer_vace_hoi.py",
+      "bytes": 5241,
+      "sha256": "0d81d38afc9952fa181b2e77f6761725248649d5c1cf0eefc8761707e0d378e1"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/judge_media_vllm.py",
+      "bytes": 11573,
+      "sha256": "c0ec3689edb6a83e60e50fb942b37e5f2cb0afb90e0e02536c23b51a76313204"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/per_clip_3metric_3variant.py",
+      "bytes": 5455,
+      "sha256": "31bd4d6398fcdeb9edfb20cb120a85932e9bbe4e092733af3f570bd64fea68b8"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/train_caption.py",
+      "bytes": 13281,
+      "sha256": "e05cd230252d1ffc64f4381d02e977b6d400e5c8a4adcb09075dcc659c8e0941"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/diffusion/train_hoi.py",
+      "bytes": 10241,
+      "sha256": "0c4fe88f770d8f4efbd06c076dad262f6edab6d2a31be79d21d666f9cd23f622"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/generator_defaults.py",
+      "bytes": 1494,
+      "sha256": "48195fb1515b26d0bc6440db02d3181bd1ab536143bc44b5c251c7247b2cdc4a"
+    },
+    {
+      "path": "source-overlays/egoexo/code/core/model_resources.py",
+      "bytes": 2861,
+      "sha256": "d7449d7c527adbe7645d1ffe1942e22cc411ad7afe59a40f652ebf5da3f1daf1"
+    },
+    {
+      "path": "tools/apply_source_overlay.py",
+      "bytes": 1279,
+      "sha256": "b8e14989be327292e0a224527dd116ba6ae9b0ff856838ff4c196b6a546cbef2"
+    },
+    {
+      "path": "tools/capture_environment.py",
+      "bytes": 5686,
+      "sha256": "ef8eb641f58f6808adcaf53f76c139f30f166c6ebd189ba93af7f5fe82929566"
+    },
+    {
+      "path": "tools/fetch_training_data.py",
+      "bytes": 3237,
+      "sha256": "67acb0f58a052ce7fd309e50cdaf473a9880bc31a8d1c7f68212d528088d275c"
+    },
+    {
+      "path": "tools/install_environment.py",
+      "bytes": 3512,
+      "sha256": "2345c6ca6c3cf06a52c57a00e14062f85fbe311b090a4c4f229ec8cca13e42fe"
+    },
+    {
+      "path": "tools/restore_legacy_mix6.py",
+      "bytes": 2642,
+      "sha256": "4cd8ba59a93ab631419f29b61153ab8024cc7392f81fca1e0470c400144c8826"
+    },
+    {
+      "path": "tools/run_training_recipe.py",
+      "bytes": 3385,
+      "sha256": "7b4bcca87edce3f261dca556dce23d6c2019ea9a4ba7ff3f470481284c004b40"
+    },
+    {
+      "path": "tools/verify_environment.py",
+      "bytes": 2225,
+      "sha256": "dc71bd874a80e1b41a3c74dcb2029a7cb45f8d14ead191fb42cbb40e1884d3ce"
+    }
+  ],
+  "original_environment_modified": false,
+  "full_GPU_training_verified": false
+}

SOURCE_LOCK.json

@@ -0,0 +1,11 @@
+{
+  "schema_version": 1,
+  "projects": {
+    "egoexo": {"repository": "https://github.com/coisini-gif/exo-ego-video-converter.git", "revision": "40bcb0fdf01be2205c79bb3e1de786fa6accf4f2", "private_source_access_required": true, "working_directory": "code/core", "post_checkout": "Apply source-overlays supplied here, then the model bundle's attach_model_bundle.py --apply for sharded Wan loading."},
+    "robot": {"repository": "https://github.com/coisini-gif/trick_prediction.git", "revision": "8ae7b90fc2e616aea3683b03e00c2ce26d66c674", "private_source_access_required": true, "working_directory": "."},
+    "diffsynth": {"repository": "https://github.com/modelscope/DiffSynth-Studio.git", "revision": "166e6d2d38764209f66a74dd0fe468226536ad0f", "unrelated_local_flux_edits_excluded": true, "install": "python -m pip install --no-deps -e SOURCE_DIR"}
+  },
+  "models": {"egoexo": "diffusion-gif/any2ego", "robot": "diffusion-gif/trick-prediction"},
+  "data_index": {"repository": "diffusion-gif/training-data-index", "revision": "04182e349e581ce2f5826eb4a7bed2803b2925fc", "path": "TRAINING_DATA_INDEX.json"},
+  "source_clone_policy": "Never use --depth 1 main/latest for reproduction. Clone and checkout the listed immutable commit. Keep third-party licenses. Private code/data access is required where indicated."
+}

SOURCE_OVERLAY_MANIFEST.json

@@ -0,0 +1,105 @@
+{
+  "repository": "https://github.com/coisini-gif/exo-ego-video-converter.git",
+  "revision": "40bcb0fdf01be2205c79bb3e1de786fa6accf4f2",
+  "files": [
+    {
+      "path": "code/core/model_resources.py",
+      "patch_file": "source-overlays/egoexo/code/core/model_resources.py",
+      "expected_git_source_sha256": null,
+      "patched_sha256": "d7449d7c527adbe7645d1ffe1942e22cc411ad7afe59a40f652ebf5da3f1daf1",
+      "reason": "sharded model-path portability overlay"
+    },
+    {
+      "path": "code/core/generator_defaults.py",
+      "patch_file": "source-overlays/egoexo/code/core/generator_defaults.py",
+      "expected_git_source_sha256": "f576a6ceaa29cfb78ba8d783ddadf52bffcfb5f30a8740d55f9e55af1b0a4369",
+      "patched_sha256": "48195fb1515b26d0bc6440db02d3181bd1ab536143bc44b5c251c7247b2cdc4a",
+      "reason": "sharded model-path portability overlay"
+    },
+    {
+      "path": "code/core/diffusion/train_hoi.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/train_hoi.py",
+      "expected_git_source_sha256": "4feb3da54ce6689a03b132ada27b75c46e61724bd474d5c49930aef6c630d92b",
+      "patched_sha256": "0c4fe88f770d8f4efbd06c076dad262f6edab6d2a31be79d21d666f9cd23f622",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/infer_caption.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/infer_caption.py",
+      "expected_git_source_sha256": "377b52f76012189f6ca779c645d2891a8769e160c27a11d09a5591889de93aeb",
+      "patched_sha256": "46701eac89b1d2efbe03efc030e25b7e5d2ee6231ee097f5b5182d452152eaeb",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/judge_media_vllm.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/judge_media_vllm.py",
+      "expected_git_source_sha256": "600967daa9e3aa8b7ab87dc701d8f7236bb73c6f7815404465a2f7ccbdba5393",
+      "patched_sha256": "c0ec3689edb6a83e60e50fb942b37e5f2cb0afb90e0e02536c23b51a76313204",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/train_caption.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/train_caption.py",
+      "expected_git_source_sha256": "626f02d058897e2cf8a8a4cfc979e429c71bf6c337d26f982067fff23aa74e77",
+      "patched_sha256": "e05cd230252d1ffc64f4381d02e977b6d400e5c8a4adcb09075dcc659c8e0941",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/gen_ego_caption.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/gen_ego_caption.py",
+      "expected_git_source_sha256": "a07ac21f6bd3bc00e675c9aad839bc93a303c5590db4089b4ad62930a64ceb90",
+      "patched_sha256": "24a5260ca673f92cee2b2da2598ba2883c07e1bc4c58d23dfef8c3290dbeffd4",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/caption_clip_probe.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/caption_clip_probe.py",
+      "expected_git_source_sha256": "5427cb8728745f5a8b627709613cf20001398898bf625898af5e09bf47f8ec70",
+      "patched_sha256": "92cd54185a0d3878cbed558ee412bbd7a4d31f74141e538aa538ae3f252d3aa9",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/infer_vace_hoi.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/infer_vace_hoi.py",
+      "expected_git_source_sha256": "37e8ecb9f92551d048524226f22f4c0f86086eb69ca6e9a7bf56d69b47c955d4",
+      "patched_sha256": "0d81d38afc9952fa181b2e77f6761725248649d5c1cf0eefc8761707e0d378e1",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/eval_masked_grpo3800_3wan.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/eval_masked_grpo3800_3wan.py",
+      "expected_git_source_sha256": "4d38357518e298b949ce20ccde5ddc42db4eb718b60c2edc9009196809d9d00c",
+      "patched_sha256": "e2245db96e1da73a54f390bc943c17c4d5d3d047b1da5996c2f9f73e695ec1b8",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/eval_masked_region_metrics.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/eval_masked_region_metrics.py",
+      "expected_git_source_sha256": "3d4b7f09f56ad54d65f8c73101e0dc2ce5d883e8787a7b40e0fe321d05a9dde8",
+      "patched_sha256": "52a4b152561e3458424ad3e133d72a63d270f27f43941cab66e6b1eabd127ce5",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/per_clip_3metric_3variant.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/per_clip_3metric_3variant.py",
+      "expected_git_source_sha256": "c12a35fa50aec9b6c8c0181f42af952732fc799b4fea7d275b19226f0841497f",
+      "patched_sha256": "31bd4d6398fcdeb9edfb20cb120a85932e9bbe4e092733af3f570bd64fea68b8",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/diffusion/eval_opensource_baselines.py",
+      "patch_file": "source-overlays/egoexo/code/core/diffusion/eval_opensource_baselines.py",
+      "expected_git_source_sha256": "4ca1b89e27aed338f8ac2a04daedf65d931ec9dd18bd69cf4975e152054f9cf6",
+      "patched_sha256": "bc208dab05bfc0310fa39be211bb6872995eb9eba22502ef65eff1f9d44be18f",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    },
+    {
+      "path": "code/core/data/build_testset_from_masks.py",
+      "patch_file": "source-overlays/egoexo/code/core/data/build_testset_from_masks.py",
+      "expected_git_source_sha256": "0fbc7ff8502de63181a76f1869f5ce23ee817d6a1043e3dba6271b9843602632",
+      "patched_sha256": "22e06189577a0c6347807f1e6209b63cf1b10778332d05266bfa5a68ca085f5c",
+      "reason": "Python3.11-compatible f-string syntax lowering; training expressions preserved"
+    }
+  ],
+  "source_originals_modified": false
+}

VALIDATION_REPORT.json

@@ -0,0 +1,17 @@
+{
+  "status": "environment_package_ready",
+  "version_capture_complete": true,
+  "original_environment_modified": false,
+  "source_revisions_locked": true,
+  "python311_source_compatibility_overlays_syntax_checked": 14,
+  "installer_dry_runs_passed": 4,
+  "CUDA_basic_tensor_checks_passed": 3,
+  "CPU_hash_dependency_lock_resolved": true,
+  "robot_dependency_check_using_uv": "all256 installed packages compatible",
+  "egoexo_pip_check_warning": "original Decord wheel internal cp36 tag; import passed",
+  "full_GPU_environment_clean_reinstall_performed": false,
+  "8GPU_training_performed": false,
+  "paper_metrics_reproduced": false,
+  "new_teacher_API_calls": 0,
+  "secrets_published": false
+}

egoexo/captioner_and_phiseg/conda-explicit-linux-64.txt

@@ -0,0 +1,152 @@
+@EXPLICIT
+https://conda.anaconda.org/conda-forge/linux-64/_openmp_mutex-4.5-20_gnu.conda#a9f577daf3de00bca7c3c76c0ecbd1de
+https://conda.anaconda.org/conda-forge/linux-64/alsa-lib-1.2.16.1-h7cc23a3_1.conda#7094e0d8d14de0eff6d83f0d2f1f661e
+https://conda.anaconda.org/conda-forge/linux-64/aom-3.14.1-pl5321h57e6904_2.conda#5fabcad67aa128f07f269066ee7fdde6
+https://conda.anaconda.org/conda-forge/linux-64/bzip2-1.0.8-hda65f42_10.conda#e675fabcf81499adc7edf58124fb1e01
+https://conda.anaconda.org/conda-forge/noarch/ca-certificates-2026.7.22-hbd8a1cb_0.conda#0f51e2391ade309db462a55611263e9c
+https://conda.anaconda.org/conda-forge/linux-64/cairo-1.18.4-h3c89d7e_3.conda#7b870cffbaa8430290e567a0facd8dca
+https://conda.anaconda.org/conda-forge/linux-64/dav1d-1.2.1-hd590300_0.conda#418c6ca5929a611cbd69204907a83995
+https://conda.anaconda.org/conda-forge/linux-64/dbus-1.16.2-he8c428d_2.conda#a6eb37b51be1a9a654dc3a8aca039b1a
+https://conda.anaconda.org/conda-forge/linux-64/ffmpeg-9.0.1-gpl_h5342cc5_901.conda#08760cc55c9985ce0bd7b2e92037e567
+https://conda.anaconda.org/conda-forge/noarch/font-ttf-dejavu-sans-mono-2.37-hab24e00_0.tar.bz2#0c96522c6bdaed4b1566d11387caaf45
+https://conda.anaconda.org/conda-forge/noarch/font-ttf-inconsolata-3.000-h77eed37_0.tar.bz2#34893075a5c9e55cdafac56607368fc6
+https://conda.anaconda.org/conda-forge/noarch/font-ttf-source-code-pro-2.038-h77eed37_0.tar.bz2#4d59c254e01d9cde7957100457e2d5fb
+https://conda.anaconda.org/conda-forge/noarch/font-ttf-ubuntu-0.83-h77eed37_3.conda#49023d73832ef61042f6a237cb2687e7
+https://conda.anaconda.org/conda-forge/linux-64/fontconfig-2.18.3-h4db4eae_1.conda#922776b528a470ab5afa81fd42abfa1d
+https://conda.anaconda.org/conda-forge/noarch/fonts-conda-ecosystem-1-0.tar.bz2#fee5683a3f04bd15cbd8318b096a27ab
+https://conda.anaconda.org/conda-forge/noarch/fonts-conda-forge-1-hc364b38_1.conda#a7970cd949a077b7cb9696379d338681
+https://conda.anaconda.org/conda-forge/linux-64/freetype-2.14.3-ha770c72_2.conda#2d0ea23b23603e07ca47f23f949f67b6
+https://conda.anaconda.org/conda-forge/linux-64/fribidi-1.0.16-hb03c661_1.conda#1cd10eda5692519d01bb20e086e214c9
+https://conda.anaconda.org/conda-forge/linux-64/gdk-pixbuf-2.44.8-h68053f1_1.conda#11adf7fc09b1146a6d91deb65d374456
+https://conda.anaconda.org/conda-forge/linux-64/glslang-16.5.0-h980caa0_2.conda#f8cf8d6c2f5a98e7263d461c8514cb8a
+https://conda.anaconda.org/conda-forge/linux-64/gmp-6.3.0-hfd2156b_3.conda#576e32739f323438bf69ab006c21be7b
+https://conda.anaconda.org/conda-forge/linux-64/graphite2-1.3.15-h54a6638_1.conda#f9fe2984587fa8235a6af6004760cd18
+https://conda.anaconda.org/conda-forge/linux-64/harfbuzz-14.4.0-ha770c72_0.conda#c2d4a4fe3216b26ef30e91d917c1b718
+https://conda.anaconda.org/conda-forge/linux-64/icu-78.3-py310h44b86e0_2.conda#72a381cbad04f24b1c2a43ef707f45b4
+https://conda.anaconda.org/conda-forge/linux-64/intel-gmmlib-22.10.0-hb700be7_0.conda#10909406c1b0e4b57f9f4f0eb0999af8
+https://conda.anaconda.org/conda-forge/linux-64/intel-media-driver-26.1.6-hecca717_0.conda#aaf7c3db8c7c4533deb5449d3ba1c51f
+https://conda.anaconda.org/conda-forge/linux-64/lame-4.0-h770b6ad_1.conda#5aecb65b6ecfee6f878e6789a8a779de
+https://conda.anaconda.org/conda-forge/linux-64/lcms2-2.19.1-h9073bf1_2.conda#b68c7304d2edb0f1a5bdfb875094d603
+https://conda.anaconda.org/conda-forge/linux-64/ld_impl_linux-64-2.46.1-default_hbd61a6d_102.conda#449500f2c089da11c40f5c21312e3e07
+https://conda.anaconda.org/conda-forge/linux-64/lerc-4.2.0-hdb68285_0.conda#fb9d356b1a57d6d54768be7ebd5fce09
+https://conda.anaconda.org/conda-forge/linux-64/level-zero-1.29.0-hb700be7_0.conda#f3c3bc77c96af553f761af0e78bc8d9d
+https://conda.anaconda.org/conda-forge/linux-64/libabseil-20260526.0-cxx17_h0dc7533_2.conda#6983bfe8e09992014b9cf393769c7a84
+https://conda.anaconda.org/conda-forge/linux-64/libass-0.17.5-h434b012_0.conda#c1cb4d6e8a6e3f724740dee5346fc8b4
+https://conda.anaconda.org/conda-forge/linux-64/libbrotlicommon-1.2.0-h39a168f_3.conda#7a2499a177753582fb7ae7e9dc4a908a
+https://conda.anaconda.org/conda-forge/linux-64/libbrotlidec-1.2.0-ha411449_3.conda#6ab3315dc56618d652c1da42a648a129
+https://conda.anaconda.org/conda-forge/linux-64/libbrotlienc-1.2.0-h018ffa1_3.conda#2ac965638d4c6b2b38383bb1aebaf543
+https://conda.anaconda.org/conda-forge/linux-64/libcap-2.78-h084b8d7_1.conda#5db514adf5f843126ff846d1510f22a4
+https://conda.anaconda.org/conda-forge/linux-64/libdeflate-1.25-hd45a770_1.conda#40f9b31aa9cf007789867df0decd0492
+https://conda.anaconda.org/conda-forge/linux-64/libdovi-3.4.0-ha23c83e_0.conda#4377d220f09344452b227d699cacce4f
+https://conda.anaconda.org/conda-forge/linux-64/libdrm-2.4.129-h7cc23a3_0.conda#64cc91512b6278c315349dc13c53f680
+https://conda.anaconda.org/conda-forge/linux-64/libegl-1.7.0-ha4b6fd6_5.conda#3a1b41c4591a0a1734382af3e2b8bc08
+https://conda.anaconda.org/conda-forge/linux-64/libexpat-2.8.1-hecca717_1.conda#b24d3c612f71e7aa74158d92106318b2
+https://conda.anaconda.org/conda-forge/linux-64/libffi-3.7.0-h81df57d_1.conda#6525a0b06aa4fd390795f0740636a9dd
+https://conda.anaconda.org/conda-forge/linux-64/libflac-1.5.0-he200343_1.conda#47595b9d53054907a00d95e4d47af1d6
+https://conda.anaconda.org/conda-forge/linux-64/libfreetype-2.14.3-ha770c72_2.conda#f8054e759d0ddddaf34a5c8fedc900a1
+https://conda.anaconda.org/conda-forge/linux-64/libfreetype6-2.14.3-h5e6c136_2.conda#b72a266a9317036fd0464cb98b027cd0
+https://conda.anaconda.org/conda-forge/linux-64/libgcc-16.2.0-ha9f2e26_4.conda#cba14d01083fc62ffd32c24d7d390633
+https://conda.anaconda.org/conda-forge/linux-64/libgcc-ng-16.2.0-h69a702a_4.conda#b3e52878163a841f6fb951989cc0b217
+https://conda.anaconda.org/conda-forge/linux-64/libgl-1.7.0-ha4b6fd6_5.conda#81141db127a106eb5a91df69a29c9918
+https://conda.anaconda.org/conda-forge/linux-64/libglib-2.88.3-he503a2a_2.conda#533d342dedadf134c67760ce6fc5b748
+https://conda.anaconda.org/conda-forge/linux-64/libglvnd-1.7.0-ha4b6fd6_5.conda#cfcf11edcfd4acf0094771a9c3b8587f
+https://conda.anaconda.org/conda-forge/linux-64/libglx-1.7.0-ha4b6fd6_5.conda#614e10aae180f87b84f0d1ca8ceb7d9e
+https://conda.anaconda.org/conda-forge/linux-64/libgomp-16.2.0-he0feb66_4.conda#89d2c1231f47bd818f5d624b9411459d
+https://conda.anaconda.org/conda-forge/linux-64/libharfbuzz-14.4.0-h23af247_0.conda#4463210c2a16ff5d3bf7f502af23f1f4
+https://conda.anaconda.org/conda-forge/linux-64/libharfbuzz-devel-14.4.0-h23af247_0.conda#26e37e05324d7bb723350d50b33057fc
+https://conda.anaconda.org/conda-forge/linux-64/libhwloc-2.13.0-default_he001693_1000.conda#c197985b58bc813d26b42881f0021c82
+https://conda.anaconda.org/conda-forge/linux-64/libhwy-1.4.0-h57c4cff_1.conda#d58bfbaaf51ed85771eae92c2e7f5816
+https://conda.anaconda.org/conda-forge/linux-64/libiconv-1.18-h0cb94f2_3.conda#f92233bf33e24a25668bb2119e2c51f9
+https://conda.anaconda.org/conda-forge/linux-64/libjpeg-turbo-3.2.0-hb03c661_1.conda#898d1c9793eaa52efc4727bd84d2e39a
+https://conda.anaconda.org/conda-forge/linux-64/libjxl-0.12.0-heb2dce7_2.conda#485788d339785bac87fe86315b4a0627
+https://conda.anaconda.org/conda-forge/linux-64/liblzma-5.8.3-hb03c661_1.conda#1390b7c5ac0b1d8e447bc5efa6d3c8c2
+https://conda.anaconda.org/conda-forge/linux-64/libnsl-2.0.1-hb9d3cd8_1.conda#d864d34357c3b65a4b731f78c0801dc4
+https://conda.anaconda.org/conda-forge/linux-64/libogg-1.3.5-hd0c01bc_1.conda#68e52064ed3897463c0e958ab5c8f91b
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-2026.3.1-hf7e0547_0.conda#8ed292d0e780b85a5d20a5a0fb28a059
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-auto-batch-plugin-2026.3.1-h202117f_0.conda#78221ffaac5f45905b32a4882f85c42d
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-auto-plugin-2026.3.1-h202117f_0.conda#b92521d029afe25bbc59434ddbe107f5
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-hetero-plugin-2026.3.1-hc0229a9_0.conda#76996545f632c1f8f1b6ba269b7efa78
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-intel-cpu-plugin-2026.3.1-hf7e0547_0.conda#d8dd693c1a0ea3ec3e684831511b8d81
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-intel-gpu-plugin-2026.3.1-hf7e0547_0.conda#c0833352f4c66f883f06f3895ac6e8f4
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-intel-npu-plugin-2026.3.1-hf7e0547_0.conda#8e2395729e3da9ae0333f6266b2e65c8
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-ir-frontend-2026.3.1-hc0229a9_0.conda#84751393936b9d374e5d5f0e424bf162
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-onnx-frontend-2026.3.1-h09f0106_0.conda#c0454189eadbc398ac8d758e1df348e1
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-paddle-frontend-2026.3.1-h09f0106_0.conda#c6217da00ade21420bf17b7763ad73ae
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-pytorch-frontend-2026.3.1-ha623fbf_0.conda#1e50f2163f66ea9ccd07dc95856bad07
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-tensorflow-frontend-2026.3.1-h9a43043_0.conda#c9352191bb377298f1d4eefa31529a6a
+https://conda.anaconda.org/conda-forge/linux-64/libopenvino-tensorflow-lite-frontend-2026.3.1-ha623fbf_0.conda#3d137fc391f46e636ca606b442a46e11
+https://conda.anaconda.org/conda-forge/linux-64/libopus-1.6.1-hebe6cf0_1.conda#8a1d977c3d77666406a40c0ab648ff52
+https://conda.anaconda.org/conda-forge/linux-64/libpciaccess-0.19-hb03c661_1.conda#35fa2b34bbced424e6976d30f5fde576
+https://conda.anaconda.org/conda-forge/linux-64/libplacebo-7.360.1-hc50b9dd_1.conda#4a750e2ae0d52d003bb1e3421581585e
+https://conda.anaconda.org/conda-forge/linux-64/libpng-1.6.58-h922cc85_1.conda#fcf71c8d979148873f6f8ad4cfc73d86
+https://conda.anaconda.org/conda-forge/linux-64/libprotobuf-7.35.1-h622638d_3.conda#fd307b61cceb36fdca38e1718bdb2893
+https://conda.anaconda.org/conda-forge/linux-64/librsvg-2.62.3-h4c96295_0.conda#492c8d9b1c564c2e948b6cb4ba0f8261
+https://conda.anaconda.org/conda-forge/linux-64/libsndfile-1.2.2-hbc6d301_3.conda#a2bc10137c845d9c45067f3c53aad6e5
+https://conda.anaconda.org/conda-forge/linux-64/libsqlite-3.53.4-h13e7031_1.conda#e72bbec309c2b0f37823ee7d4fabfcd3
+https://conda.anaconda.org/conda-forge/linux-64/libstdcxx-16.2.0-h934c35e_4.conda#2f2ef0d96de5bdd8c1270ff22fdf9352
+https://conda.anaconda.org/conda-forge/linux-64/libstdcxx-ng-16.2.0-hdf11a46_4.conda#de0dceacf3e33c5fc88e167885dc8274
+https://conda.anaconda.org/conda-forge/linux-64/libsystemd0-261.2-h6f4a2f1_0.conda#592a8bd3952e6da248e3a40fdd612ed1
+https://conda.anaconda.org/conda-forge/linux-64/libtiff-4.7.2-hcc2c06a_1.conda#0907d876f460ebc941ae590338b6102f
+https://conda.anaconda.org/conda-forge/linux-64/libudev1-261.2-h6f4a2f1_0.conda#6f0fe22fab3d7e1fa244c6b59c710530
+https://conda.anaconda.org/conda-forge/linux-64/libunwind-1.8.3-h65a8314_0.conda#e179a69edd30d75c0144d7a380b88f28
+https://conda.anaconda.org/conda-forge/linux-64/liburing-2.14-hb700be7_0.conda#56f65185b520e016d29d01657ac02c0d
+https://conda.anaconda.org/conda-forge/linux-64/libusb-1.0.29-h73b1eb8_0.conda#d17e3fb595a9f24fa9e149239a33475d
+https://conda.anaconda.org/conda-forge/linux-64/libuuid-2.42.2-h5347b49_0.conda#01bb81d12c957de066ea7362007df642
+https://conda.anaconda.org/conda-forge/linux-64/libva-2.24.1-hb83e432_1.conda#c3e6df2790fff34ba708722052f02c0e
+https://conda.anaconda.org/conda-forge/linux-64/libvorbis-1.3.7-h54a6638_2.conda#b4ecbefe517ed0157c37f8182768271c
+https://conda.anaconda.org/conda-forge/linux-64/libvpl-2.16.0-h54a6638_0.conda#9f6b0090c3902b2c763a16f7dace7b6e
+https://conda.anaconda.org/conda-forge/linux-64/libvpx-1.17.0-hd2095e1_0.conda#8762883a9fae4c96eabcdfcec145149f
+https://conda.anaconda.org/conda-forge/linux-64/libvulkan-loader-1.4.357.0-h0e34353_2.conda#c5f5f159b66332152197893427071695
+https://conda.anaconda.org/conda-forge/linux-64/libwebp-base-1.6.0-hd42ef1d_1.conda#9332b53d0ea93c5d39e33be03a0c611a
+https://conda.anaconda.org/conda-forge/linux-64/libxcb-1.17.0-hb83e432_2.conda#61f0ffba9161cbb3a77722869b21e4bb
+https://conda.anaconda.org/conda-forge/linux-64/libxcrypt-4.4.38-h280c20c_0.conda#f7a7ff5a6ab331e037abd34f379a631d
+https://conda.anaconda.org/conda-forge/linux-64/libxkbcommon-1.13.2-h51789e4_1.conda#f888787e0eab7a8076a67d197e155ffc
+https://conda.anaconda.org/conda-forge/linux-64/libxml2-16-2.15.3-hca6bf5a_1.conda#20e4d67ec908f0405124f11612c48305
+https://conda.anaconda.org/conda-forge/linux-64/libxml2-2.15.3-h49c6c72_1.conda#2d34cdb31014cbc8f1e9384c89ede0a5
+https://conda.anaconda.org/conda-forge/linux-64/libzlib-1.3.2-h25fd6f3_3.conda#0de0122d9570a8ab637c6b73db268389
+https://conda.anaconda.org/conda-forge/linux-64/mpg123-1.33.7-h877a99e_1.conda#e7cc06dba8d5fb83d9ae033225ffa458
+https://conda.anaconda.org/conda-forge/linux-64/ncurses-6.6-hdb14827_1.conda#ee6c0cd80a60961a1f48aa3e0b91f986
+https://conda.anaconda.org/conda-forge/linux-64/ocl-icd-2.3.4-hb03c661_1.conda#c2871ba95727fd1382c05db66048b64c
+https://conda.anaconda.org/conda-forge/linux-64/opencl-headers-2025.06.13-hecca717_0.conda#c930c8052d780caa41216af7de472226
+https://conda.anaconda.org/conda-forge/linux-64/openh264-2.6.0-h8c49934_2.conda#2f0b27ebfcbc15298dd99205e91288ba
+https://conda.anaconda.org/conda-forge/linux-64/openssl-3.6.4-h781a0a9_0.conda#16e3034a330cc625f2c685edec60cf4e
+https://conda.anaconda.org/conda-forge/noarch/packaging-26.3-pyhc364b38_0.conda#936687ed80f295a1f5dbcf8bd34c252c
+https://conda.anaconda.org/conda-forge/linux-64/pango-1.58.2-hda50119_0.conda#6a2822aaf9a34ac3708904a47ff3dd7e
+https://conda.anaconda.org/conda-forge/linux-64/pcre2-10.47-h8b3dc9c_1.conda#06f6113cf1ff4a54b65f87ead132a5f1
+https://conda.anaconda.org/conda-forge/noarch/pip-26.2.1-pyh8b19718_0.conda#77b6cf3b0689d9fc68561df0db9b856d
+https://conda.anaconda.org/conda-forge/linux-64/pixman-0.46.4-h54a6638_3.conda#0ee5bb30034b081a1386c1e2c98ab0a7
+https://conda.anaconda.org/conda-forge/linux-64/pthread-stubs-0.4-hb03c661_1003.conda#df2c27f36bdb0dde779f55b5df76a352
+https://conda.anaconda.org/conda-forge/linux-64/pugixml-1.15-h3f63f65_0.conda#b11a4c6bf6f6f44e5e143f759ffa2087
+https://conda.anaconda.org/conda-forge/linux-64/pulseaudio-client-17.0-h9a6aba3_3.conda#b8ea447fdf62e3597cb8d2fae4eb1a90
+https://conda.anaconda.org/conda-forge/linux-64/python-3.12.14-h8ab3286_0_cpython.conda#e9dcdd23a1c68738eb3257d23d5f7285
+https://conda.anaconda.org/conda-forge/linux-64/readline-8.3-hd6e31c0_1.conda#69c01c781e7c8190bbdaf79e3848c5ca
+https://conda.anaconda.org/conda-forge/linux-64/sdl2-2.32.56-h54a6638_0.conda#cdd138897d94dc07d99afe7113a07bec
+https://conda.anaconda.org/conda-forge/linux-64/sdl3-3.4.14-hdeec2a5_0.conda#1ef39a7b42a06e262723fa7937210639
+https://conda.anaconda.org/conda-forge/noarch/setuptools-84.0.0-pyh332efcf_0.conda#62ac906f1cd582c6c264c95625cb9d6f
+https://conda.anaconda.org/conda-forge/linux-64/shaderc-2026.3-hcebf71c_1.conda#8f3979185375f21da886fb6cc2630005
+https://conda.anaconda.org/conda-forge/linux-64/snappy-1.2.2-h03e3b7b_1.conda#98b6c9dc80eb87b2519b97bcf7e578dd
+https://conda.anaconda.org/conda-forge/linux-64/spirv-tools-2026.3-h7148c6a_1.conda#ac9adda1573683c31a8c58850e911273
+https://conda.anaconda.org/conda-forge/linux-64/svt-av1-4.2.0-hd2095e1_1.conda#3bd5f5f3f6f6431b9f19b35ad4a8ed21
+https://conda.anaconda.org/conda-forge/linux-64/tbb-2023.0.0-hab88423_2.conda#7073b15f9364ebc118998601ac6ca6a6
+https://conda.anaconda.org/conda-forge/linux-64/tk-8.6.13-noxft_h1df4ec4_4.conda#676a2f9b1c4fcf57b70aa5c65f6c60d0
+https://conda.anaconda.org/conda-forge/noarch/tzdata-2026c-h151e31d_0.conda#fcb489df604d100968b737f2cb6076c6
+https://conda.anaconda.org/conda-forge/linux-64/wayland-1.26.0-hc1c935e_2.conda#88a47e22b53e50865b1cabe2eb648352
+https://conda.anaconda.org/conda-forge/noarch/wayland-protocols-1.49-hd8ed1ab_0.conda#0839a3421140d4a9ba93fb988698fc00
+https://conda.anaconda.org/conda-forge/noarch/wheel-0.48.0-pyhd8ed1ab_0.conda#9a2124517bfd5d792e0f7b86eefdfeb8
+https://conda.anaconda.org/conda-forge/linux-64/x264-1!164.3095-h166bdaf_2.tar.bz2#6c99772d483f566d59e25037fea2c4b1
+https://conda.anaconda.org/conda-forge/linux-64/x265-3.5-h924138e_3.tar.bz2#e7f6ed84d4623d52ee581325c1587a6b
+https://conda.anaconda.org/conda-forge/linux-64/xkeyboard-config-2.48-h280c20c_0.conda#b233b41be0bf210989d57160ed39b394
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libice-1.1.2-h280c20c_0.conda#85c9442aec283b4e464fa9ecc484a2f3
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libsm-1.2.6-h0d788c3_1.conda#aa7459ed9ad086ba11dda843d764b33d
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libx11-1.8.13-he1eb515_1.conda#8c282bbe4808a3cc80a5c98e9aec1cfc
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxau-1.0.12-hb03c661_2.conda#f06ef439c280a5f90b8bf62355008dbc
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxcursor-1.2.3-hb9d3cd8_0.conda#2ccd714aa2242315acaf0a67faea780b
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxdmcp-1.1.5-hb03c661_2.conda#2e66c929f3d879708335b6ea4557c838
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxext-1.3.7-h7cc23a3_1.conda#e5b6b28536b81b3f4cb20db4668a4642
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxfixes-6.0.2-h7cc23a3_1.conda#09132e874fe0e1f5e4492b0b6b904b6e
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxi-1.8.3-h7cc23a3_1.conda#a1412b2b1184dacda45f8476fef2cc25
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxrandr-1.5.5-h7cc23a3_1.conda#798a8c9d171859a022e1cb89e7d0eb10
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxrender-0.9.12-hb03c661_1.conda#e470d224a7a5be1b1d021bded7abb536
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxscrnsaver-1.2.4-hb9d3cd8_0.conda#303f7a0e9e0cd7d250bb6b952cecda90
+https://conda.anaconda.org/conda-forge/linux-64/xorg-libxtst-1.2.5-h7cc23a3_4.conda#752a5ac9322e0fbbc24146c1ce3ae44e
+https://conda.anaconda.org/conda-forge/linux-64/zstd-1.5.7-hb78ec9c_7.conda#aa459086047c0e5e27023ab19f8cb86a

egoexo/captioner_and_phiseg/environment.snapshot.json

@@ -0,0 +1,1725 @@
+{
+  "profile": "egoexo_captioner_and_phiseg",
+  "capture_scope": "observed current retained runtime; not proof of original launch-time environment",
+  "python": "3.12.14",
+  "implementation": "CPython",
+  "SOABI": "cpython-312-x86_64-linux-gnu",
+  "machine": "x86_64",
+  "platform": "Linux",
+  "libc": [
+    "glibc",
+    "2.35"
+  ],
+  "packages": {
+    "absl-py": "2.5.0",
+    "accelerate": "1.14.0",
+    "addict": "2.4.0",
+    "agent-detector": "1.1.0",
+    "aiofiles": "24.1.0",
+    "aiohappyeyeballs": "2.7.1",
+    "aiohttp": "3.14.3",
+    "aiosignal": "1.4.0",
+    "aliyun-python-sdk-core": "2.16.1",
+    "aliyun-python-sdk-kms": "2.16.5",
+    "annotated-doc": "0.0.5",
+    "annotated-types": "0.8.0",
+    "anthropic": "1.2.0",
+    "anyio": "4.14.2",
+    "apache-tvm-ffi": "0.1.13.post3",
+    "astor": "0.8.1",
+    "attrdict": "2.0.1",
+    "attrs": "26.1.0",
+    "audioread": "3.1.0",
+    "av": "18.1.0",
+    "binpacking": "2.0.1",
+    "blake3": "1.0.9",
+    "brotli": "1.2.0",
+    "cachetools": "7.1.8",
+    "cbor2": "6.1.4",
+    "certifi": "2026.7.22",
+    "cffi": "2.1.1",
+    "charset-normalizer": "3.5.1",
+    "click": "8.5.0",
+    "cloudpickle": "3.1.2",
+    "compressed-tensors": "0.14.0.1",
+    "contourpy": "1.3.3",
+    "cpm-kernels": "1.0.11",
+    "crcmod": "1.7",
+    "cryptography": "50.0.1",
+    "cuda-bindings": "12.9.4",
+    "cuda-pathfinder": "1.8.0",
+    "cuda-python": "12.9.4",
+    "cycler": "0.12.1",
+    "dacite": "1.9.2",
+    "datasets": "4.8.4",
+    "decorator": "5.3.1",
+    "decord": "0.6.0",
+    "deepspeed": "0.19.1",
+    "defusedxml": "0.7.1",
+    "depyf": "0.20.0",
+    "detect-installer": "0.1.0",
+    "dill": "0.4.1",
+    "diskcache": "5.6.3",
+    "dnspython": "2.8.0",
+    "docstring-parser": "0.18.0",
+    "einops": "0.8.2",
+    "email-validator": "2.3.0",
+    "fastapi": "0.141.1",
+    "fastapi-cli": "0.0.32",
+    "fastapi-cloud-cli": "0.24.0",
+    "fastar": "0.12.0",
+    "ffmpy": "1.0.0",
+    "filelock": "3.32.5",
+    "flashinfer-cubin": "0.6.6",
+    "flashinfer-python": "0.6.6",
+    "fonttools": "4.64.0",
+    "frozenlist": "1.8.0",
+    "fsspec": "2026.2.0",
+    "ftfy": "6.3.1",
+    "gguf": "0.19.0",
+    "googleapis-common-protos": "1.75.2",
+    "gradio": "5.50.0",
+    "gradio-client": "1.14.0",
+    "groovy": "0.1.2",
+    "grpcio": "1.83.1",
+    "h11": "0.16.0",
+    "hf-xet": "1.6.0",
+    "hjson": "3.1.0",
+    "httpcore": "1.0.9",
+    "httpcore2": "2.12.0",
+    "httptools": "0.8.0",
+    "httpx": "0.28.1",
+    "httpx2": "2.12.0",
+    "huggingface-hub": "0.36.2",
+    "idna": "3.19",
+    "ijson": "3.5.1",
+    "imageio": "2.37.4",
+    "imageio-ffmpeg": "0.6.0",
+    "importlib-metadata": "9.0.1",
+    "interegular": "0.3.3",
+    "jinja2": "3.1.6",
+    "jiter": "0.16.0",
+    "jmespath": "1.1.0",
+    "joblib": "1.6.0",
+    "json-repair": "0.63.4",
+    "jsonschema": "4.26.0",
+    "jsonschema-specifications": "2025.9.1",
+    "kiwisolver": "1.5.1",
+    "lark": "1.2.2",
+    "lazy-loader": "0.5",
+    "librosa": "0.10.2.post1",
+    "llguidance": "1.3.0",
+    "llvmlite": "0.44.0",
+    "lm-format-enforcer": "0.11.3",
+    "loguru": "0.7.3",
+    "markdown": "3.10.3",
+    "markdown-it-py": "4.2.0",
+    "markupsafe": "3.0.3",
+    "matplotlib": "3.11.1",
+    "mcp": "2.1.1",
+    "mcp-types": "2.1.1",
+    "mdurl": "0.1.2",
+    "mistral-common": "1.11.7",
+    "model-hosting-container-standards": "0.1.16",
+    "modelscope": "1.39.1",
+    "modelscope-hub": "0.3.1",
+    "mpmath": "1.3.0",
+    "ms-swift": "4.4.2",
+    "msgpack": "1.2.2",
+    "msgspec": "0.21.1",
+    "multidict": "6.7.1",
+    "multiprocess": "0.70.19",
+    "narwhals": "2.25.0",
+    "networkx": "3.6.1",
+    "ninja": "1.13.2",
+    "nltk": "3.10.3",
+    "numba": "0.61.2",
+    "numpy": "2.2.6",
+    "nvidia-cublas-cu12": "12.8.4.1",
+    "nvidia-cuda-cupti-cu12": "12.8.90",
+    "nvidia-cuda-nvdisasm": "13.3.73",
+    "nvidia-cuda-nvrtc-cu12": "12.8.93",
+    "nvidia-cuda-runtime-cu12": "12.8.90",
+    "nvidia-cudnn-cu12": "9.10.2.21",
+    "nvidia-cudnn-frontend": "1.18.0",
+    "nvidia-cufft-cu12": "11.3.3.83",
+    "nvidia-cufile-cu12": "1.13.1.3",
+    "nvidia-curand-cu12": "10.3.9.90",
+    "nvidia-cusolver-cu12": "11.7.3.90",
+    "nvidia-cusparse-cu12": "12.5.8.93",
+    "nvidia-cusparselt-cu12": "0.7.1",
+    "nvidia-cutlass-dsl": "4.8.0.dev0",
+    "nvidia-cutlass-dsl-libs-base": "4.8.0.dev0",
+    "nvidia-cutlass-dsl-libs-core": "4.8.0.dev0",
+    "nvidia-cutlass-dsl-libs-cu12": "4.8.0.dev0",
+    "nvidia-ml-py": "13.610.43",
+    "nvidia-nccl-cu12": "2.27.5",
+    "nvidia-nvjitlink-cu12": "12.8.93",
+    "nvidia-nvshmem-cu12": "3.4.5",
+    "nvidia-nvtx-cu12": "12.8.90",
+    "openai": "3.6.0",
+    "openai-harmony": "0.0.8",
+    "opencv-python-headless": "5.0.0.93",
+    "opentelemetry-api": "1.44.0",
+    "opentelemetry-exporter-otlp": "1.44.0",
+    "opentelemetry-exporter-otlp-proto-common": "1.44.0",
+    "opentelemetry-exporter-otlp-proto-grpc": "1.44.0",
+    "opentelemetry-exporter-otlp-proto-http": "1.44.0",
+    "opentelemetry-proto": "1.44.0",
+    "opentelemetry-sdk": "1.44.0",
+    "opentelemetry-semantic-conventions": "0.65b0",
+    "opentelemetry-semantic-conventions-ai": "0.5.1",
+    "orjson": "3.12.0",
+    "oss2": "2.19.1",
+    "outlines-core": "0.2.11",
+    "packaging": "26.3",
+    "pandas": "2.3.3",
+    "partial-json-parser": "0.2.1.1.post7",
+    "peft": "0.19.1",
+    "pillow": "11.3.0",
+    "pip": "26.2.1",
+    "platformdirs": "4.11.7",
+    "pooch": "1.9.0",
+    "prometheus-client": "0.26.0",
+    "prometheus-fastapi-instrumentator": "7.1.0",
+    "propcache": "0.5.2",
+    "protobuf": "7.36.1",
+    "psutil": "7.2.2",
+    "py-cpuinfo": "9.0.0",
+    "pyarrow": "25.0.1",
+    "pybase64": "1.5.0",
+    "pycountry": "26.2.16",
+    "pycparser": "3.0",
+    "pycryptodome": "3.23.0",
+    "pydantic": "2.12.3",
+    "pydantic-core": "2.41.4",
+    "pydantic-extra-types": "2.11.1",
+    "pydantic-settings": "2.15.0",
+    "pydub": "0.25.1",
+    "pygments": "2.21.0",
+    "pyjwt": "2.13.0",
+    "pyparsing": "3.3.2",
+    "python-dateutil": "2.9.0.post0",
+    "python-dotenv": "1.2.3",
+    "python-json-logger": "4.2.0",
+    "python-multipart": "0.0.32",
+    "pytz": "2026.3.post1",
+    "pyyaml": "6.0.3",
+    "pyzmq": "27.2.0",
+    "quack-kernels": "0.5.0",
+    "qwen-omni-utils": "0.0.9",
+    "qwen-vl-utils": "0.0.14",
+    "referencing": "0.37.0",
+    "regex": "2026.9.3",
+    "requests": "2.34.2",
+    "rich": "15.0.0",
+    "rich-toolkit": "0.20.3",
+    "rignore": "0.8.1",
+    "rouge": "1.0.1",
+    "rpds-py": "2026.6.3",
+    "ruff": "0.16.5",
+    "safehttpx": "0.1.7",
+    "safetensors": "0.8.0",
+    "scikit-learn": "1.9.0",
+    "scipy": "1.18.1",
+    "semantic-version": "2.10.0",
+    "sentencepiece": "0.2.2",
+    "sentry-sdk": "2.68.1",
+    "setproctitle": "1.3.7",
+    "setuptools": "84.0.0",
+    "shellingham": "1.5.4",
+    "simplejson": "4.1.2",
+    "six": "1.17.0",
+    "sniffio": "1.3.1",
+    "sortedcontainers": "2.4.0",
+    "soundfile": "0.14.0",
+    "soxr": "1.1.0",
+    "sse-starlette": "3.4.8",
+    "starlette": "0.52.1",
+    "supervisor": "4.3.0",
+    "sympy": "1.14.0",
+    "tabulate": "0.10.0",
+    "tensorboard": "2.21.0",
+    "tensorboard-data-server": "0.7.2",
+    "threadpoolctl": "3.6.0",
+    "tiktoken": "0.14.0",
+    "tokenizers": "0.22.2",
+    "tomlkit": "0.13.3",
+    "torch": "2.10.0",
+    "torch-c-dlpack-ext": "0.1.5",
+    "torchaudio": "2.10.0",
+    "torchvision": "0.25.0",
+    "tqdm": "4.70.0",
+    "transformers": "4.57.6",
+    "transformers-stream-generator": "0.0.5",
+    "triton": "3.6.0",
+    "trl": "0.29.1",
+    "truststore": "0.10.4",
+    "typer": "0.27.2",
+    "typing-extensions": "4.16.0",
+    "typing-inspection": "0.4.4",
+    "tzdata": "2026.3",
+    "urllib3": "2.7.0",
+    "uvicorn": "0.52.4",
+    "uvloop": "0.22.1",
+    "vllm": "0.19.0",
+    "watchfiles": "1.2.0",
+    "wcwidth": "0.8.3",
+    "websockets": "15.0.1",
+    "werkzeug": "3.1.8",
+    "wheel": "0.48.0",
+    "xgrammar": "0.2.3",
+    "xxhash": "4.0.1",
+    "yarl": "1.24.5",
+    "zipp": "4.1.0",
+    "zstandard": "0.25.0"
+  },
+  "core_package_provenance": {
+    "triton": {
+      "version": "3.6.0",
+      "requires_dist": [
+        "importlib-metadata; python_version < \"3.10\"",
+        "cmake<4.0,>=3.20; extra == \"build\"",
+        "lit; extra == \"build\"",
+        "autopep8; extra == \"tests\"",
+        "isort; extra == \"tests\"",
+        "numpy; extra == \"tests\"",
+        "pytest; extra == \"tests\"",
+        "pytest-forked; extra == \"tests\"",
+        "pytest-xdist; extra == \"tests\"",
+        "scipy>=1.7.1; extra == \"tests\"",
+        "llnl-hatchet; extra == \"tests\"",
+        "matplotlib; extra == \"tutorials\"",
+        "pandas; extra == \"tutorials\"",
+        "tabulate; extra == \"tutorials\""
+      ],
+      "installer": "pip"
+    },
+    "ms-swift": {
+      "version": "4.4.2",
+      "requires_dist": [
+        "accelerate",
+        "addict",
+        "aiohttp",
+        "attrdict",
+        "binpacking",
+        "charset-normalizer",
+        "cpm-kernels",
+        "dacite",
+        "datasets<4.8.5,>=3.0",
+        "einops",
+        "fastapi",
+        "gradio<6.0,>=3.40.0",
+        "importlib-metadata",
+        "json-repair",
+        "matplotlib",
+        "modelscope>=1.23",
+        "nltk",
+        "numpy",
+        "openai",
+        "oss2",
+        "pandas",
+        "peft<0.20,>=0.11",
+        "pillow",
+        "PyYAML>=5.4",
+        "requests",
+        "rouge",
+        "safetensors",
+        "scipy",
+        "sentencepiece",
+        "simplejson>=3.3.0",
+        "sortedcontainers>=1.5.9",
+        "tensorboard",
+        "tiktoken",
+        "tqdm",
+        "transformers<5.13.0,>=4.33",
+        "transformers-stream-generator",
+        "trl<1.0,>=0.15",
+        "uvicorn",
+        "zstandard",
+        "accelerate; extra == \"all\"",
+        "addict; extra == \"all\"",
+        "aiohttp; extra == \"all\"",
+        "attrdict; extra == \"all\"",
+        "binpacking; extra == \"all\"",
+        "charset-normalizer; extra == \"all\"",
+        "cpm-kernels; extra == \"all\"",
+        "dacite; extra == \"all\"",
+        "datasets<4.8.5,>=3.0; extra == \"all\"",
+        "einops; extra == \"all\"",
+        "fastapi; extra == \"all\"",
+        "gradio<6.0,>=3.40.0; extra == \"all\"",
+        "importlib-metadata; extra == \"all\"",
+        "json-repair; extra == \"all\"",
+        "matplotlib; extra == \"all\"",
+        "modelscope>=1.23; extra == \"all\"",
+        "nltk; extra == \"all\"",
+        "numpy; extra == \"all\"",
+        "openai; extra == \"all\"",
+        "oss2; extra == \"all\"",
+        "pandas; extra == \"all\"",
+        "peft<0.20,>=0.11; extra == \"all\"",
+        "pillow; extra == \"all\"",
+        "PyYAML>=5.4; extra == \"all\"",
+        "requests; extra == \"all\"",
+        "rouge; extra == \"all\"",
+        "safetensors; extra == \"all\"",
+        "scipy; extra == \"all\"",
+        "sentencepiece; extra == \"all\"",
+        "simplejson>=3.3.0; extra == \"all\"",
+        "sortedcontainers>=1.5.9; extra == \"all\"",
+        "tensorboard; extra == \"all\"",
+        "tiktoken; extra == \"all\"",
+        "tqdm; extra == \"all\"",
+        "transformers<5.13.0,>=4.33; extra == \"all\"",
+        "transformers-stream-generator; extra == \"all\"",
+        "trl<1.0,>=0.15; extra == \"all\"",
+        "uvicorn; extra == \"all\"",
+        "zstandard; extra == \"all\"",
+        "evalscope>=1.0.0; extra == \"all\"",
+        "evalscope[opencompass]; extra == \"all\"",
+        "evalscope[vlmeval]; extra == \"all\"",
+        "swanlab; extra == \"all\"",
+        "ray; extra == \"all\"",
+        "evalscope>=1.0.0; extra == \"eval\"",
+        "evalscope[opencompass]; extra == \"eval\"",
+        "evalscope[vlmeval]; extra == \"eval\"",
+        "mcore-bridge>=1.5.0; extra == \"megatron\"",
+        "megatron-core>=0.16; extra == \"megatron\"",
+        "peft>=0.15; extra == \"megatron\"",
+        "ray; extra == \"ray\"",
+        "swanlab; extra == \"swanlab\""
+      ],
+      "installer": "pip"
+    },
+    "peft": {
+      "version": "0.19.1",
+      "requires_dist": [
+        "numpy>=1.17",
+        "packaging>=20.0",
+        "psutil",
+        "pyyaml",
+        "torch>=1.13.0",
+        "transformers",
+        "tqdm",
+        "accelerate>=0.21.0",
+        "safetensors",
+        "huggingface_hub>=0.25.0",
+        "black; extra == \"quality\"",
+        "hf-doc-builder; extra == \"quality\"",
+        "ruff~=0.12.8; extra == \"quality\"",
+        "black; extra == \"docs-specific\"",
+        "requests; extra == \"docs-specific\"",
+        "hf-doc-builder; extra == \"docs-specific\"",
+        "black; extra == \"dev\"",
+        "hf-doc-builder; extra == \"dev\"",
+        "ruff~=0.12.8; extra == \"dev\"",
+        "black; extra == \"dev\"",
+        "requests; extra == \"dev\"",
+        "hf-doc-builder; extra == \"dev\"",
+        "black; extra == \"test\"",
+        "hf-doc-builder; extra == \"test\"",
+        "ruff~=0.12.8; extra == \"test\"",
+        "black; extra == \"test\"",
+        "requests; extra == \"test\"",
+        "hf-doc-builder; extra == \"test\"",
+        "pytest; extra == \"test\"",
+        "pytest-cov; extra == \"test\"",
+        "pytest-xdist; extra == \"test\"",
+        "parameterized; extra == \"test\"",
+        "datasets; extra == \"test\"",
+        "diffusers; extra == \"test\"",
+        "scipy; extra == \"test\"",
+        "scikit-learn; extra == \"test\"",
+        "protobuf; extra == \"test\"",
+        "sentencepiece; extra == \"test\"",
+        "torchvision; extra == \"test\""
+      ],
+      "installer": "pip"
+    },
+    "pandas": {
+      "version": "2.3.3",
+      "requires_dist": [
+        "numpy>=1.22.4; python_version < \"3.11\"",
+        "numpy>=1.23.2; python_version == \"3.11\"",
+        "numpy>=1.26.0; python_version >= \"3.12\"",
+        "python-dateutil>=2.8.2",
+        "pytz>=2020.1",
+        "tzdata>=2022.7",
+        "hypothesis>=6.46.1; extra == \"test\"",
+        "pytest>=7.3.2; extra == \"test\"",
+        "pytest-xdist>=2.2.0; extra == \"test\"",
+        "pyarrow>=10.0.1; extra == \"pyarrow\"",
+        "bottleneck>=1.3.6; extra == \"performance\"",
+        "numba>=0.56.4; extra == \"performance\"",
+        "numexpr>=2.8.4; extra == \"performance\"",
+        "scipy>=1.10.0; extra == \"computation\"",
+        "xarray>=2022.12.0; extra == \"computation\"",
+        "fsspec>=2022.11.0; extra == \"fss\"",
+        "s3fs>=2022.11.0; extra == \"aws\"",
+        "gcsfs>=2022.11.0; extra == \"gcp\"",
+        "pandas-gbq>=0.19.0; extra == \"gcp\"",
+        "odfpy>=1.4.1; extra == \"excel\"",
+        "openpyxl>=3.1.0; extra == \"excel\"",
+        "python-calamine>=0.1.7; extra == \"excel\"",
+        "pyxlsb>=1.0.10; extra == \"excel\"",
+        "xlrd>=2.0.1; extra == \"excel\"",
+        "xlsxwriter>=3.0.5; extra == \"excel\"",
+        "pyarrow>=10.0.1; extra == \"parquet\"",
+        "pyarrow>=10.0.1; extra == \"feather\"",
+        "tables>=3.8.0; extra == \"hdf5\"",
+        "pyreadstat>=1.2.0; extra == \"spss\"",
+        "SQLAlchemy>=2.0.0; extra == \"postgresql\"",
+        "psycopg2>=2.9.6; extra == \"postgresql\"",
+        "adbc-driver-postgresql>=0.8.0; extra == \"postgresql\"",
+        "SQLAlchemy>=2.0.0; extra == \"mysql\"",
+        "pymysql>=1.0.2; extra == \"mysql\"",
+        "SQLAlchemy>=2.0.0; extra == \"sql-other\"",
+        "adbc-driver-postgresql>=0.8.0; extra == \"sql-other\"",
+        "adbc-driver-sqlite>=0.8.0; extra == \"sql-other\"",
+        "beautifulsoup4>=4.11.2; extra == \"html\"",
+        "html5lib>=1.1; extra == \"html\"",
+        "lxml>=4.9.2; extra == \"html\"",
+        "lxml>=4.9.2; extra == \"xml\"",
+        "matplotlib>=3.6.3; extra == \"plot\"",
+        "jinja2>=3.1.2; extra == \"output-formatting\"",
+        "tabulate>=0.9.0; extra == \"output-formatting\"",
+        "PyQt5>=5.15.9; extra == \"clipboard\"",
+        "qtpy>=2.3.0; extra == \"clipboard\"",
+        "zstandard>=0.19.0; extra == \"compression\"",
+        "dataframe-api-compat>=0.1.7; extra == \"consortium-standard\"",
+        "adbc-driver-postgresql>=0.8.0; extra == \"all\"",
+        "adbc-driver-sqlite>=0.8.0; extra == \"all\"",
+        "beautifulsoup4>=4.11.2; extra == \"all\"",
+        "bottleneck>=1.3.6; extra == \"all\"",
+        "dataframe-api-compat>=0.1.7; extra == \"all\"",
+        "fastparquet>=2022.12.0; extra == \"all\"",
+        "fsspec>=2022.11.0; extra == \"all\"",
+        "gcsfs>=2022.11.0; extra == \"all\"",
+        "html5lib>=1.1; extra == \"all\"",
+        "hypothesis>=6.46.1; extra == \"all\"",
+        "jinja2>=3.1.2; extra == \"all\"",
+        "lxml>=4.9.2; extra == \"all\"",
+        "matplotlib>=3.6.3; extra == \"all\"",
+        "numba>=0.56.4; extra == \"all\"",
+        "numexpr>=2.8.4; extra == \"all\"",
+        "odfpy>=1.4.1; extra == \"all\"",
+        "openpyxl>=3.1.0; extra == \"all\"",
+        "pandas-gbq>=0.19.0; extra == \"all\"",
+        "psycopg2>=2.9.6; extra == \"all\"",
+        "pyarrow>=10.0.1; extra == \"all\"",
+        "pymysql>=1.0.2; extra == \"all\"",
+        "PyQt5>=5.15.9; extra == \"all\"",
+        "pyreadstat>=1.2.0; extra == \"all\"",
+        "pytest>=7.3.2; extra == \"all\"",
+        "pytest-xdist>=2.2.0; extra == \"all\"",
+        "python-calamine>=0.1.7; extra == \"all\"",
+        "pyxlsb>=1.0.10; extra == \"all\"",
+        "qtpy>=2.3.0; extra == \"all\"",
+        "scipy>=1.10.0; extra == \"all\"",
+        "s3fs>=2022.11.0; extra == \"all\"",
+        "SQLAlchemy>=2.0.0; extra == \"all\"",
+        "tables>=3.8.0; extra == \"all\"",
+        "tabulate>=0.9.0; extra == \"all\"",
+        "xarray>=2022.12.0; extra == \"all\"",
+        "xlrd>=2.0.1; extra == \"all\"",
+        "xlsxwriter>=3.0.5; extra == \"all\"",
+        "zstandard>=0.19.0; extra == \"all\""
+      ],
+      "installer": "pip"
+    },
+    "transformers": {
+      "version": "4.57.6",
+      "requires_dist": [
+        "filelock",
+        "huggingface-hub<1.0,>=0.34.0",
+        "numpy>=1.17",
+        "packaging>=20.0",
+        "pyyaml>=5.1",
+        "regex!=2019.12.17",
+        "requests",
+        "tokenizers<=0.23.0,>=0.22.0",
+        "safetensors>=0.4.3",
+        "tqdm>=4.27",
+        "fugashi>=1.0; extra == \"ja\"",
+        "ipadic<2.0,>=1.0.0; extra == \"ja\"",
+        "unidic_lite>=1.0.7; extra == \"ja\"",
+        "unidic>=1.0.2; extra == \"ja\"",
+        "sudachipy>=0.6.6; extra == \"ja\"",
+        "sudachidict_core>=20220729; extra == \"ja\"",
+        "rhoknp<1.3.1,>=1.1.0; extra == \"ja\"",
+        "scikit-learn; extra == \"sklearn\"",
+        "tensorflow<2.16,>2.9; extra == \"tf\"",
+        "onnxconverter-common; extra == \"tf\"",
+        "tf2onnx; extra == \"tf\"",
+        "tensorflow-text<2.16; extra == \"tf\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"tf\"",
+        "keras<2.16,>2.9; extra == \"tf-cpu\"",
+        "tensorflow-cpu<2.16,>2.9; extra == \"tf-cpu\"",
+        "onnxconverter-common; extra == \"tf-cpu\"",
+        "tf2onnx; extra == \"tf-cpu\"",
+        "tensorflow-text<2.16; extra == \"tf-cpu\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"tf-cpu\"",
+        "tensorflow-probability<0.24; extra == \"tf-cpu\"",
+        "torch>=2.2; extra == \"torch\"",
+        "accelerate>=0.26.0; extra == \"torch\"",
+        "accelerate>=0.26.0; extra == \"accelerate\"",
+        "hf_xet; extra == \"hf-xet\"",
+        "faiss-cpu; extra == \"retrieval\"",
+        "datasets>=2.15.0; extra == \"retrieval\"",
+        "jax<=0.4.13,>=0.4.1; extra == \"flax\"",
+        "jaxlib<=0.4.13,>=0.4.1; extra == \"flax\"",
+        "flax<=0.7.0,>=0.4.1; extra == \"flax\"",
+        "optax<=0.1.4,>=0.0.8; extra == \"flax\"",
+        "scipy<1.13.0; extra == \"flax\"",
+        "tokenizers<=0.23.0,>=0.22.0; extra == \"tokenizers\"",
+        "ftfy; extra == \"ftfy\"",
+        "onnxruntime>=1.4.0; extra == \"onnxruntime\"",
+        "onnxruntime-tools>=1.4.2; extra == \"onnxruntime\"",
+        "onnxconverter-common; extra == \"onnx\"",
+        "tf2onnx; extra == \"onnx\"",
+        "onnxruntime>=1.4.0; extra == \"onnx\"",
+        "onnxruntime-tools>=1.4.2; extra == \"onnx\"",
+        "cookiecutter==1.7.3; extra == \"modelcreation\"",
+        "sagemaker>=2.31.0; extra == \"sagemaker\"",
+        "deepspeed>=0.9.3; extra == \"deepspeed\"",
+        "accelerate>=0.26.0; extra == \"deepspeed\"",
+        "optuna; extra == \"optuna\"",
+        "ray[tune]>=2.7.0; extra == \"ray\"",
+        "sigopt; extra == \"sigopt\"",
+        "kernels<=0.9,>=0.6.1; extra == \"hub-kernels\"",
+        "kernels<=0.9,>=0.6.1; extra == \"integrations\"",
+        "optuna; extra == \"integrations\"",
+        "ray[tune]>=2.7.0; extra == \"integrations\"",
+        "openai>=1.98.0; extra == \"serving\"",
+        "pydantic>=2; extra == \"serving\"",
+        "uvicorn; extra == \"serving\"",
+        "fastapi; extra == \"serving\"",
+        "starlette; extra == \"serving\"",
+        "torch>=2.2; extra == \"serving\"",
+        "accelerate>=0.26.0; extra == \"serving\"",
+        "librosa; extra == \"audio\"",
+        "pyctcdecode>=0.4.0; extra == \"audio\"",
+        "phonemizer; extra == \"audio\"",
+        "kenlm; extra == \"audio\"",
+        "torchaudio; extra == \"speech\"",
+        "librosa; extra == \"speech\"",
+        "pyctcdecode>=0.4.0; extra == \"speech\"",
+        "phonemizer; extra == \"speech\"",
+        "kenlm; extra == \"speech\"",
+        "torchaudio; extra == \"torch-speech\"",
+        "librosa; extra == \"torch-speech\"",
+        "pyctcdecode>=0.4.0; extra == \"torch-speech\"",
+        "phonemizer; extra == \"torch-speech\"",
+        "kenlm; extra == \"torch-speech\"",
+        "librosa; extra == \"tf-speech\"",
+        "pyctcdecode>=0.4.0; extra == \"tf-speech\"",
+        "phonemizer; extra == \"tf-speech\"",
+        "kenlm; extra == \"tf-speech\"",
+        "librosa; extra == \"flax-speech\"",
+        "pyctcdecode>=0.4.0; extra == \"flax-speech\"",
+        "phonemizer; extra == \"flax-speech\"",
+        "kenlm; extra == \"flax-speech\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"vision\"",
+        "timm!=1.0.18,<=1.0.19; extra == \"timm\"",
+        "torchvision; extra == \"torch-vision\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"torch-vision\"",
+        "natten<0.15.0,>=0.14.6; extra == \"natten\"",
+        "codecarbon>=2.8.1; extra == \"codecarbon\"",
+        "av; extra == \"video\"",
+        "num2words; extra == \"num2words\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"sentencepiece\"",
+        "protobuf; extra == \"sentencepiece\"",
+        "tiktoken; extra == \"tiktoken\"",
+        "blobfile; extra == \"tiktoken\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"mistral-common\"",
+        "jinja2>=3.1.0; extra == \"chat-template\"",
+        "pytest>=7.2.0; extra == \"testing\"",
+        "pytest-asyncio; extra == \"testing\"",
+        "pytest-rich; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-order; extra == \"testing\"",
+        "pytest-rerunfailures<16.0; extra == \"testing\"",
+        "timeout-decorator; extra == \"testing\"",
+        "parameterized>=0.9; extra == \"testing\"",
+        "psutil; extra == \"testing\"",
+        "datasets>=2.15.0; extra == \"testing\"",
+        "dill<0.3.5; extra == \"testing\"",
+        "evaluate>=0.2.0; extra == \"testing\"",
+        "pytest-timeout; extra == \"testing\"",
+        "ruff==0.13.1; extra == \"testing\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"testing\"",
+        "nltk<=3.8.1; extra == \"testing\"",
+        "GitPython<3.1.19; extra == \"testing\"",
+        "sacremoses; extra == \"testing\"",
+        "rjieba; extra == \"testing\"",
+        "beautifulsoup4; extra == \"testing\"",
+        "tensorboard; extra == \"testing\"",
+        "pydantic>=2; extra == \"testing\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"testing\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"testing\"",
+        "libcst; extra == \"testing\"",
+        "faiss-cpu; extra == \"testing\"",
+        "datasets>=2.15.0; extra == \"testing\"",
+        "cookiecutter==1.7.3; extra == \"testing\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"testing\"",
+        "openai>=1.98.0; extra == \"testing\"",
+        "pydantic>=2; extra == \"testing\"",
+        "uvicorn; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "starlette; extra == \"testing\"",
+        "torch>=2.2; extra == \"testing\"",
+        "accelerate>=0.26.0; extra == \"testing\"",
+        "deepspeed>=0.9.3; extra == \"deepspeed-testing\"",
+        "accelerate>=0.26.0; extra == \"deepspeed-testing\"",
+        "pytest>=7.2.0; extra == \"deepspeed-testing\"",
+        "pytest-asyncio; extra == \"deepspeed-testing\"",
+        "pytest-rich; extra == \"deepspeed-testing\"",
+        "pytest-xdist; extra == \"deepspeed-testing\"",
+        "pytest-order; extra == \"deepspeed-testing\"",
+        "pytest-rerunfailures<16.0; extra == \"deepspeed-testing\"",
+        "timeout-decorator; extra == \"deepspeed-testing\"",
+        "parameterized>=0.9; extra == \"deepspeed-testing\"",
+        "psutil; extra == \"deepspeed-testing\"",
+        "datasets>=2.15.0; extra == \"deepspeed-testing\"",
+        "dill<0.3.5; extra == \"deepspeed-testing\"",
+        "evaluate>=0.2.0; extra == \"deepspeed-testing\"",
+        "pytest-timeout; extra == \"deepspeed-testing\"",
+        "ruff==0.13.1; extra == \"deepspeed-testing\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"deepspeed-testing\"",
+        "nltk<=3.8.1; extra == \"deepspeed-testing\"",
+        "GitPython<3.1.19; extra == \"deepspeed-testing\"",
+        "sacremoses; extra == \"deepspeed-testing\"",
+        "rjieba; extra == \"deepspeed-testing\"",
+        "beautifulsoup4; extra == \"deepspeed-testing\"",
+        "tensorboard; extra == \"deepspeed-testing\"",
+        "pydantic>=2; extra == \"deepspeed-testing\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"deepspeed-testing\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"deepspeed-testing\"",
+        "libcst; extra == \"deepspeed-testing\"",
+        "faiss-cpu; extra == \"deepspeed-testing\"",
+        "datasets>=2.15.0; extra == \"deepspeed-testing\"",
+        "cookiecutter==1.7.3; extra == \"deepspeed-testing\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"deepspeed-testing\"",
+        "openai>=1.98.0; extra == \"deepspeed-testing\"",
+        "pydantic>=2; extra == \"deepspeed-testing\"",
+        "uvicorn; extra == \"deepspeed-testing\"",
+        "fastapi; extra == \"deepspeed-testing\"",
+        "starlette; extra == \"deepspeed-testing\"",
+        "torch>=2.2; extra == \"deepspeed-testing\"",
+        "accelerate>=0.26.0; extra == \"deepspeed-testing\"",
+        "optuna; extra == \"deepspeed-testing\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"deepspeed-testing\"",
+        "protobuf; extra == \"deepspeed-testing\"",
+        "ruff==0.13.1; extra == \"ruff\"",
+        "datasets>=2.15.0; extra == \"quality\"",
+        "ruff==0.13.1; extra == \"quality\"",
+        "GitPython<3.1.19; extra == \"quality\"",
+        "urllib3<2.0.0; extra == \"quality\"",
+        "libcst; extra == \"quality\"",
+        "rich; extra == \"quality\"",
+        "pandas<2.3.0; extra == \"quality\"",
+        "tensorflow<2.16,>2.9; extra == \"all\"",
+        "onnxconverter-common; extra == \"all\"",
+        "tf2onnx; extra == \"all\"",
+        "tensorflow-text<2.16; extra == \"all\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"all\"",
+        "torch>=2.2; extra == \"all\"",
+        "accelerate>=0.26.0; extra == \"all\"",
+        "jax<=0.4.13,>=0.4.1; extra == \"all\"",
+        "jaxlib<=0.4.13,>=0.4.1; extra == \"all\"",
+        "flax<=0.7.0,>=0.4.1; extra == \"all\"",
+        "optax<=0.1.4,>=0.0.8; extra == \"all\"",
+        "scipy<1.13.0; extra == \"all\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"all\"",
+        "protobuf; extra == \"all\"",
+        "tokenizers<=0.23.0,>=0.22.0; extra == \"all\"",
+        "torchaudio; extra == \"all\"",
+        "librosa; extra == \"all\"",
+        "pyctcdecode>=0.4.0; extra == \"all\"",
+        "phonemizer; extra == \"all\"",
+        "kenlm; extra == \"all\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"all\"",
+        "kernels<=0.9,>=0.6.1; extra == \"all\"",
+        "optuna; extra == \"all\"",
+        "ray[tune]>=2.7.0; extra == \"all\"",
+        "timm!=1.0.18,<=1.0.19; extra == \"all\"",
+        "torchvision; extra == \"all\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"all\"",
+        "codecarbon>=2.8.1; extra == \"all\"",
+        "accelerate>=0.26.0; extra == \"all\"",
+        "av; extra == \"all\"",
+        "num2words; extra == \"all\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"all\"",
+        "jinja2>=3.1.0; extra == \"all\"",
+        "pytest>=7.2.0; extra == \"dev-torch\"",
+        "pytest-asyncio; extra == \"dev-torch\"",
+        "pytest-rich; extra == \"dev-torch\"",
+        "pytest-xdist; extra == \"dev-torch\"",
+        "pytest-order; extra == \"dev-torch\"",
+        "pytest-rerunfailures<16.0; extra == \"dev-torch\"",
+        "timeout-decorator; extra == \"dev-torch\"",
+        "parameterized>=0.9; extra == \"dev-torch\"",
+        "psutil; extra == \"dev-torch\"",
+        "datasets>=2.15.0; extra == \"dev-torch\"",
+        "dill<0.3.5; extra == \"dev-torch\"",
+        "evaluate>=0.2.0; extra == \"dev-torch\"",
+        "pytest-timeout; extra == \"dev-torch\"",
+        "ruff==0.13.1; extra == \"dev-torch\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"dev-torch\"",
+        "nltk<=3.8.1; extra == \"dev-torch\"",
+        "GitPython<3.1.19; extra == \"dev-torch\"",
+        "sacremoses; extra == \"dev-torch\"",
+        "rjieba; extra == \"dev-torch\"",
+        "beautifulsoup4; extra == \"dev-torch\"",
+        "tensorboard; extra == \"dev-torch\"",
+        "pydantic>=2; extra == \"dev-torch\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev-torch\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"dev-torch\"",
+        "libcst; extra == \"dev-torch\"",
+        "faiss-cpu; extra == \"dev-torch\"",
+        "datasets>=2.15.0; extra == \"dev-torch\"",
+        "cookiecutter==1.7.3; extra == \"dev-torch\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"dev-torch\"",
+        "openai>=1.98.0; extra == \"dev-torch\"",
+        "pydantic>=2; extra == \"dev-torch\"",
+        "uvicorn; extra == \"dev-torch\"",
+        "fastapi; extra == \"dev-torch\"",
+        "starlette; extra == \"dev-torch\"",
+        "torch>=2.2; extra == \"dev-torch\"",
+        "accelerate>=0.26.0; extra == \"dev-torch\"",
+        "torch>=2.2; extra == \"dev-torch\"",
+        "accelerate>=0.26.0; extra == \"dev-torch\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev-torch\"",
+        "protobuf; extra == \"dev-torch\"",
+        "tokenizers<=0.23.0,>=0.22.0; extra == \"dev-torch\"",
+        "torchaudio; extra == \"dev-torch\"",
+        "librosa; extra == \"dev-torch\"",
+        "pyctcdecode>=0.4.0; extra == \"dev-torch\"",
+        "phonemizer; extra == \"dev-torch\"",
+        "kenlm; extra == \"dev-torch\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev-torch\"",
+        "kernels<=0.9,>=0.6.1; extra == \"dev-torch\"",
+        "optuna; extra == \"dev-torch\"",
+        "ray[tune]>=2.7.0; extra == \"dev-torch\"",
+        "timm!=1.0.18,<=1.0.19; extra == \"dev-torch\"",
+        "torchvision; extra == \"dev-torch\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev-torch\"",
+        "codecarbon>=2.8.1; extra == \"dev-torch\"",
+        "datasets>=2.15.0; extra == \"dev-torch\"",
+        "ruff==0.13.1; extra == \"dev-torch\"",
+        "GitPython<3.1.19; extra == \"dev-torch\"",
+        "urllib3<2.0.0; extra == \"dev-torch\"",
+        "libcst; extra == \"dev-torch\"",
+        "rich; extra == \"dev-torch\"",
+        "pandas<2.3.0; extra == \"dev-torch\"",
+        "fugashi>=1.0; extra == \"dev-torch\"",
+        "ipadic<2.0,>=1.0.0; extra == \"dev-torch\"",
+        "unidic_lite>=1.0.7; extra == \"dev-torch\"",
+        "unidic>=1.0.2; extra == \"dev-torch\"",
+        "sudachipy>=0.6.6; extra == \"dev-torch\"",
+        "sudachidict_core>=20220729; extra == \"dev-torch\"",
+        "rhoknp<1.3.1,>=1.1.0; extra == \"dev-torch\"",
+        "scikit-learn; extra == \"dev-torch\"",
+        "cookiecutter==1.7.3; extra == \"dev-torch\"",
+        "onnxruntime>=1.4.0; extra == \"dev-torch\"",
+        "onnxruntime-tools>=1.4.2; extra == \"dev-torch\"",
+        "num2words; extra == \"dev-torch\"",
+        "pytest>=7.2.0; extra == \"dev-tensorflow\"",
+        "pytest-asyncio; extra == \"dev-tensorflow\"",
+        "pytest-rich; extra == \"dev-tensorflow\"",
+        "pytest-xdist; extra == \"dev-tensorflow\"",
+        "pytest-order; extra == \"dev-tensorflow\"",
+        "pytest-rerunfailures<16.0; extra == \"dev-tensorflow\"",
+        "timeout-decorator; extra == \"dev-tensorflow\"",
+        "parameterized>=0.9; extra == \"dev-tensorflow\"",
+        "psutil; extra == \"dev-tensorflow\"",
+        "datasets>=2.15.0; extra == \"dev-tensorflow\"",
+        "dill<0.3.5; extra == \"dev-tensorflow\"",
+        "evaluate>=0.2.0; extra == \"dev-tensorflow\"",
+        "pytest-timeout; extra == \"dev-tensorflow\"",
+        "ruff==0.13.1; extra == \"dev-tensorflow\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"dev-tensorflow\"",
+        "nltk<=3.8.1; extra == \"dev-tensorflow\"",
+        "GitPython<3.1.19; extra == \"dev-tensorflow\"",
+        "sacremoses; extra == \"dev-tensorflow\"",
+        "rjieba; extra == \"dev-tensorflow\"",
+        "beautifulsoup4; extra == \"dev-tensorflow\"",
+        "tensorboard; extra == \"dev-tensorflow\"",
+        "pydantic>=2; extra == \"dev-tensorflow\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev-tensorflow\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"dev-tensorflow\"",
+        "libcst; extra == \"dev-tensorflow\"",
+        "faiss-cpu; extra == \"dev-tensorflow\"",
+        "datasets>=2.15.0; extra == \"dev-tensorflow\"",
+        "cookiecutter==1.7.3; extra == \"dev-tensorflow\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"dev-tensorflow\"",
+        "openai>=1.98.0; extra == \"dev-tensorflow\"",
+        "pydantic>=2; extra == \"dev-tensorflow\"",
+        "uvicorn; extra == \"dev-tensorflow\"",
+        "fastapi; extra == \"dev-tensorflow\"",
+        "starlette; extra == \"dev-tensorflow\"",
+        "torch>=2.2; extra == \"dev-tensorflow\"",
+        "accelerate>=0.26.0; extra == \"dev-tensorflow\"",
+        "tensorflow<2.16,>2.9; extra == \"dev-tensorflow\"",
+        "onnxconverter-common; extra == \"dev-tensorflow\"",
+        "tf2onnx; extra == \"dev-tensorflow\"",
+        "tensorflow-text<2.16; extra == \"dev-tensorflow\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"dev-tensorflow\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev-tensorflow\"",
+        "protobuf; extra == \"dev-tensorflow\"",
+        "tokenizers<=0.23.0,>=0.22.0; extra == \"dev-tensorflow\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev-tensorflow\"",
+        "datasets>=2.15.0; extra == \"dev-tensorflow\"",
+        "ruff==0.13.1; extra == \"dev-tensorflow\"",
+        "GitPython<3.1.19; extra == \"dev-tensorflow\"",
+        "urllib3<2.0.0; extra == \"dev-tensorflow\"",
+        "libcst; extra == \"dev-tensorflow\"",
+        "rich; extra == \"dev-tensorflow\"",
+        "pandas<2.3.0; extra == \"dev-tensorflow\"",
+        "scikit-learn; extra == \"dev-tensorflow\"",
+        "cookiecutter==1.7.3; extra == \"dev-tensorflow\"",
+        "onnxconverter-common; extra == \"dev-tensorflow\"",
+        "tf2onnx; extra == \"dev-tensorflow\"",
+        "onnxruntime>=1.4.0; extra == \"dev-tensorflow\"",
+        "onnxruntime-tools>=1.4.2; extra == \"dev-tensorflow\"",
+        "librosa; extra == \"dev-tensorflow\"",
+        "pyctcdecode>=0.4.0; extra == \"dev-tensorflow\"",
+        "phonemizer; extra == \"dev-tensorflow\"",
+        "kenlm; extra == \"dev-tensorflow\"",
+        "tensorflow<2.16,>2.9; extra == \"dev\"",
+        "onnxconverter-common; extra == \"dev\"",
+        "tf2onnx; extra == \"dev\"",
+        "tensorflow-text<2.16; extra == \"dev\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"dev\"",
+        "torch>=2.2; extra == \"dev\"",
+        "accelerate>=0.26.0; extra == \"dev\"",
+        "jax<=0.4.13,>=0.4.1; extra == \"dev\"",
+        "jaxlib<=0.4.13,>=0.4.1; extra == \"dev\"",
+        "flax<=0.7.0,>=0.4.1; extra == \"dev\"",
+        "optax<=0.1.4,>=0.0.8; extra == \"dev\"",
+        "scipy<1.13.0; extra == \"dev\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev\"",
+        "protobuf; extra == \"dev\"",
+        "tokenizers<=0.23.0,>=0.22.0; extra == \"dev\"",
+        "torchaudio; extra == \"dev\"",
+        "librosa; extra == \"dev\"",
+        "pyctcdecode>=0.4.0; extra == \"dev\"",
+        "phonemizer; extra == \"dev\"",
+        "kenlm; extra == \"dev\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev\"",
+        "kernels<=0.9,>=0.6.1; extra == \"dev\"",
+        "optuna; extra == \"dev\"",
+        "ray[tune]>=2.7.0; extra == \"dev\"",
+        "timm!=1.0.18,<=1.0.19; extra == \"dev\"",
+        "torchvision; extra == \"dev\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev\"",
+        "codecarbon>=2.8.1; extra == \"dev\"",
+        "accelerate>=0.26.0; extra == \"dev\"",
+        "av; extra == \"dev\"",
+        "num2words; extra == \"dev\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"dev\"",
+        "jinja2>=3.1.0; extra == \"dev\"",
+        "pytest>=7.2.0; extra == \"dev\"",
+        "pytest-asyncio; extra == \"dev\"",
+        "pytest-rich; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-order; extra == \"dev\"",
+        "pytest-rerunfailures<16.0; extra == \"dev\"",
+        "timeout-decorator; extra == \"dev\"",
+        "parameterized>=0.9; extra == \"dev\"",
+        "psutil; extra == \"dev\"",
+        "datasets>=2.15.0; extra == \"dev\"",
+        "dill<0.3.5; extra == \"dev\"",
+        "evaluate>=0.2.0; extra == \"dev\"",
+        "pytest-timeout; extra == \"dev\"",
+        "ruff==0.13.1; extra == \"dev\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"dev\"",
+        "nltk<=3.8.1; extra == \"dev\"",
+        "GitPython<3.1.19; extra == \"dev\"",
+        "sacremoses; extra == \"dev\"",
+        "rjieba; extra == \"dev\"",
+        "beautifulsoup4; extra == \"dev\"",
+        "tensorboard; extra == \"dev\"",
+        "pydantic>=2; extra == \"dev\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"dev\"",
+        "libcst; extra == \"dev\"",
+        "faiss-cpu; extra == \"dev\"",
+        "datasets>=2.15.0; extra == \"dev\"",
+        "cookiecutter==1.7.3; extra == \"dev\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"dev\"",
+        "openai>=1.98.0; extra == \"dev\"",
+        "pydantic>=2; extra == \"dev\"",
+        "uvicorn; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "starlette; extra == \"dev\"",
+        "torch>=2.2; extra == \"dev\"",
+        "accelerate>=0.26.0; extra == \"dev\"",
+        "datasets>=2.15.0; extra == \"dev\"",
+        "ruff==0.13.1; extra == \"dev\"",
+        "GitPython<3.1.19; extra == \"dev\"",
+        "urllib3<2.0.0; extra == \"dev\"",
+        "libcst; extra == \"dev\"",
+        "rich; extra == \"dev\"",
+        "pandas<2.3.0; extra == \"dev\"",
+        "fugashi>=1.0; extra == \"dev\"",
+        "ipadic<2.0,>=1.0.0; extra == \"dev\"",
+        "unidic_lite>=1.0.7; extra == \"dev\"",
+        "unidic>=1.0.2; extra == \"dev\"",
+        "sudachipy>=0.6.6; extra == \"dev\"",
+        "sudachidict_core>=20220729; extra == \"dev\"",
+        "rhoknp<1.3.1,>=1.1.0; extra == \"dev\"",
+        "scikit-learn; extra == \"dev\"",
+        "cookiecutter==1.7.3; extra == \"dev\"",
+        "filelock; extra == \"torchhub\"",
+        "huggingface-hub<1.0,>=0.34.0; extra == \"torchhub\"",
+        "importlib_metadata; extra == \"torchhub\"",
+        "numpy>=1.17; extra == \"torchhub\"",
+        "packaging>=20.0; extra == \"torchhub\"",
+        "protobuf; extra == \"torchhub\"",
+        "regex!=2019.12.17; extra == \"torchhub\"",
+        "requests; extra == \"torchhub\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"torchhub\"",
+        "torch>=2.2; extra == \"torchhub\"",
+        "tokenizers<=0.23.0,>=0.22.0; extra == \"torchhub\"",
+        "tqdm>=4.27; extra == \"torchhub\"",
+        "optimum-benchmark>=0.3.0; extra == \"benchmark\"",
+        "opentelemetry-api; extra == \"open-telemetry\"",
+        "opentelemetry-exporter-otlp; extra == \"open-telemetry\"",
+        "opentelemetry-sdk; extra == \"open-telemetry\""
+      ],
+      "installer": "pip"
+    },
+    "pyarrow": {
+      "version": "25.0.1",
+      "requires_dist": [],
+      "installer": "pip"
+    },
+    "deepspeed": {
+      "version": "0.19.1",
+      "requires_dist": [
+        "einops",
+        "hjson",
+        "msgpack",
+        "ninja",
+        "numpy",
+        "packaging>=20.0",
+        "psutil",
+        "py-cpuinfo",
+        "pydantic>=2.0.0",
+        "torch>=2.0.0",
+        "tqdm",
+        "mpi4py; extra == \"1bit-mpi\"",
+        "autodoc_pydantic>=2.0.0; extra == \"readthedocs\"",
+        "docutils<0.18; extra == \"readthedocs\"",
+        "hjson; extra == \"readthedocs\"",
+        "packaging; extra == \"readthedocs\"",
+        "psutil; extra == \"readthedocs\"",
+        "py-cpuinfo; extra == \"readthedocs\"",
+        "pydantic>=2.0.0; extra == \"readthedocs\"",
+        "recommonmark; extra == \"readthedocs\"",
+        "sphinx_rtd_theme; extra == \"readthedocs\"",
+        "torch>=2.0.0; extra == \"readthedocs\"",
+        "tqdm; extra == \"readthedocs\"",
+        "accelerate; extra == \"dev\"",
+        "clang-format==18.1.3; extra == \"dev\"",
+        "comet_ml>=3.41.0; extra == \"dev\"",
+        "deepspeed-kernels; sys_platform == \"linux\" and extra == \"dev\"",
+        "docutils<0.18; extra == \"dev\"",
+        "future; extra == \"dev\"",
+        "importlib-metadata>=4; extra == \"dev\"",
+        "mup; extra == \"dev\"",
+        "pre-commit>=3.2.0; extra == \"dev\"",
+        "pytest<8.4.0,>=7.2.0; extra == \"dev\"",
+        "pytest-forked; extra == \"dev\"",
+        "pytest-randomly; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "qtorch==0.3.0; extra == \"dev\"",
+        "recommonmark; extra == \"dev\"",
+        "sphinx; extra == \"dev\"",
+        "sphinx-rtd-theme; extra == \"dev\"",
+        "tensorboard; extra == \"dev\"",
+        "torchvision; extra == \"dev\"",
+        "transformers>=4.51.3; extra == \"dev\"",
+        "wandb; extra == \"dev\"",
+        "tabulate; extra == \"autotuning\"",
+        "hjson; extra == \"autotuning-ml\"",
+        "tabulate; extra == \"autotuning-ml\"",
+        "xgboost; extra == \"autotuning-ml\"",
+        "triton==1.0.0; extra == \"sparse-attn\"",
+        "neural-compressor==2.1.0; extra == \"sparse\"",
+        "google; extra == \"inf\"",
+        "lm-eval==0.3.0; extra == \"inf\"",
+        "protobuf; extra == \"inf\"",
+        "qtorch; extra == \"inf\"",
+        "safetensors; extra == \"inf\"",
+        "sentencepiece; extra == \"inf\"",
+        "transformers>=4.32.1; extra == \"inf\"",
+        "diffusers>=0.25.0; extra == \"sd\"",
+        "triton>=2.1.0; extra == \"sd\"",
+        "triton==2.1.0; extra == \"triton\"",
+        "scipy; extra == \"deepcompile\"",
+        "transformers>=4.32.1; extra == \"all\"",
+        "pytest-forked; extra == \"all\"",
+        "lm-eval==0.3.0; extra == \"all\"",
+        "pydantic>=2.0.0; extra == \"all\"",
+        "sphinx_rtd_theme; extra == \"all\"",
+        "torchvision; extra == \"all\"",
+        "qtorch; extra == \"all\"",
+        "docutils<0.18; extra == \"all\"",
+        "neural-compressor==2.1.0; extra == \"all\"",
+        "triton==2.1.0; extra == \"all\"",
+        "qtorch==0.3.0; extra == \"all\"",
+        "mpi4py; extra == \"all\"",
+        "scipy; extra == \"all\"",
+        "recommonmark; extra == \"all\"",
+        "clang-format==18.1.3; extra == \"all\"",
+        "importlib-metadata>=4; extra == \"all\"",
+        "wandb; extra == \"all\"",
+        "xgboost; extra == \"all\"",
+        "future; extra == \"all\"",
+        "google; extra == \"all\"",
+        "sphinx; extra == \"all\"",
+        "packaging; extra == \"all\"",
+        "torch>=2.0.0; extra == \"all\"",
+        "comet_ml>=3.41.0; extra == \"all\"",
+        "diffusers>=0.25.0; extra == \"all\"",
+        "mup; extra == \"all\"",
+        "tabulate; extra == \"all\"",
+        "tqdm; extra == \"all\"",
+        "sphinx-rtd-theme; extra == \"all\"",
+        "pre-commit>=3.2.0; extra == \"all\"",
+        "protobuf; extra == \"all\"",
+        "psutil; extra == \"all\"",
+        "tensorboard; extra == \"all\"",
+        "triton==1.0.0; extra == \"all\"",
+        "pytest<8.4.0,>=7.2.0; extra == \"all\"",
+        "transformers>=4.51.3; extra == \"all\"",
+        "deepspeed-kernels; sys_platform == \"linux\" and extra == \"all\"",
+        "pytest-xdist; extra == \"all\"",
+        "sentencepiece; extra == \"all\"",
+        "autodoc_pydantic>=2.0.0; extra == \"all\"",
+        "pytest-randomly; extra == \"all\"",
+        "triton>=2.1.0; extra == \"all\"",
+        "py-cpuinfo; extra == \"all\"",
+        "hjson; extra == \"all\"",
+        "safetensors; extra == \"all\"",
+        "accelerate; extra == \"all\""
+      ],
+      "installer": "pip"
+    },
+    "torchvision": {
+      "version": "0.25.0",
+      "requires_dist": [
+        "numpy",
+        "torch (==2.10.0)",
+        "pillow (!=8.3.*,>=5.3.0)",
+        "gdown (>=4.7.3) ; extra == 'gdown'",
+        "scipy ; extra == 'scipy'"
+      ],
+      "installer": "pip"
+    },
+    "huggingface-hub": {
+      "version": "0.36.2",
+      "requires_dist": [
+        "filelock",
+        "fsspec>=2023.5.0",
+        "hf-xet<2.0.0,>=1.1.3; platform_machine == \"x86_64\" or platform_machine == \"amd64\" or platform_machine == \"arm64\" or platform_machine == \"aarch64\"",
+        "packaging>=20.9",
+        "pyyaml>=5.1",
+        "requests",
+        "tqdm>=4.42.1",
+        "typing-extensions>=3.7.4.3",
+        "InquirerPy==0.3.4; extra == \"cli\"",
+        "aiohttp; extra == \"inference\"",
+        "authlib>=1.3.2; extra == \"oauth\"",
+        "fastapi; extra == \"oauth\"",
+        "httpx; extra == \"oauth\"",
+        "itsdangerous; extra == \"oauth\"",
+        "torch; extra == \"torch\"",
+        "safetensors[torch]; extra == \"torch\"",
+        "hf_transfer>=0.1.4; extra == \"hf-transfer\"",
+        "toml; extra == \"fastai\"",
+        "fastai>=2.4; extra == \"fastai\"",
+        "fastcore>=1.3.27; extra == \"fastai\"",
+        "tensorflow; extra == \"tensorflow\"",
+        "pydot; extra == \"tensorflow\"",
+        "graphviz; extra == \"tensorflow\"",
+        "tensorflow; extra == \"tensorflow-testing\"",
+        "keras<3.0; extra == \"tensorflow-testing\"",
+        "hf-xet<2.0.0,>=1.1.2; extra == \"hf-xet\"",
+        "mcp>=1.8.0; extra == \"mcp\"",
+        "typer; extra == \"mcp\"",
+        "aiohttp; extra == \"mcp\"",
+        "InquirerPy==0.3.4; extra == \"testing\"",
+        "aiohttp; extra == \"testing\"",
+        "authlib>=1.3.2; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "httpx; extra == \"testing\"",
+        "itsdangerous; extra == \"testing\"",
+        "jedi; extra == \"testing\"",
+        "Jinja2; extra == \"testing\"",
+        "pytest<8.2.2,>=8.1.1; extra == \"testing\"",
+        "pytest-cov; extra == \"testing\"",
+        "pytest-env; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-vcr; extra == \"testing\"",
+        "pytest-asyncio; extra == \"testing\"",
+        "pytest-rerunfailures<16.0; extra == \"testing\"",
+        "pytest-mock; extra == \"testing\"",
+        "urllib3<2.0; extra == \"testing\"",
+        "soundfile; extra == \"testing\"",
+        "Pillow; extra == \"testing\"",
+        "gradio>=4.0.0; extra == \"testing\"",
+        "numpy; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "typing-extensions>=4.8.0; extra == \"typing\"",
+        "types-PyYAML; extra == \"typing\"",
+        "types-requests; extra == \"typing\"",
+        "types-simplejson; extra == \"typing\"",
+        "types-toml; extra == \"typing\"",
+        "types-tqdm; extra == \"typing\"",
+        "types-urllib3; extra == \"typing\"",
+        "ruff>=0.9.0; extra == \"quality\"",
+        "mypy<1.15.0,>=1.14.1; python_version == \"3.8\" and extra == \"quality\"",
+        "mypy==1.15.0; python_version >= \"3.9\" and extra == \"quality\"",
+        "libcst>=1.4.0; extra == \"quality\"",
+        "ty; extra == \"quality\"",
+        "InquirerPy==0.3.4; extra == \"all\"",
+        "aiohttp; extra == \"all\"",
+        "authlib>=1.3.2; extra == \"all\"",
+        "fastapi; extra == \"all\"",
+        "httpx; extra == \"all\"",
+        "itsdangerous; extra == \"all\"",
+        "jedi; extra == \"all\"",
+        "Jinja2; extra == \"all\"",
+        "pytest<8.2.2,>=8.1.1; extra == \"all\"",
+        "pytest-cov; extra == \"all\"",
+        "pytest-env; extra == \"all\"",
+        "pytest-xdist; extra == \"all\"",
+        "pytest-vcr; extra == \"all\"",
+        "pytest-asyncio; extra == \"all\"",
+        "pytest-rerunfailures<16.0; extra == \"all\"",
+        "pytest-mock; extra == \"all\"",
+        "urllib3<2.0; extra == \"all\"",
+        "soundfile; extra == \"all\"",
+        "Pillow; extra == \"all\"",
+        "gradio>=4.0.0; extra == \"all\"",
+        "numpy; extra == \"all\"",
+        "fastapi; extra == \"all\"",
+        "ruff>=0.9.0; extra == \"all\"",
+        "mypy<1.15.0,>=1.14.1; python_version == \"3.8\" and extra == \"all\"",
+        "mypy==1.15.0; python_version >= \"3.9\" and extra == \"all\"",
+        "libcst>=1.4.0; extra == \"all\"",
+        "ty; extra == \"all\"",
+        "typing-extensions>=4.8.0; extra == \"all\"",
+        "types-PyYAML; extra == \"all\"",
+        "types-requests; extra == \"all\"",
+        "types-simplejson; extra == \"all\"",
+        "types-toml; extra == \"all\"",
+        "types-tqdm; extra == \"all\"",
+        "types-urllib3; extra == \"all\"",
+        "InquirerPy==0.3.4; extra == \"dev\"",
+        "aiohttp; extra == \"dev\"",
+        "authlib>=1.3.2; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "httpx; extra == \"dev\"",
+        "itsdangerous; extra == \"dev\"",
+        "jedi; extra == \"dev\"",
+        "Jinja2; extra == \"dev\"",
+        "pytest<8.2.2,>=8.1.1; extra == \"dev\"",
+        "pytest-cov; extra == \"dev\"",
+        "pytest-env; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-vcr; extra == \"dev\"",
+        "pytest-asyncio; extra == \"dev\"",
+        "pytest-rerunfailures<16.0; extra == \"dev\"",
+        "pytest-mock; extra == \"dev\"",
+        "urllib3<2.0; extra == \"dev\"",
+        "soundfile; extra == \"dev\"",
+        "Pillow; extra == \"dev\"",
+        "gradio>=4.0.0; extra == \"dev\"",
+        "numpy; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "ruff>=0.9.0; extra == \"dev\"",
+        "mypy<1.15.0,>=1.14.1; python_version == \"3.8\" and extra == \"dev\"",
+        "mypy==1.15.0; python_version >= \"3.9\" and extra == \"dev\"",
+        "libcst>=1.4.0; extra == \"dev\"",
+        "ty; extra == \"dev\"",
+        "typing-extensions>=4.8.0; extra == \"dev\"",
+        "types-PyYAML; extra == \"dev\"",
+        "types-requests; extra == \"dev\"",
+        "types-simplejson; extra == \"dev\"",
+        "types-toml; extra == \"dev\"",
+        "types-tqdm; extra == \"dev\"",
+        "types-urllib3; extra == \"dev\""
+      ],
+      "installer": "pip"
+    },
+    "pillow": {
+      "version": "11.3.0",
+      "requires_dist": [
+        "furo; extra == \"docs\"",
+        "olefile; extra == \"docs\"",
+        "sphinx>=8.2; extra == \"docs\"",
+        "sphinx-autobuild; extra == \"docs\"",
+        "sphinx-copybutton; extra == \"docs\"",
+        "sphinx-inline-tabs; extra == \"docs\"",
+        "sphinxext-opengraph; extra == \"docs\"",
+        "olefile; extra == \"fpx\"",
+        "olefile; extra == \"mic\"",
+        "pyarrow; extra == \"test-arrow\"",
+        "check-manifest; extra == \"tests\"",
+        "coverage>=7.4.2; extra == \"tests\"",
+        "defusedxml; extra == \"tests\"",
+        "markdown2; extra == \"tests\"",
+        "olefile; extra == \"tests\"",
+        "packaging; extra == \"tests\"",
+        "pyroma; extra == \"tests\"",
+        "pytest; extra == \"tests\"",
+        "pytest-cov; extra == \"tests\"",
+        "pytest-timeout; extra == \"tests\"",
+        "pytest-xdist; extra == \"tests\"",
+        "trove-classifiers>=2024.10.12; extra == \"tests\"",
+        "typing-extensions; python_version < \"3.10\" and extra == \"typing\"",
+        "defusedxml; extra == \"xmp\""
+      ],
+      "installer": "pip"
+    },
+    "accelerate": {
+      "version": "1.14.0",
+      "requires_dist": [
+        "numpy>=1.17",
+        "packaging>=20.0",
+        "psutil",
+        "pyyaml",
+        "torch>=2.0.0",
+        "huggingface_hub>=0.21.0",
+        "safetensors>=0.4.3",
+        "ruff==0.13.1; extra == \"quality\"",
+        "pytest>=7.2.0; extra == \"test-prod\"",
+        "pytest-xdist; extra == \"test-prod\"",
+        "pytest-subtests; extra == \"test-prod\"",
+        "parameterized; extra == \"test-prod\"",
+        "pytest-order; extra == \"test-prod\"",
+        "datasets; extra == \"test-dev\"",
+        "diffusers; extra == \"test-dev\"",
+        "evaluate; extra == \"test-dev\"",
+        "torchdata>=0.8.0; extra == \"test-dev\"",
+        "torchpippy>=0.2.0; extra == \"test-dev\"",
+        "transformers; extra == \"test-dev\"",
+        "scipy; extra == \"test-dev\"",
+        "scikit-learn; extra == \"test-dev\"",
+        "tqdm; extra == \"test-dev\"",
+        "bitsandbytes; extra == \"test-dev\"",
+        "timm; extra == \"test-dev\"",
+        "pytest>=7.2.0; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-subtests; extra == \"testing\"",
+        "parameterized; extra == \"testing\"",
+        "pytest-order; extra == \"testing\"",
+        "datasets; extra == \"testing\"",
+        "diffusers; extra == \"testing\"",
+        "evaluate; extra == \"testing\"",
+        "torchdata>=0.8.0; extra == \"testing\"",
+        "torchpippy>=0.2.0; extra == \"testing\"",
+        "transformers; extra == \"testing\"",
+        "scipy; extra == \"testing\"",
+        "scikit-learn; extra == \"testing\"",
+        "tqdm; extra == \"testing\"",
+        "bitsandbytes; extra == \"testing\"",
+        "timm; extra == \"testing\"",
+        "deepspeed; extra == \"deepspeed\"",
+        "rich; extra == \"rich\"",
+        "torchao; extra == \"test-fp8\"",
+        "wandb; extra == \"test-trackers\"",
+        "tensorboard; extra == \"test-trackers\"",
+        "dvclive; extra == \"test-trackers\"",
+        "matplotlib; extra == \"test-trackers\"",
+        "swanlab[dashboard]; extra == \"test-trackers\"",
+        "trackio; extra == \"test-trackers\"",
+        "ruff==0.13.1; extra == \"dev\"",
+        "pytest>=7.2.0; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-subtests; extra == \"dev\"",
+        "parameterized; extra == \"dev\"",
+        "pytest-order; extra == \"dev\"",
+        "datasets; extra == \"dev\"",
+        "diffusers; extra == \"dev\"",
+        "evaluate; extra == \"dev\"",
+        "torchdata>=0.8.0; extra == \"dev\"",
+        "torchpippy>=0.2.0; extra == \"dev\"",
+        "transformers; extra == \"dev\"",
+        "scipy; extra == \"dev\"",
+        "scikit-learn; extra == \"dev\"",
+        "tqdm; extra == \"dev\"",
+        "bitsandbytes; extra == \"dev\"",
+        "timm; extra == \"dev\"",
+        "rich; extra == \"dev\"",
+        "sagemaker; extra == \"sagemaker\""
+      ],
+      "installer": "pip"
+    },
+    "qwen-vl-utils": {
+      "version": "0.0.14",
+      "requires_dist": [
+        "av",
+        "packaging",
+        "pillow",
+        "requests",
+        "decord; extra == 'decord'"
+      ],
+      "installer": "pip"
+    },
+    "torchaudio": {
+      "version": "2.10.0",
+      "requires_dist": [
+        "torch==2.10.0"
+      ],
+      "installer": "pip"
+    },
+    "tokenizers": {
+      "version": "0.22.2",
+      "requires_dist": [
+        "huggingface-hub>=0.16.4,<2.0",
+        "pytest ; extra == 'testing'",
+        "pytest-asyncio ; extra == 'testing'",
+        "requests ; extra == 'testing'",
+        "numpy ; extra == 'testing'",
+        "datasets ; extra == 'testing'",
+        "ruff ; extra == 'testing'",
+        "ty ; extra == 'testing'",
+        "sphinx ; extra == 'docs'",
+        "sphinx-rtd-theme ; extra == 'docs'",
+        "setuptools-rust ; extra == 'docs'",
+        "tokenizers[testing] ; extra == 'dev'"
+      ],
+      "installer": "pip"
+    },
+    "numpy": {
+      "version": "2.2.6",
+      "requires_dist": [],
+      "installer": "pip"
+    },
+    "opencv-python-headless": {
+      "version": "5.0.0.93",
+      "requires_dist": [
+        "numpy<2.0; python_version < \"3.9\"",
+        "numpy>=2; python_version >= \"3.9\""
+      ],
+      "installer": "pip"
+    },
+    "vllm": {
+      "version": "0.19.0",
+      "requires_dist": [
+        "regex",
+        "cachetools",
+        "psutil",
+        "sentencepiece",
+        "numpy",
+        "requests>=2.26.0",
+        "tqdm",
+        "blake3",
+        "py-cpuinfo",
+        "transformers<5,>=4.56.0",
+        "tokenizers>=0.21.1",
+        "protobuf!=6.30.*,!=6.31.*,!=6.32.*,!=6.33.0.*,!=6.33.1.*,!=6.33.2.*,!=6.33.3.*,!=6.33.4.*,>=5.29.6",
+        "fastapi[standard]>=0.115.0",
+        "aiohttp>=3.13.3",
+        "openai>=2.0.0",
+        "pydantic>=2.12.0",
+        "prometheus_client>=0.18.0",
+        "pillow",
+        "prometheus-fastapi-instrumentator>=7.0.0",
+        "tiktoken>=0.6.0",
+        "lm-format-enforcer==0.11.3",
+        "llguidance<1.4.0,>=1.3.0; platform_machine == \"x86_64\" or platform_machine == \"arm64\" or platform_machine == \"aarch64\" or platform_machine == \"s390x\" or platform_machine == \"ppc64le\"",
+        "outlines_core==0.2.11",
+        "diskcache==5.6.3",
+        "lark==1.2.2",
+        "xgrammar<1.0.0,>=0.1.32; platform_machine == \"x86_64\" or platform_machine == \"aarch64\" or platform_machine == \"arm64\" or platform_machine == \"s390x\" or platform_machine == \"ppc64le\"",
+        "typing_extensions>=4.10",
+        "filelock>=3.16.1",
+        "partial-json-parser",
+        "pyzmq>=25.0.0",
+        "msgspec",
+        "gguf>=0.17.0",
+        "mistral_common[image]>=1.10.0",
+        "opencv-python-headless>=4.13.0",
+        "pyyaml",
+        "six>=1.16.0; python_version > \"3.11\"",
+        "setuptools<81.0.0,>=77.0.3; python_version > \"3.11\"",
+        "einops",
+        "compressed-tensors==0.14.0.1",
+        "depyf==0.20.0",
+        "cloudpickle",
+        "watchfiles",
+        "python-json-logger",
+        "ninja",
+        "pybase64",
+        "cbor2",
+        "ijson",
+        "setproctitle",
+        "openai-harmony>=0.0.3",
+        "anthropic>=0.71.0",
+        "model-hosting-container-standards<1.0.0,>=0.1.13",
+        "mcp",
+        "opentelemetry-sdk>=1.27.0",
+        "opentelemetry-api>=1.27.0",
+        "opentelemetry-exporter-otlp>=1.27.0",
+        "opentelemetry-semantic-conventions-ai>=0.4.1",
+        "numba==0.61.2",
+        "torch==2.10.0",
+        "torchaudio==2.10.0",
+        "torchvision==0.25.0",
+        "flashinfer-python==0.6.6",
+        "flashinfer-cubin==0.6.6",
+        "nvidia-cudnn-frontend<1.19.0,>=1.13.0",
+        "nvidia-cutlass-dsl>=4.4.0.dev1",
+        "quack-kernels>=0.2.7",
+        "zentorch; extra == \"zen\"",
+        "pandas; extra == \"bench\"",
+        "matplotlib; extra == \"bench\"",
+        "seaborn; extra == \"bench\"",
+        "datasets; extra == \"bench\"",
+        "scipy; extra == \"bench\"",
+        "plotly; extra == \"bench\"",
+        "tensorizer==2.10.1; extra == \"tensorizer\"",
+        "fastsafetensors>=0.2.2; extra == \"fastsafetensors\"",
+        "instanttensor>=0.1.5; extra == \"instanttensor\"",
+        "runai-model-streamer[azure,gcs,s3]>=0.15.7; extra == \"runai\"",
+        "av; extra == \"audio\"",
+        "resampy; extra == \"audio\"",
+        "scipy; extra == \"audio\"",
+        "soundfile; extra == \"audio\"",
+        "mistral_common[audio]; extra == \"audio\"",
+        "petit-kernel; extra == \"petit-kernel\"",
+        "helion==0.3.2; extra == \"helion\"",
+        "smg-grpc-servicer[vllm]>=0.5.0; extra == \"grpc\"",
+        "opentelemetry-sdk>=1.26.0; extra == \"otel\"",
+        "opentelemetry-api>=1.26.0; extra == \"otel\"",
+        "opentelemetry-exporter-otlp>=1.26.0; extra == \"otel\"",
+        "opentelemetry-semantic-conventions-ai>=0.4.1; extra == \"otel\""
+      ],
+      "installer": "pip"
+    },
+    "torch": {
+      "version": "2.10.0",
+      "requires_dist": [
+        "filelock",
+        "typing-extensions>=4.10.0",
+        "setuptools; python_version >= \"3.12\"",
+        "sympy>=1.13.3",
+        "networkx>=2.5.1",
+        "jinja2",
+        "fsspec>=0.8.5",
+        "cuda-bindings==12.9.4; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cuda-nvrtc-cu12==12.8.93; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cuda-runtime-cu12==12.8.90; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cuda-cupti-cu12==12.8.90; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cudnn-cu12==9.10.2.21; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cublas-cu12==12.8.4.1; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cufft-cu12==11.3.3.83; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-curand-cu12==10.3.9.90; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cusolver-cu12==11.7.3.90; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cusparse-cu12==12.5.8.93; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cusparselt-cu12==0.7.1; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-nccl-cu12==2.27.5; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-nvshmem-cu12==3.4.5; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-nvtx-cu12==12.8.90; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-nvjitlink-cu12==12.8.93; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "nvidia-cufile-cu12==1.13.1.3; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "triton==3.6.0; platform_system == \"Linux\" and platform_machine == \"x86_64\"",
+        "optree>=0.13.0; extra == \"optree\"",
+        "opt-einsum>=3.3; extra == \"opt-einsum\"",
+        "pyyaml; extra == \"pyyaml\""
+      ],
+      "installer": "pip"
+    },
+    "scipy": {
+      "version": "1.18.1",
+      "requires_dist": [
+        "numpy<2.8,>=2.0.0",
+        "pytest>=8.0.0; extra == \"test\"",
+        "pytest-cov; extra == \"test\"",
+        "pytest-timeout; extra == \"test\"",
+        "pytest-xdist; extra == \"test\"",
+        "asv; extra == \"test\"",
+        "mpmath; extra == \"test\"",
+        "gmpy2; extra == \"test\"",
+        "threadpoolctl; extra == \"test\"",
+        "scikit-umfpack; extra == \"test\"",
+        "pooch; extra == \"test\"",
+        "hypothesis>=6.30; extra == \"test\"",
+        "array-api-strict>=2.3.1; extra == \"test\"",
+        "Cython; extra == \"test\"",
+        "meson; extra == \"test\"",
+        "ninja; sys_platform != \"emscripten\" and extra == \"test\"",
+        "scipy-doctest>=2.0.0; extra == \"test\"",
+        "sphinx<8.2.0,>=5.0.0; extra == \"doc\"",
+        "intersphinx_registry; extra == \"doc\"",
+        "pydata-sphinx-theme>=0.15.2; extra == \"doc\"",
+        "sphinx-copybutton; extra == \"doc\"",
+        "sphinx-design>=0.4.0; extra == \"doc\"",
+        "matplotlib>=3.5; extra == \"doc\"",
+        "numpydoc; extra == \"doc\"",
+        "jupytext; extra == \"doc\"",
+        "myst-nb>=1.2.0; extra == \"doc\"",
+        "pooch; extra == \"doc\"",
+        "jupyterlite-sphinx>=0.19.1; extra == \"doc\"",
+        "jupyterlite-pyodide-kernel; extra == \"doc\"",
+        "linkify-it-py; extra == \"doc\"",
+        "tabulate; extra == \"doc\"",
+        "click<8.3.0; extra == \"dev\"",
+        "spin; extra == \"dev\"",
+        "mypy==1.19.1; extra == \"dev\"",
+        "pyrefly==0.63.0; extra == \"dev\"",
+        "typing_extensions; extra == \"dev\"",
+        "types-psutil; extra == \"dev\"",
+        "pycodestyle; extra == \"dev\"",
+        "ruff>=0.12.0; extra == \"dev\"",
+        "cython-lint>=0.12.2; extra == \"dev\""
+      ],
+      "installer": "pip"
+    },
+    "safetensors": {
+      "version": "0.8.0",
+      "requires_dist": [
+        "safetensors[torch] ; extra == 'all'",
+        "safetensors[numpy] ; extra == 'all'",
+        "safetensors[jax] ; extra == 'all'",
+        "safetensors[paddlepaddle] ; extra == 'all'",
+        "safetensors[convert] ; extra == 'all'",
+        "safetensors[quality] ; extra == 'all'",
+        "safetensors[testing] ; extra == 'all'",
+        "safetensors[torch] ; extra == 'convert'",
+        "huggingface-hub>=1.4 ; extra == 'convert'",
+        "safetensors[all] ; extra == 'dev'",
+        "safetensors[pinned-tf] ; extra == 'dev'",
+        "safetensors[numpy] ; extra == 'jax'",
+        "flax>=0.6.3 ; extra == 'jax'",
+        "jax>=0.3.25 ; extra == 'jax'",
+        "jaxlib>=0.3.25 ; extra == 'jax'",
+        "mlx>=0.0.9 ; extra == 'mlx'",
+        "numpy>=1.24.6 ; extra == 'numpy'",
+        "safetensors[numpy] ; extra == 'paddlepaddle'",
+        "paddlepaddle>=2.4.1 ; extra == 'paddlepaddle'",
+        "safetensors[numpy] ; extra == 'pinned-tf'",
+        "tensorflow==2.18.0 ; extra == 'pinned-tf'",
+        "ruff ; extra == 'quality'",
+        "safetensors[numpy] ; extra == 'tensorflow'",
+        "tensorflow>=2.11.0 ; extra == 'tensorflow'",
+        "safetensors[numpy] ; extra == 'testing'",
+        "h5py>=3.7.0 ; extra == 'testing'",
+        "setuptools-rust>=1.12.0 ; extra == 'testing'",
+        "pytest>=9.0 ; extra == 'testing'",
+        "pytest-benchmark>=5.2 ; extra == 'testing'",
+        "hypothesis>=6.70.2 ; extra == 'testing'",
+        "fsspec>=2024.6.0 ; extra == 'testing'",
+        "s3fs>=2024.6.0 ; extra == 'testing'",
+        "safetensors[numpy] ; extra == 'tf-nightly'",
+        "tf-nightly ; extra == 'tf-nightly'",
+        "safetensors[numpy] ; extra == 'torch'",
+        "torch>=2.4 ; extra == 'torch'"
+      ],
+      "installer": "pip"
+    }
+  },
+  "module_presence": {
+    "torch": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "d6388b1e3b16c1f9a22a87898e5f205fa13a97bd65c163f4829a157e0d9d602f"
+    },
+    "torchvision": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "ee2c9f4110cf1203db48c42601607329ac1f19709fa91c152f8d95eb53437a73"
+    },
+    "torchaudio": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "f0e07710f18256217b2e0056cbf6d4c99505e872542296d323ca2e446c27ea55"
+    },
+    "transformers": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "aa8cb54da488cf43ba37b9955bb6c0d84d2000db0306f6a5f1c20b6842482d04"
+    },
+    "tokenizers": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "148ecb122f3fee03be9abb1fe213dd85fbc691bf2ad2b437ff531533781a6a3a"
+    },
+    "huggingface_hub": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "e586b64612482128fd24119364bc4dc6dca938407134e454a5e876341b364e35"
+    },
+    "diffusers": {
+      "found": false
+    },
+    "diffsynth": {
+      "found": false
+    },
+    "accelerate": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "79eabeeaf005b8b0f98dc833610e8c073f33e9653f3630c4ed22c30808052d27"
+    },
+    "peft": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "69de31a0c28be2a6e22d48a0cff469c8e98a67eece5a0bfccc37e52cee004751"
+    },
+    "swift": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "d345fd8f68077d11730ffe56747a52b1858550e8c21067ed55a1db2f79ab5caf"
+    },
+    "verl": {
+      "found": false
+    },
+    "vllm": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "fd3708e5a13abe98c566afd79a6b1987cef70c0548bb19c1218ff6a7f9d43346"
+    },
+    "fla": {
+      "found": false
+    },
+    "tilelang": {
+      "found": false
+    },
+    "triton": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "96e68192cab5dd11fa30d5dc95d56c29ceecda53477fc3af8f37bee026b7a021"
+    },
+    "cv2": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "69192904dead3ada37c8b305ee850ac7474924730881a9d5e0b2390944eb5aae"
+    },
+    "numpy": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "6ae17b070c0f70a8e3cad89a510a256942e5a1f37ea5feb120cec167ed2a6236"
+    }
+  },
+  "pip_check": {
+    "exit_code": 1,
+    "output": "decord 0.6.0 is not supported on this platform\n"
+  },
+  "GPU": {
+    "checked": true,
+    "basic_tensor_check": {
+      "exit_code": 0,
+      "output": "{\"torch\": \"2.10.0+cu128\", \"cuda_runtime\": \"12.8\", \"cuda_available\": true, \"arch_list\": [\"sm_70\", \"sm_75\", \"sm_80\", \"sm_86\", \"sm_90\", \"sm_100\", \"sm_120\"], \"gpu\": \"NVIDIA RPBZZZ6\", \"capability\": [12, 0], \"tiny_bfloat16_result\": 4.0}\n"
+    },
+    "training_or_model_inference_run": false
+  },
+  "original_environment_modified": false,
+  "credentials_included": false
+}

egoexo/captioner_and_phiseg/import_validation.json

@@ -0,0 +1,53 @@
+{
+  "torch": {
+    "ok": true,
+    "version": "2.10.0+cu128"
+  },
+  "torchvision": {
+    "ok": true,
+    "version": "0.25.0+cu128"
+  },
+  "torchaudio": {
+    "ok": true,
+    "version": "2.10.0+cu128"
+  },
+  "transformers": {
+    "ok": true,
+    "version": "4.57.6"
+  },
+  "peft": {
+    "ok": true,
+    "version": "0.19.1"
+  },
+  "accelerate": {
+    "ok": true,
+    "version": "1.14.0"
+  },
+  "safetensors": {
+    "ok": true,
+    "version": "0.8.0"
+  },
+  "decord": {
+    "ok": true,
+    "version": "0.6.0"
+  },
+  "vllm": {
+    "ok": true,
+    "version": "0.19.0"
+  },
+  "verl": {
+    "ok": false,
+    "error_type": "ModuleNotFoundError",
+    "error": "No module named 'verl'"
+  },
+  "fla": {
+    "ok": false,
+    "error_type": "ModuleNotFoundError",
+    "error": "No module named 'fla'"
+  },
+  "tilelang": {
+    "ok": false,
+    "error_type": "ModuleNotFoundError",
+    "error": "No module named 'tilelang'"
+  }
+}

egoexo/captioner_and_phiseg/requirements.core.in

@@ -0,0 +1,24 @@
+# Observed role-specific root requirements; transitive dependencies resolved/locked separately.
+torch==2.10.0+cu128
+torchvision==0.25.0+cu128
+torchaudio==2.10.0+cu128
+triton==3.6.0
+transformers==4.57.6
+tokenizers==0.22.2
+huggingface-hub==0.36.2
+accelerate==1.14.0
+ms-swift==4.4.2
+deepspeed==0.19.1
+vllm==0.19.0
+peft==0.19.1
+safetensors==0.8.0
+qwen-vl-utils==0.0.14
+numpy==2.2.6
+pillow==11.3.0
+pandas==2.3.3
+pyarrow==25.0.1
+opencv-python-headless==5.0.0.93
+scipy==1.18.1
+einops==0.8.2
+datasets==4.8.4
+tqdm==4.70.0

egoexo/captioner_and_phiseg/requirements.observed-full.lock.txt

@@ -0,0 +1,262 @@
+absl-py==2.5.0
+accelerate==1.14.0
+addict==2.4.0
+agent-detector==1.1.0
+aiofiles==24.1.0
+aiohappyeyeballs==2.7.1
+aiohttp==3.14.3
+aiosignal==1.4.0
+aliyun-python-sdk-core==2.16.1
+aliyun-python-sdk-kms==2.16.5
+annotated-doc==0.0.5
+annotated-types==0.8.0
+anthropic==1.2.0
+anyio==4.14.2
+apache-tvm-ffi==0.1.13.post3
+astor==0.8.1
+attrdict==2.0.1
+attrs==26.1.0
+audioread==3.1.0
+av==18.1.0
+binpacking==2.0.1
+blake3==1.0.9
+brotli==1.2.0
+cachetools==7.1.8
+cbor2==6.1.4
+certifi==2026.7.22
+cffi==2.1.1
+charset-normalizer==3.5.1
+click==8.5.0
+cloudpickle==3.1.2
+compressed-tensors==0.14.0.1
+contourpy==1.3.3
+cpm-kernels==1.0.11
+crcmod==1.7
+cryptography==50.0.1
+cuda-bindings==12.9.4
+cuda-pathfinder==1.8.0
+cuda-python==12.9.4
+cycler==0.12.1
+dacite==1.9.2
+datasets==4.8.4
+decorator==5.3.1
+decord==0.6.0
+deepspeed==0.19.1
+defusedxml==0.7.1
+depyf==0.20.0
+detect-installer==0.1.0
+dill==0.4.1
+diskcache==5.6.3
+dnspython==2.8.0
+docstring-parser==0.18.0
+einops==0.8.2
+email-validator==2.3.0
+fastapi==0.141.1
+fastapi-cli==0.0.32
+fastapi-cloud-cli==0.24.0
+fastar==0.12.0
+ffmpy==1.0.0
+filelock==3.32.5
+flashinfer-cubin==0.6.6
+flashinfer-python==0.6.6
+fonttools==4.64.0
+frozenlist==1.8.0
+fsspec==2026.2.0
+ftfy==6.3.1
+gguf==0.19.0
+googleapis-common-protos==1.75.2
+gradio==5.50.0
+gradio-client==1.14.0
+groovy==0.1.2
+grpcio==1.83.1
+h11==0.16.0
+hf-xet==1.6.0
+hjson==3.1.0
+httpcore==1.0.9
+httpcore2==2.12.0
+httptools==0.8.0
+httpx==0.28.1
+httpx2==2.12.0
+huggingface-hub==0.36.2
+idna==3.19
+ijson==3.5.1
+imageio==2.37.4
+imageio-ffmpeg==0.6.0
+importlib-metadata==9.0.1
+interegular==0.3.3
+jinja2==3.1.6
+jiter==0.16.0
+jmespath==1.1.0
+joblib==1.6.0
+json-repair==0.63.4
+jsonschema==4.26.0
+jsonschema-specifications==2025.9.1
+kiwisolver==1.5.1
+lark==1.2.2
+lazy-loader==0.5
+librosa==0.10.2.post1
+llguidance==1.3.0
+llvmlite==0.44.0
+lm-format-enforcer==0.11.3
+loguru==0.7.3
+markdown==3.10.3
+markdown-it-py==4.2.0
+markupsafe==3.0.3
+matplotlib==3.11.1
+mcp==2.1.1
+mcp-types==2.1.1
+mdurl==0.1.2
+mistral-common==1.11.7
+model-hosting-container-standards==0.1.16
+modelscope==1.39.1
+modelscope-hub==0.3.1
+mpmath==1.3.0
+ms-swift==4.4.2
+msgpack==1.2.2
+msgspec==0.21.1
+multidict==6.7.1
+multiprocess==0.70.19
+narwhals==2.25.0
+networkx==3.6.1
+ninja==1.13.2
+nltk==3.10.3
+numba==0.61.2
+numpy==2.2.6
+nvidia-cublas-cu12==12.8.4.1
+nvidia-cuda-cupti-cu12==12.8.90
+nvidia-cuda-nvdisasm==13.3.73
+nvidia-cuda-nvrtc-cu12==12.8.93
+nvidia-cuda-runtime-cu12==12.8.90
+nvidia-cudnn-cu12==9.10.2.21
+nvidia-cudnn-frontend==1.18.0
+nvidia-cufft-cu12==11.3.3.83
+nvidia-cufile-cu12==1.13.1.3
+nvidia-curand-cu12==10.3.9.90
+nvidia-cusolver-cu12==11.7.3.90
+nvidia-cusparse-cu12==12.5.8.93
+nvidia-cusparselt-cu12==0.7.1
+nvidia-cutlass-dsl==4.8.0.dev0
+nvidia-cutlass-dsl-libs-base==4.8.0.dev0
+nvidia-cutlass-dsl-libs-core==4.8.0.dev0
+nvidia-cutlass-dsl-libs-cu12==4.8.0.dev0
+nvidia-ml-py==13.610.43
+nvidia-nccl-cu12==2.27.5
+nvidia-nvjitlink-cu12==12.8.93
+nvidia-nvshmem-cu12==3.4.5
+nvidia-nvtx-cu12==12.8.90
+openai==3.6.0
+openai-harmony==0.0.8
+opencv-python-headless==5.0.0.93
+opentelemetry-api==1.44.0
+opentelemetry-exporter-otlp==1.44.0
+opentelemetry-exporter-otlp-proto-common==1.44.0
+opentelemetry-exporter-otlp-proto-grpc==1.44.0
+opentelemetry-exporter-otlp-proto-http==1.44.0
+opentelemetry-proto==1.44.0
+opentelemetry-sdk==1.44.0
+opentelemetry-semantic-conventions==0.65b0
+opentelemetry-semantic-conventions-ai==0.5.1
+orjson==3.12.0
+oss2==2.19.1
+outlines-core==0.2.11
+packaging==26.3
+pandas==2.3.3
+partial-json-parser==0.2.1.1.post7
+peft==0.19.1
+pillow==11.3.0
+pip==26.2.1
+platformdirs==4.11.7
+pooch==1.9.0
+prometheus-client==0.26.0
+prometheus-fastapi-instrumentator==7.1.0
+propcache==0.5.2
+protobuf==7.36.1
+psutil==7.2.2
+py-cpuinfo==9.0.0
+pyarrow==25.0.1
+pybase64==1.5.0
+pycountry==26.2.16
+pycparser==3.0
+pycryptodome==3.23.0
+pydantic==2.12.3
+pydantic-core==2.41.4
+pydantic-extra-types==2.11.1
+pydantic-settings==2.15.0
+pydub==0.25.1
+pygments==2.21.0
+pyjwt==2.13.0
+pyparsing==3.3.2
+python-dateutil==2.9.0.post0
+python-dotenv==1.2.3
+python-json-logger==4.2.0
+python-multipart==0.0.32
+pytz==2026.3.post1
+pyyaml==6.0.3
+pyzmq==27.2.0
+quack-kernels==0.5.0
+qwen-omni-utils==0.0.9
+qwen-vl-utils==0.0.14
+referencing==0.37.0
+regex==2026.9.3
+requests==2.34.2
+rich==15.0.0
+rich-toolkit==0.20.3
+rignore==0.8.1
+rouge==1.0.1
+rpds-py==2026.6.3
+ruff==0.16.5
+safehttpx==0.1.7
+safetensors==0.8.0
+scikit-learn==1.9.0
+scipy==1.18.1
+semantic-version==2.10.0
+sentencepiece==0.2.2
+sentry-sdk==2.68.1
+setproctitle==1.3.7
+setuptools==84.0.0
+shellingham==1.5.4
+simplejson==4.1.2
+six==1.17.0
+sniffio==1.3.1
+sortedcontainers==2.4.0
+soundfile==0.14.0
+soxr==1.1.0
+sse-starlette==3.4.8
+starlette==0.52.1
+supervisor==4.3.0
+sympy==1.14.0
+tabulate==0.10.0
+tensorboard==2.21.0
+tensorboard-data-server==0.7.2
+threadpoolctl==3.6.0
+tiktoken==0.14.0
+tokenizers==0.22.2
+tomlkit==0.13.3
+torch==2.10.0
+torch-c-dlpack-ext==0.1.5
+torchaudio==2.10.0
+torchvision==0.25.0
+tqdm==4.70.0
+transformers==4.57.6
+transformers-stream-generator==0.0.5
+triton==3.6.0
+trl==0.29.1
+truststore==0.10.4
+typer==0.27.2
+typing-extensions==4.16.0
+typing-inspection==0.4.4
+tzdata==2026.3
+urllib3==2.7.0
+uvicorn==0.52.4
+uvloop==0.22.1
+vllm==0.19.0
+watchfiles==1.2.0
+wcwidth==0.8.3
+websockets==15.0.1
+werkzeug==3.1.8
+wheel==0.48.0
+xgrammar==0.2.3
+xxhash==4.0.1
+yarl==1.24.5
+zipp==4.1.0
+zstandard==0.25.0

egoexo/generator/conda-explicit-linux-64.txt

@@ -0,0 +1,33 @@
+@EXPLICIT
+https://repo.anaconda.com/pkgs/main/linux-64/_libgcc_mutex-0.1-main.conda#c3473ff8bdb3d124ed5ff11ec380d6f9
+https://repo.anaconda.com/pkgs/main/linux-64/_openmp_mutex-5.1-52_gnu.conda#8c74a3d91bfc4b484c91ed7201f84fba
+https://repo.anaconda.com/pkgs/main/linux-64/bzip2-1.0.8-h5eee18b_6.conda#f21a3ff51c1b271977f53ce956a69297
+https://repo.anaconda.com/pkgs/main/linux-64/ca-certificates-2026.5.14-h06a4308_0.conda#16c8384e6e546a181720f7f2f2da717b
+https://repo.anaconda.com/pkgs/main/linux-64/ld_impl_linux-64-2.44-h9e0c5a2_3.conda#c563bb71f4df90fc3b92cde204827564
+https://repo.anaconda.com/pkgs/main/linux-64/libexpat-2.8.2-h7354ed3_1.conda#bcf2aaf1e184a7b1311418e3e62f650e
+https://repo.anaconda.com/pkgs/main/linux-64/libffi-3.4.8-h06d3fd0_3.conda#cc7c0dee0dd82e6cb7ac6bc41abb68ea
+https://repo.anaconda.com/pkgs/main/linux-64/libgcc-15.2.0-h69a1729_8.conda#44a53b93cf0953bde4b0f4e610cef1c3
+https://repo.anaconda.com/pkgs/main/linux-64/libgcc-ng-15.2.0-h166f726_8.conda#b3f01816cd1e046648fff257be7b0248
+https://repo.anaconda.com/pkgs/main/linux-64/libnsl-2.0.0-h5eee18b_0.conda#55e3f1b0c2a70454cdf1a5d64fb44752
+https://repo.anaconda.com/pkgs/main/linux-64/libstdcxx-15.2.0-h39759b7_8.conda#157776d72d1eba48854127f461b729cc
+https://repo.anaconda.com/pkgs/main/linux-64/libuuid-1.41.5-h5eee18b_0.conda#4a6a2354414c9080327274aa514e5299
+https://repo.anaconda.com/pkgs/main/linux-64/libxcb-1.17.0-h9b100fa_0.conda#fdf0d380fa3809a301e2dbc0d5183883
+https://repo.anaconda.com/pkgs/main/linux-64/libzlib-1.3.2-h47b2149_0.conda#d78e716b8773cada946b76eaa4e6a471
+https://repo.anaconda.com/pkgs/main/linux-64/ncurses-6.5-h7934f7d_0.conda#0abfc090299da4bb031b84c64309757b
+https://repo.anaconda.com/pkgs/main/linux-64/openssl-3.5.7-h1b28b03_0.conda#ae75e695171e5244c01c5548e2d72595
+https://repo.anaconda.com/pkgs/main/linux-64/packaging-26.0-py311h06a4308_0.conda#3036509f70fd50149c31a1905fb712c6
+https://repo.anaconda.com/pkgs/main/noarch/pip-26.1.2-pyhc872135_0.conda#679b1339b53c0bbc44960f048b5a7b80
+https://repo.anaconda.com/pkgs/main/linux-64/pthread-stubs-0.3-h0ce48e5_1.conda#973a642312d2a28927aaf5b477c67250
+https://repo.anaconda.com/pkgs/main/linux-64/python-3.11.15-h17756b0_1.conda#940d282fb8c4a925669c399f44691916
+https://repo.anaconda.com/pkgs/main/linux-64/readline-8.3-hc2a1206_0.conda#8578e006d4ef5cb98a6cda232b3490f6
+https://repo.anaconda.com/pkgs/main/linux-64/setuptools-82.0.1-py311h06a4308_0.conda#64536881db87c7133c0a95f05c1063f5
+https://repo.anaconda.com/pkgs/main/linux-64/sqlite-3.53.2-h795bf6d_0.conda#1b4196e3df4c99c2df5a98e5d5b912a7
+https://repo.anaconda.com/pkgs/main/linux-64/tk-8.6.15-h54e0aa7_0.conda#1fa91e0c4fc9c9435eda3f1a25a676fd
+https://repo.anaconda.com/pkgs/main/noarch/tzdata-2026b-he532380_0.conda#805dd0dbe5ecde428f3c6c1afb6d94e2
+https://repo.anaconda.com/pkgs/main/linux-64/wheel-0.47.0-py311h06a4308_0.conda#c799fc602bbe1ca25c465839ba76da9f
+https://repo.anaconda.com/pkgs/main/linux-64/xorg-libx11-1.8.12-h9b100fa_1.conda#6298b27afae6f49f03765b2a03df2fcb
+https://repo.anaconda.com/pkgs/main/linux-64/xorg-libxau-1.0.12-h9b100fa_0.conda#a8005a9f6eb903e113cd5363e8a11459
+https://repo.anaconda.com/pkgs/main/linux-64/xorg-libxdmcp-1.1.5-h9b100fa_0.conda#c284a09ddfba81d9c4e740110f09ea06
+https://repo.anaconda.com/pkgs/main/linux-64/xorg-xorgproto-2024.1-h47b2149_2.conda#eea4ccf000bb5092da6dc999abf66377
+https://repo.anaconda.com/pkgs/main/linux-64/xz-5.8.2-h448239c_0.conda#3cf8f0b9f45cd775a7ee57ede84f827f
+https://repo.anaconda.com/pkgs/main/linux-64/zlib-1.3.2-h47b2149_0.conda#95257497fc7a2390d61d46eebd18bdb5

egoexo/generator/environment.snapshot.json

@@ -0,0 +1,1318 @@
+{
+  "profile": "egoexo_generator",
+  "capture_scope": "observed current retained runtime; not proof of original launch-time environment",
+  "python": "3.11.15",
+  "implementation": "CPython",
+  "SOABI": "cpython-311-x86_64-linux-gnu",
+  "machine": "x86_64",
+  "platform": "Linux",
+  "libc": [
+    "glibc",
+    "2.35"
+  ],
+  "packages": {
+    "absl-py": "2.5.0",
+    "accelerate": "1.14.0",
+    "aiohappyeyeballs": "2.7.1",
+    "aiohttp": "3.14.1",
+    "aiosignal": "1.4.0",
+    "antlr4-python3-runtime": "4.9.3",
+    "attrs": "26.1.0",
+    "cd-fvd": "0.1.1",
+    "certifi": "2026.6.17",
+    "charset-normalizer": "3.4.8",
+    "datasets": "4.0.0",
+    "decord": "0.6.0",
+    "diffusers": "0.34.0",
+    "dill": "0.3.8",
+    "einops": "0.8.1",
+    "filelock": "3.29.5",
+    "frozenlist": "1.8.0",
+    "fsspec": "2025.3.0",
+    "ftfy": "6.3.1",
+    "grpcio": "1.83.0",
+    "hf-xet": "1.5.1",
+    "huggingface-hub": "0.36.2",
+    "hydra-core": "1.3.4",
+    "idna": "3.18",
+    "imageio": "2.37.0",
+    "imageio-ffmpeg": "0.6.0",
+    "importlib-metadata": "9.0.0",
+    "iopath": "0.1.10",
+    "jinja2": "3.1.6",
+    "lazy-loader": "0.5",
+    "lightning-utilities": "0.15.3",
+    "loguru": "0.7.3",
+    "lpips": "0.1.4",
+    "markdown": "3.10.3",
+    "markupsafe": "3.0.3",
+    "modelscope": "1.36.3",
+    "mpmath": "1.3.0",
+    "multidict": "6.7.1",
+    "multiprocess": "0.70.16",
+    "networkx": "3.6.1",
+    "numpy": "2.2.6",
+    "nvidia-cublas": "13.0.0.19",
+    "nvidia-cublas-cu12": "12.6.4.1",
+    "nvidia-cuda-cupti": "13.0.48",
+    "nvidia-cuda-cupti-cu12": "12.6.80",
+    "nvidia-cuda-nvrtc": "13.0.48",
+    "nvidia-cuda-nvrtc-cu12": "12.6.77",
+    "nvidia-cuda-runtime": "13.0.48",
+    "nvidia-cuda-runtime-cu12": "12.6.77",
+    "nvidia-cudnn-cu12": "9.5.1.17",
+    "nvidia-cudnn-cu13": "9.13.0.50",
+    "nvidia-cufft": "12.0.0.15",
+    "nvidia-cufft-cu12": "11.3.0.4",
+    "nvidia-cufile": "1.15.0.42",
+    "nvidia-cufile-cu12": "1.11.1.6",
+    "nvidia-curand": "10.4.0.35",
+    "nvidia-curand-cu12": "10.3.7.77",
+    "nvidia-cusolver": "12.0.3.29",
+    "nvidia-cusolver-cu12": "11.7.1.2",
+    "nvidia-cusparse": "12.6.2.49",
+    "nvidia-cusparse-cu12": "12.5.4.2",
+    "nvidia-cusparselt-cu12": "0.6.3",
+    "nvidia-cusparselt-cu13": "0.8.0",
+    "nvidia-nccl-cu12": "2.26.2",
+    "nvidia-nccl-cu13": "2.27.7",
+    "nvidia-nvjitlink": "13.0.39",
+    "nvidia-nvjitlink-cu12": "12.6.85",
+    "nvidia-nvshmem-cu13": "3.3.24",
+    "nvidia-nvtx": "13.0.39",
+    "nvidia-nvtx-cu12": "12.6.77",
+    "omegaconf": "2.3.1",
+    "opencv-python": "4.12.0.88",
+    "opencv-python-headless": "4.12.0.88",
+    "packaging": "26.0",
+    "pandas": "2.3.2",
+    "peft": "0.19.1",
+    "pillow": "11.0.0",
+    "pip": "26.1.2",
+    "portalocker": "3.2.0",
+    "propcache": "0.5.2",
+    "protobuf": "6.32.0",
+    "psutil": "7.2.2",
+    "pyarrow": "24.0.0",
+    "pymupdf": "1.28.2",
+    "pypdf": "6.14.2",
+    "python-dateutil": "2.9.0.post0",
+    "pytz": "2026.2",
+    "pyyaml": "6.0.3",
+    "regex": "2026.6.28",
+    "reportlab": "5.0.1",
+    "requests": "2.34.2",
+    "safetensors": "0.6.2",
+    "sam-2": "1.0",
+    "scikit-image": "0.26.0",
+    "scipy": "1.17.1",
+    "sentencepiece": "0.2.0",
+    "setuptools": "82.0.1",
+    "six": "1.17.0",
+    "sympy": "1.14.0",
+    "tensorboard": "2.21.0",
+    "tensorboard-data-server": "0.7.2",
+    "tifffile": "2026.3.3",
+    "tokenizers": "0.21.4",
+    "torch": "2.9.1+cu130",
+    "torchaudio": "2.9.1+cu130",
+    "torchmetrics": "1.9.0",
+    "torchvision": "0.24.1+cu130",
+    "tqdm": "4.68.3",
+    "transformers": "4.54.0",
+    "triton": "3.5.1",
+    "typing-extensions": "4.16.0",
+    "tzdata": "2026.2",
+    "urllib3": "2.7.0",
+    "wcwidth": "0.8.2",
+    "werkzeug": "3.1.8",
+    "wheel": "0.47.0",
+    "xxhash": "3.8.1",
+    "yarl": "1.24.2",
+    "zipp": "4.1.0"
+  },
+  "core_package_provenance": {
+    "peft": {
+      "version": "0.19.1",
+      "requires_dist": [
+        "numpy>=1.17",
+        "packaging>=20.0",
+        "psutil",
+        "pyyaml",
+        "torch>=1.13.0",
+        "transformers",
+        "tqdm",
+        "accelerate>=0.21.0",
+        "safetensors",
+        "huggingface_hub>=0.25.0",
+        "black; extra == \"quality\"",
+        "hf-doc-builder; extra == \"quality\"",
+        "ruff~=0.12.8; extra == \"quality\"",
+        "black; extra == \"docs-specific\"",
+        "requests; extra == \"docs-specific\"",
+        "hf-doc-builder; extra == \"docs-specific\"",
+        "black; extra == \"dev\"",
+        "hf-doc-builder; extra == \"dev\"",
+        "ruff~=0.12.8; extra == \"dev\"",
+        "black; extra == \"dev\"",
+        "requests; extra == \"dev\"",
+        "hf-doc-builder; extra == \"dev\"",
+        "black; extra == \"test\"",
+        "hf-doc-builder; extra == \"test\"",
+        "ruff~=0.12.8; extra == \"test\"",
+        "black; extra == \"test\"",
+        "requests; extra == \"test\"",
+        "hf-doc-builder; extra == \"test\"",
+        "pytest; extra == \"test\"",
+        "pytest-cov; extra == \"test\"",
+        "pytest-xdist; extra == \"test\"",
+        "parameterized; extra == \"test\"",
+        "datasets; extra == \"test\"",
+        "diffusers; extra == \"test\"",
+        "scipy; extra == \"test\"",
+        "scikit-learn; extra == \"test\"",
+        "protobuf; extra == \"test\"",
+        "sentencepiece; extra == \"test\"",
+        "torchvision; extra == \"test\""
+      ],
+      "installer": "pip"
+    },
+    "safetensors": {
+      "version": "0.6.2",
+      "requires_dist": [
+        "numpy>=1.21.6 ; extra == 'numpy'",
+        "safetensors[numpy] ; extra == 'torch'",
+        "torch>=1.10 ; extra == 'torch'",
+        "safetensors[numpy] ; extra == 'tensorflow'",
+        "tensorflow>=2.11.0 ; extra == 'tensorflow'",
+        "safetensors[numpy] ; extra == 'pinned-tf'",
+        "tensorflow==2.18.0 ; extra == 'pinned-tf'",
+        "safetensors[numpy] ; extra == 'jax'",
+        "flax>=0.6.3 ; extra == 'jax'",
+        "jax>=0.3.25 ; extra == 'jax'",
+        "jaxlib>=0.3.25 ; extra == 'jax'",
+        "mlx>=0.0.9 ; extra == 'mlx'",
+        "safetensors[numpy] ; extra == 'paddlepaddle'",
+        "paddlepaddle>=2.4.1 ; extra == 'paddlepaddle'",
+        "ruff ; extra == 'quality'",
+        "safetensors[numpy] ; extra == 'testing'",
+        "h5py>=3.7.0 ; extra == 'testing'",
+        "huggingface-hub>=0.12.1 ; extra == 'testing'",
+        "setuptools-rust>=1.5.2 ; extra == 'testing'",
+        "pytest>=7.2.0 ; extra == 'testing'",
+        "pytest-benchmark>=4.0.0 ; extra == 'testing'",
+        "hypothesis>=6.70.2 ; extra == 'testing'",
+        "safetensors[numpy] ; extra == 'testingfree'",
+        "huggingface-hub>=0.12.1 ; extra == 'testingfree'",
+        "setuptools-rust>=1.5.2 ; extra == 'testingfree'",
+        "pytest>=7.2.0 ; extra == 'testingfree'",
+        "pytest-benchmark>=4.0.0 ; extra == 'testingfree'",
+        "hypothesis>=6.70.2 ; extra == 'testingfree'",
+        "safetensors[torch] ; extra == 'all'",
+        "safetensors[numpy] ; extra == 'all'",
+        "safetensors[pinned-tf] ; extra == 'all'",
+        "safetensors[jax] ; extra == 'all'",
+        "safetensors[paddlepaddle] ; extra == 'all'",
+        "safetensors[quality] ; extra == 'all'",
+        "safetensors[testing] ; extra == 'all'",
+        "safetensors[all] ; extra == 'dev'"
+      ],
+      "installer": "pip"
+    },
+    "triton": {
+      "version": "3.5.1",
+      "requires_dist": [
+        "importlib-metadata; python_version < \"3.10\"",
+        "cmake<4.0,>=3.20; extra == \"build\"",
+        "lit; extra == \"build\"",
+        "autopep8; extra == \"tests\"",
+        "isort; extra == \"tests\"",
+        "numpy; extra == \"tests\"",
+        "pytest; extra == \"tests\"",
+        "pytest-forked; extra == \"tests\"",
+        "pytest-xdist; extra == \"tests\"",
+        "scipy>=1.7.1; extra == \"tests\"",
+        "llnl-hatchet; extra == \"tests\"",
+        "matplotlib; extra == \"tutorials\"",
+        "pandas; extra == \"tutorials\"",
+        "tabulate; extra == \"tutorials\""
+      ],
+      "installer": "pip"
+    },
+    "scipy": {
+      "version": "1.17.1",
+      "requires_dist": [
+        "numpy<2.7,>=1.26.4",
+        "pytest>=8.0.0; extra == \"test\"",
+        "pytest-cov; extra == \"test\"",
+        "pytest-timeout; extra == \"test\"",
+        "pytest-xdist; extra == \"test\"",
+        "asv; extra == \"test\"",
+        "mpmath; extra == \"test\"",
+        "gmpy2; extra == \"test\"",
+        "threadpoolctl; extra == \"test\"",
+        "scikit-umfpack; extra == \"test\"",
+        "pooch; extra == \"test\"",
+        "hypothesis>=6.30; extra == \"test\"",
+        "array-api-strict>=2.3.1; extra == \"test\"",
+        "Cython; extra == \"test\"",
+        "meson; extra == \"test\"",
+        "ninja; sys_platform != \"emscripten\" and extra == \"test\"",
+        "sphinx<8.2.0,>=5.0.0; extra == \"doc\"",
+        "intersphinx_registry; extra == \"doc\"",
+        "pydata-sphinx-theme>=0.15.2; extra == \"doc\"",
+        "sphinx-copybutton; extra == \"doc\"",
+        "sphinx-design>=0.4.0; extra == \"doc\"",
+        "matplotlib>=3.5; extra == \"doc\"",
+        "numpydoc; extra == \"doc\"",
+        "jupytext; extra == \"doc\"",
+        "myst-nb>=1.2.0; extra == \"doc\"",
+        "pooch; extra == \"doc\"",
+        "jupyterlite-sphinx>=0.19.1; extra == \"doc\"",
+        "jupyterlite-pyodide-kernel; extra == \"doc\"",
+        "linkify-it-py; extra == \"doc\"",
+        "tabulate; extra == \"doc\"",
+        "click<8.3.0; extra == \"dev\"",
+        "spin; extra == \"dev\"",
+        "mypy==1.10.0; extra == \"dev\"",
+        "typing_extensions; extra == \"dev\"",
+        "types-psutil; extra == \"dev\"",
+        "pycodestyle; extra == \"dev\"",
+        "ruff>=0.12.0; extra == \"dev\"",
+        "cython-lint>=0.12.2; extra == \"dev\""
+      ],
+      "installer": "pip"
+    },
+    "pillow": {
+      "version": "11.0.0",
+      "requires_dist": [
+        "furo ; extra == 'docs'",
+        "olefile ; extra == 'docs'",
+        "sphinx >=8.1 ; extra == 'docs'",
+        "sphinx-copybutton ; extra == 'docs'",
+        "sphinx-inline-tabs ; extra == 'docs'",
+        "sphinxext-opengraph ; extra == 'docs'",
+        "olefile ; extra == 'fpx'",
+        "olefile ; extra == 'mic'",
+        "check-manifest ; extra == 'tests'",
+        "coverage ; extra == 'tests'",
+        "defusedxml ; extra == 'tests'",
+        "markdown2 ; extra == 'tests'",
+        "olefile ; extra == 'tests'",
+        "packaging ; extra == 'tests'",
+        "pyroma ; extra == 'tests'",
+        "pytest ; extra == 'tests'",
+        "pytest-cov ; extra == 'tests'",
+        "pytest-timeout ; extra == 'tests'",
+        "typing-extensions ; (python_version < \"3.10\") and extra == 'typing'",
+        "defusedxml ; extra == 'xmp'"
+      ],
+      "installer": "pip"
+    },
+    "huggingface-hub": {
+      "version": "0.36.2",
+      "requires_dist": [
+        "filelock",
+        "fsspec>=2023.5.0",
+        "hf-xet<2.0.0,>=1.1.3; platform_machine == \"x86_64\" or platform_machine == \"amd64\" or platform_machine == \"arm64\" or platform_machine == \"aarch64\"",
+        "packaging>=20.9",
+        "pyyaml>=5.1",
+        "requests",
+        "tqdm>=4.42.1",
+        "typing-extensions>=3.7.4.3",
+        "InquirerPy==0.3.4; extra == \"cli\"",
+        "aiohttp; extra == \"inference\"",
+        "authlib>=1.3.2; extra == \"oauth\"",
+        "fastapi; extra == \"oauth\"",
+        "httpx; extra == \"oauth\"",
+        "itsdangerous; extra == \"oauth\"",
+        "torch; extra == \"torch\"",
+        "safetensors[torch]; extra == \"torch\"",
+        "hf_transfer>=0.1.4; extra == \"hf-transfer\"",
+        "toml; extra == \"fastai\"",
+        "fastai>=2.4; extra == \"fastai\"",
+        "fastcore>=1.3.27; extra == \"fastai\"",
+        "tensorflow; extra == \"tensorflow\"",
+        "pydot; extra == \"tensorflow\"",
+        "graphviz; extra == \"tensorflow\"",
+        "tensorflow; extra == \"tensorflow-testing\"",
+        "keras<3.0; extra == \"tensorflow-testing\"",
+        "hf-xet<2.0.0,>=1.1.2; extra == \"hf-xet\"",
+        "mcp>=1.8.0; extra == \"mcp\"",
+        "typer; extra == \"mcp\"",
+        "aiohttp; extra == \"mcp\"",
+        "InquirerPy==0.3.4; extra == \"testing\"",
+        "aiohttp; extra == \"testing\"",
+        "authlib>=1.3.2; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "httpx; extra == \"testing\"",
+        "itsdangerous; extra == \"testing\"",
+        "jedi; extra == \"testing\"",
+        "Jinja2; extra == \"testing\"",
+        "pytest<8.2.2,>=8.1.1; extra == \"testing\"",
+        "pytest-cov; extra == \"testing\"",
+        "pytest-env; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-vcr; extra == \"testing\"",
+        "pytest-asyncio; extra == \"testing\"",
+        "pytest-rerunfailures<16.0; extra == \"testing\"",
+        "pytest-mock; extra == \"testing\"",
+        "urllib3<2.0; extra == \"testing\"",
+        "soundfile; extra == \"testing\"",
+        "Pillow; extra == \"testing\"",
+        "gradio>=4.0.0; extra == \"testing\"",
+        "numpy; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "typing-extensions>=4.8.0; extra == \"typing\"",
+        "types-PyYAML; extra == \"typing\"",
+        "types-requests; extra == \"typing\"",
+        "types-simplejson; extra == \"typing\"",
+        "types-toml; extra == \"typing\"",
+        "types-tqdm; extra == \"typing\"",
+        "types-urllib3; extra == \"typing\"",
+        "ruff>=0.9.0; extra == \"quality\"",
+        "mypy<1.15.0,>=1.14.1; python_version == \"3.8\" and extra == \"quality\"",
+        "mypy==1.15.0; python_version >= \"3.9\" and extra == \"quality\"",
+        "libcst>=1.4.0; extra == \"quality\"",
+        "ty; extra == \"quality\"",
+        "InquirerPy==0.3.4; extra == \"all\"",
+        "aiohttp; extra == \"all\"",
+        "authlib>=1.3.2; extra == \"all\"",
+        "fastapi; extra == \"all\"",
+        "httpx; extra == \"all\"",
+        "itsdangerous; extra == \"all\"",
+        "jedi; extra == \"all\"",
+        "Jinja2; extra == \"all\"",
+        "pytest<8.2.2,>=8.1.1; extra == \"all\"",
+        "pytest-cov; extra == \"all\"",
+        "pytest-env; extra == \"all\"",
+        "pytest-xdist; extra == \"all\"",
+        "pytest-vcr; extra == \"all\"",
+        "pytest-asyncio; extra == \"all\"",
+        "pytest-rerunfailures<16.0; extra == \"all\"",
+        "pytest-mock; extra == \"all\"",
+        "urllib3<2.0; extra == \"all\"",
+        "soundfile; extra == \"all\"",
+        "Pillow; extra == \"all\"",
+        "gradio>=4.0.0; extra == \"all\"",
+        "numpy; extra == \"all\"",
+        "fastapi; extra == \"all\"",
+        "ruff>=0.9.0; extra == \"all\"",
+        "mypy<1.15.0,>=1.14.1; python_version == \"3.8\" and extra == \"all\"",
+        "mypy==1.15.0; python_version >= \"3.9\" and extra == \"all\"",
+        "libcst>=1.4.0; extra == \"all\"",
+        "ty; extra == \"all\"",
+        "typing-extensions>=4.8.0; extra == \"all\"",
+        "types-PyYAML; extra == \"all\"",
+        "types-requests; extra == \"all\"",
+        "types-simplejson; extra == \"all\"",
+        "types-toml; extra == \"all\"",
+        "types-tqdm; extra == \"all\"",
+        "types-urllib3; extra == \"all\"",
+        "InquirerPy==0.3.4; extra == \"dev\"",
+        "aiohttp; extra == \"dev\"",
+        "authlib>=1.3.2; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "httpx; extra == \"dev\"",
+        "itsdangerous; extra == \"dev\"",
+        "jedi; extra == \"dev\"",
+        "Jinja2; extra == \"dev\"",
+        "pytest<8.2.2,>=8.1.1; extra == \"dev\"",
+        "pytest-cov; extra == \"dev\"",
+        "pytest-env; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-vcr; extra == \"dev\"",
+        "pytest-asyncio; extra == \"dev\"",
+        "pytest-rerunfailures<16.0; extra == \"dev\"",
+        "pytest-mock; extra == \"dev\"",
+        "urllib3<2.0; extra == \"dev\"",
+        "soundfile; extra == \"dev\"",
+        "Pillow; extra == \"dev\"",
+        "gradio>=4.0.0; extra == \"dev\"",
+        "numpy; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "ruff>=0.9.0; extra == \"dev\"",
+        "mypy<1.15.0,>=1.14.1; python_version == \"3.8\" and extra == \"dev\"",
+        "mypy==1.15.0; python_version >= \"3.9\" and extra == \"dev\"",
+        "libcst>=1.4.0; extra == \"dev\"",
+        "ty; extra == \"dev\"",
+        "typing-extensions>=4.8.0; extra == \"dev\"",
+        "types-PyYAML; extra == \"dev\"",
+        "types-requests; extra == \"dev\"",
+        "types-simplejson; extra == \"dev\"",
+        "types-toml; extra == \"dev\"",
+        "types-tqdm; extra == \"dev\"",
+        "types-urllib3; extra == \"dev\""
+      ],
+      "installer": "pip"
+    },
+    "opencv-python-headless": {
+      "version": "4.12.0.88",
+      "requires_dist": [
+        "numpy<2.0; python_version < \"3.9\"",
+        "numpy<2.3.0,>=2; python_version >= \"3.9\""
+      ],
+      "installer": "pip"
+    },
+    "pyarrow": {
+      "version": "24.0.0",
+      "requires_dist": [],
+      "installer": "pip"
+    },
+    "accelerate": {
+      "version": "1.14.0",
+      "requires_dist": [
+        "numpy>=1.17",
+        "packaging>=20.0",
+        "psutil",
+        "pyyaml",
+        "torch>=2.0.0",
+        "huggingface_hub>=0.21.0",
+        "safetensors>=0.4.3",
+        "ruff==0.13.1; extra == \"quality\"",
+        "pytest>=7.2.0; extra == \"test-prod\"",
+        "pytest-xdist; extra == \"test-prod\"",
+        "pytest-subtests; extra == \"test-prod\"",
+        "parameterized; extra == \"test-prod\"",
+        "pytest-order; extra == \"test-prod\"",
+        "datasets; extra == \"test-dev\"",
+        "diffusers; extra == \"test-dev\"",
+        "evaluate; extra == \"test-dev\"",
+        "torchdata>=0.8.0; extra == \"test-dev\"",
+        "torchpippy>=0.2.0; extra == \"test-dev\"",
+        "transformers; extra == \"test-dev\"",
+        "scipy; extra == \"test-dev\"",
+        "scikit-learn; extra == \"test-dev\"",
+        "tqdm; extra == \"test-dev\"",
+        "bitsandbytes; extra == \"test-dev\"",
+        "timm; extra == \"test-dev\"",
+        "pytest>=7.2.0; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-subtests; extra == \"testing\"",
+        "parameterized; extra == \"testing\"",
+        "pytest-order; extra == \"testing\"",
+        "datasets; extra == \"testing\"",
+        "diffusers; extra == \"testing\"",
+        "evaluate; extra == \"testing\"",
+        "torchdata>=0.8.0; extra == \"testing\"",
+        "torchpippy>=0.2.0; extra == \"testing\"",
+        "transformers; extra == \"testing\"",
+        "scipy; extra == \"testing\"",
+        "scikit-learn; extra == \"testing\"",
+        "tqdm; extra == \"testing\"",
+        "bitsandbytes; extra == \"testing\"",
+        "timm; extra == \"testing\"",
+        "deepspeed; extra == \"deepspeed\"",
+        "rich; extra == \"rich\"",
+        "torchao; extra == \"test-fp8\"",
+        "wandb; extra == \"test-trackers\"",
+        "tensorboard; extra == \"test-trackers\"",
+        "dvclive; extra == \"test-trackers\"",
+        "matplotlib; extra == \"test-trackers\"",
+        "swanlab[dashboard]; extra == \"test-trackers\"",
+        "trackio; extra == \"test-trackers\"",
+        "ruff==0.13.1; extra == \"dev\"",
+        "pytest>=7.2.0; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-subtests; extra == \"dev\"",
+        "parameterized; extra == \"dev\"",
+        "pytest-order; extra == \"dev\"",
+        "datasets; extra == \"dev\"",
+        "diffusers; extra == \"dev\"",
+        "evaluate; extra == \"dev\"",
+        "torchdata>=0.8.0; extra == \"dev\"",
+        "torchpippy>=0.2.0; extra == \"dev\"",
+        "transformers; extra == \"dev\"",
+        "scipy; extra == \"dev\"",
+        "scikit-learn; extra == \"dev\"",
+        "tqdm; extra == \"dev\"",
+        "bitsandbytes; extra == \"dev\"",
+        "timm; extra == \"dev\"",
+        "rich; extra == \"dev\"",
+        "sagemaker; extra == \"sagemaker\""
+      ],
+      "installer": "pip"
+    },
+    "tokenizers": {
+      "version": "0.21.4",
+      "requires_dist": [
+        "huggingface-hub>=0.16.4,<1.0",
+        "pytest ; extra == 'testing'",
+        "requests ; extra == 'testing'",
+        "numpy ; extra == 'testing'",
+        "datasets ; extra == 'testing'",
+        "black==22.3 ; extra == 'testing'",
+        "ruff ; extra == 'testing'",
+        "sphinx ; extra == 'docs'",
+        "sphinx-rtd-theme ; extra == 'docs'",
+        "setuptools-rust ; extra == 'docs'",
+        "tokenizers[testing] ; extra == 'dev'"
+      ],
+      "installer": "pip"
+    },
+    "opencv-python": {
+      "version": "4.12.0.88",
+      "requires_dist": [
+        "numpy<2.0; python_version < \"3.9\"",
+        "numpy<2.3.0,>=2; python_version >= \"3.9\""
+      ],
+      "installer": "pip"
+    },
+    "torchvision": {
+      "version": "0.24.1+cu130",
+      "requires_dist": [
+        "numpy",
+        "torch (==2.9.1)",
+        "pillow (!=8.3.*,>=5.3.0)",
+        "gdown (>=4.7.3) ; extra == 'gdown'",
+        "scipy ; extra == 'scipy'"
+      ],
+      "installer": "pip"
+    },
+    "diffusers": {
+      "version": "0.34.0",
+      "requires_dist": [
+        "importlib_metadata",
+        "filelock",
+        "huggingface-hub>=0.27.0",
+        "numpy",
+        "regex!=2019.12.17",
+        "requests",
+        "safetensors>=0.3.1",
+        "Pillow",
+        "urllib3<=2.0.0; extra == \"quality\"",
+        "isort>=5.5.4; extra == \"quality\"",
+        "ruff==0.9.10; extra == \"quality\"",
+        "hf-doc-builder>=0.3.0; extra == \"quality\"",
+        "hf-doc-builder>=0.3.0; extra == \"docs\"",
+        "accelerate>=0.31.0; extra == \"training\"",
+        "datasets; extra == \"training\"",
+        "protobuf<4,>=3.20.3; extra == \"training\"",
+        "tensorboard; extra == \"training\"",
+        "Jinja2; extra == \"training\"",
+        "peft>=0.15.0; extra == \"training\"",
+        "compel==0.1.8; extra == \"test\"",
+        "GitPython<3.1.19; extra == \"test\"",
+        "datasets; extra == \"test\"",
+        "Jinja2; extra == \"test\"",
+        "invisible-watermark>=0.2.0; extra == \"test\"",
+        "k-diffusion>=0.0.12; extra == \"test\"",
+        "librosa; extra == \"test\"",
+        "parameterized; extra == \"test\"",
+        "pytest; extra == \"test\"",
+        "pytest-timeout; extra == \"test\"",
+        "pytest-xdist; extra == \"test\"",
+        "requests-mock==1.10.0; extra == \"test\"",
+        "safetensors>=0.3.1; extra == \"test\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"test\"",
+        "scipy; extra == \"test\"",
+        "tiktoken>=0.7.0; extra == \"test\"",
+        "torchvision; extra == \"test\"",
+        "transformers>=4.41.2; extra == \"test\"",
+        "phonemizer; extra == \"test\"",
+        "torch>=1.4; extra == \"torch\"",
+        "accelerate>=0.31.0; extra == \"torch\"",
+        "bitsandbytes>=0.43.3; extra == \"bitsandbytes\"",
+        "accelerate>=0.31.0; extra == \"bitsandbytes\"",
+        "gguf>=0.10.0; extra == \"gguf\"",
+        "accelerate>=0.31.0; extra == \"gguf\"",
+        "optimum_quanto>=0.2.6; extra == \"optimum-quanto\"",
+        "accelerate>=0.31.0; extra == \"optimum-quanto\"",
+        "torchao>=0.7.0; extra == \"torchao\"",
+        "accelerate>=0.31.0; extra == \"torchao\"",
+        "jax>=0.4.1; extra == \"flax\"",
+        "jaxlib>=0.4.1; extra == \"flax\"",
+        "flax>=0.4.1; extra == \"flax\"",
+        "urllib3<=2.0.0; extra == \"dev\"",
+        "isort>=5.5.4; extra == \"dev\"",
+        "ruff==0.9.10; extra == \"dev\"",
+        "hf-doc-builder>=0.3.0; extra == \"dev\"",
+        "compel==0.1.8; extra == \"dev\"",
+        "GitPython<3.1.19; extra == \"dev\"",
+        "datasets; extra == \"dev\"",
+        "Jinja2; extra == \"dev\"",
+        "invisible-watermark>=0.2.0; extra == \"dev\"",
+        "k-diffusion>=0.0.12; extra == \"dev\"",
+        "librosa; extra == \"dev\"",
+        "parameterized; extra == \"dev\"",
+        "pytest; extra == \"dev\"",
+        "pytest-timeout; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "requests-mock==1.10.0; extra == \"dev\"",
+        "safetensors>=0.3.1; extra == \"dev\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev\"",
+        "scipy; extra == \"dev\"",
+        "tiktoken>=0.7.0; extra == \"dev\"",
+        "torchvision; extra == \"dev\"",
+        "transformers>=4.41.2; extra == \"dev\"",
+        "phonemizer; extra == \"dev\"",
+        "accelerate>=0.31.0; extra == \"dev\"",
+        "datasets; extra == \"dev\"",
+        "protobuf<4,>=3.20.3; extra == \"dev\"",
+        "tensorboard; extra == \"dev\"",
+        "Jinja2; extra == \"dev\"",
+        "peft>=0.15.0; extra == \"dev\"",
+        "hf-doc-builder>=0.3.0; extra == \"dev\"",
+        "torch>=1.4; extra == \"dev\"",
+        "accelerate>=0.31.0; extra == \"dev\"",
+        "jax>=0.4.1; extra == \"dev\"",
+        "jaxlib>=0.4.1; extra == \"dev\"",
+        "flax>=0.4.1; extra == \"dev\""
+      ],
+      "installer": "pip"
+    },
+    "numpy": {
+      "version": "2.2.6",
+      "requires_dist": [],
+      "installer": "pip"
+    },
+    "torchaudio": {
+      "version": "2.9.1+cu130",
+      "requires_dist": [
+        "torch==2.9.1"
+      ],
+      "installer": "pip"
+    },
+    "torch": {
+      "version": "2.9.1+cu130",
+      "requires_dist": [
+        "filelock",
+        "typing-extensions>=4.10.0",
+        "setuptools; python_version >= \"3.12\"",
+        "sympy>=1.13.3",
+        "networkx>=2.5.1",
+        "jinja2",
+        "fsspec>=0.8.5",
+        "nvidia-cuda-nvrtc==13.0.48; platform_system == \"Linux\"",
+        "nvidia-cuda-runtime==13.0.48; platform_system == \"Linux\"",
+        "nvidia-cuda-cupti==13.0.48; platform_system == \"Linux\"",
+        "nvidia-cudnn-cu13==9.13.0.50; platform_system == \"Linux\"",
+        "nvidia-cublas==13.0.0.19; platform_system == \"Linux\"",
+        "nvidia-cufft==12.0.0.15; platform_system == \"Linux\"",
+        "nvidia-curand==10.4.0.35; platform_system == \"Linux\"",
+        "nvidia-cusolver==12.0.3.29; platform_system == \"Linux\"",
+        "nvidia-cusparse==12.6.2.49; platform_system == \"Linux\"",
+        "nvidia-cusparselt-cu13==0.8.0; platform_system == \"Linux\"",
+        "nvidia-nccl-cu13==2.27.7; platform_system == \"Linux\"",
+        "nvidia-nvshmem-cu13==3.3.24; platform_system == \"Linux\"",
+        "nvidia-nvtx==13.0.39; platform_system == \"Linux\"",
+        "nvidia-nvjitlink==13.0.39; platform_system == \"Linux\"",
+        "nvidia-cufile==1.15.0.42; platform_system == \"Linux\"",
+        "triton==3.5.1; platform_system == \"Linux\"",
+        "optree>=0.13.0; extra == \"optree\"",
+        "opt-einsum>=3.3; extra == \"opt-einsum\"",
+        "pyyaml; extra == \"pyyaml\""
+      ],
+      "installer": "pip"
+    },
+    "transformers": {
+      "version": "4.54.0",
+      "requires_dist": [
+        "filelock",
+        "huggingface-hub<1.0,>=0.34.0",
+        "numpy>=1.17",
+        "packaging>=20.0",
+        "pyyaml>=5.1",
+        "regex!=2019.12.17",
+        "requests",
+        "tokenizers<0.22,>=0.21",
+        "safetensors>=0.4.3",
+        "tqdm>=4.27",
+        "fugashi>=1.0; extra == \"ja\"",
+        "ipadic<2.0,>=1.0.0; extra == \"ja\"",
+        "unidic_lite>=1.0.7; extra == \"ja\"",
+        "unidic>=1.0.2; extra == \"ja\"",
+        "sudachipy>=0.6.6; extra == \"ja\"",
+        "sudachidict_core>=20220729; extra == \"ja\"",
+        "rhoknp<1.3.1,>=1.1.0; extra == \"ja\"",
+        "scikit-learn; extra == \"sklearn\"",
+        "tensorflow<2.16,>2.9; extra == \"tf\"",
+        "onnxconverter-common; extra == \"tf\"",
+        "tf2onnx; extra == \"tf\"",
+        "tensorflow-text<2.16; extra == \"tf\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"tf\"",
+        "keras<2.16,>2.9; extra == \"tf-cpu\"",
+        "tensorflow-cpu<2.16,>2.9; extra == \"tf-cpu\"",
+        "onnxconverter-common; extra == \"tf-cpu\"",
+        "tf2onnx; extra == \"tf-cpu\"",
+        "tensorflow-text<2.16; extra == \"tf-cpu\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"tf-cpu\"",
+        "tensorflow-probability<0.24; extra == \"tf-cpu\"",
+        "torch>=2.1; extra == \"torch\"",
+        "accelerate>=0.26.0; extra == \"torch\"",
+        "accelerate>=0.26.0; extra == \"accelerate\"",
+        "hf_xet; extra == \"hf-xet\"",
+        "faiss-cpu; extra == \"retrieval\"",
+        "datasets>=2.15.0; extra == \"retrieval\"",
+        "jax<=0.4.13,>=0.4.1; extra == \"flax\"",
+        "jaxlib<=0.4.13,>=0.4.1; extra == \"flax\"",
+        "flax<=0.7.0,>=0.4.1; extra == \"flax\"",
+        "optax<=0.1.4,>=0.0.8; extra == \"flax\"",
+        "scipy<1.13.0; extra == \"flax\"",
+        "tokenizers<0.22,>=0.21; extra == \"tokenizers\"",
+        "ftfy; extra == \"ftfy\"",
+        "onnxruntime>=1.4.0; extra == \"onnxruntime\"",
+        "onnxruntime-tools>=1.4.2; extra == \"onnxruntime\"",
+        "onnxconverter-common; extra == \"onnx\"",
+        "tf2onnx; extra == \"onnx\"",
+        "onnxruntime>=1.4.0; extra == \"onnx\"",
+        "onnxruntime-tools>=1.4.2; extra == \"onnx\"",
+        "cookiecutter==1.7.3; extra == \"modelcreation\"",
+        "sagemaker>=2.31.0; extra == \"sagemaker\"",
+        "deepspeed>=0.9.3; extra == \"deepspeed\"",
+        "accelerate>=0.26.0; extra == \"deepspeed\"",
+        "optuna; extra == \"optuna\"",
+        "ray[tune]>=2.7.0; extra == \"ray\"",
+        "sigopt; extra == \"sigopt\"",
+        "kernels<0.7,>=0.6.1; extra == \"hub-kernels\"",
+        "kernels<0.7,>=0.6.1; extra == \"integrations\"",
+        "optuna; extra == \"integrations\"",
+        "ray[tune]>=2.7.0; extra == \"integrations\"",
+        "sigopt; extra == \"integrations\"",
+        "openai; extra == \"serving\"",
+        "pydantic>=2; extra == \"serving\"",
+        "uvicorn; extra == \"serving\"",
+        "fastapi; extra == \"serving\"",
+        "starlette; extra == \"serving\"",
+        "torch>=2.1; extra == \"serving\"",
+        "accelerate>=0.26.0; extra == \"serving\"",
+        "librosa; extra == \"audio\"",
+        "pyctcdecode>=0.4.0; extra == \"audio\"",
+        "phonemizer; extra == \"audio\"",
+        "kenlm; extra == \"audio\"",
+        "torchaudio; extra == \"speech\"",
+        "librosa; extra == \"speech\"",
+        "pyctcdecode>=0.4.0; extra == \"speech\"",
+        "phonemizer; extra == \"speech\"",
+        "kenlm; extra == \"speech\"",
+        "torchaudio; extra == \"torch-speech\"",
+        "librosa; extra == \"torch-speech\"",
+        "pyctcdecode>=0.4.0; extra == \"torch-speech\"",
+        "phonemizer; extra == \"torch-speech\"",
+        "kenlm; extra == \"torch-speech\"",
+        "librosa; extra == \"tf-speech\"",
+        "pyctcdecode>=0.4.0; extra == \"tf-speech\"",
+        "phonemizer; extra == \"tf-speech\"",
+        "kenlm; extra == \"tf-speech\"",
+        "librosa; extra == \"flax-speech\"",
+        "pyctcdecode>=0.4.0; extra == \"flax-speech\"",
+        "phonemizer; extra == \"flax-speech\"",
+        "kenlm; extra == \"flax-speech\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"vision\"",
+        "timm!=1.0.18,<=1.0.19; extra == \"timm\"",
+        "torchvision; extra == \"torch-vision\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"torch-vision\"",
+        "natten<0.15.0,>=0.14.6; extra == \"natten\"",
+        "codecarbon>=2.8.1; extra == \"codecarbon\"",
+        "av; extra == \"video\"",
+        "num2words; extra == \"num2words\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"sentencepiece\"",
+        "protobuf; extra == \"sentencepiece\"",
+        "tiktoken; extra == \"tiktoken\"",
+        "blobfile; extra == \"tiktoken\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"mistral-common\"",
+        "pytest>=7.2.0; extra == \"testing\"",
+        "pytest-asyncio; extra == \"testing\"",
+        "pytest-rich; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-order; extra == \"testing\"",
+        "pytest-rerunfailures; extra == \"testing\"",
+        "timeout-decorator; extra == \"testing\"",
+        "parameterized; extra == \"testing\"",
+        "psutil; extra == \"testing\"",
+        "datasets>=2.15.0; extra == \"testing\"",
+        "dill<0.3.5; extra == \"testing\"",
+        "evaluate>=0.2.0; extra == \"testing\"",
+        "pytest-timeout; extra == \"testing\"",
+        "ruff==0.11.2; extra == \"testing\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"testing\"",
+        "nltk<=3.8.1; extra == \"testing\"",
+        "GitPython<3.1.19; extra == \"testing\"",
+        "sacremoses; extra == \"testing\"",
+        "rjieba; extra == \"testing\"",
+        "beautifulsoup4; extra == \"testing\"",
+        "tensorboard; extra == \"testing\"",
+        "pydantic>=2; extra == \"testing\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"testing\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"testing\"",
+        "faiss-cpu; extra == \"testing\"",
+        "datasets>=2.15.0; extra == \"testing\"",
+        "cookiecutter==1.7.3; extra == \"testing\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"testing\"",
+        "deepspeed>=0.9.3; extra == \"deepspeed-testing\"",
+        "accelerate>=0.26.0; extra == \"deepspeed-testing\"",
+        "pytest>=7.2.0; extra == \"deepspeed-testing\"",
+        "pytest-asyncio; extra == \"deepspeed-testing\"",
+        "pytest-rich; extra == \"deepspeed-testing\"",
+        "pytest-xdist; extra == \"deepspeed-testing\"",
+        "pytest-order; extra == \"deepspeed-testing\"",
+        "pytest-rerunfailures; extra == \"deepspeed-testing\"",
+        "timeout-decorator; extra == \"deepspeed-testing\"",
+        "parameterized; extra == \"deepspeed-testing\"",
+        "psutil; extra == \"deepspeed-testing\"",
+        "datasets>=2.15.0; extra == \"deepspeed-testing\"",
+        "dill<0.3.5; extra == \"deepspeed-testing\"",
+        "evaluate>=0.2.0; extra == \"deepspeed-testing\"",
+        "pytest-timeout; extra == \"deepspeed-testing\"",
+        "ruff==0.11.2; extra == \"deepspeed-testing\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"deepspeed-testing\"",
+        "nltk<=3.8.1; extra == \"deepspeed-testing\"",
+        "GitPython<3.1.19; extra == \"deepspeed-testing\"",
+        "sacremoses; extra == \"deepspeed-testing\"",
+        "rjieba; extra == \"deepspeed-testing\"",
+        "beautifulsoup4; extra == \"deepspeed-testing\"",
+        "tensorboard; extra == \"deepspeed-testing\"",
+        "pydantic>=2; extra == \"deepspeed-testing\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"deepspeed-testing\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"deepspeed-testing\"",
+        "faiss-cpu; extra == \"deepspeed-testing\"",
+        "datasets>=2.15.0; extra == \"deepspeed-testing\"",
+        "cookiecutter==1.7.3; extra == \"deepspeed-testing\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"deepspeed-testing\"",
+        "optuna; extra == \"deepspeed-testing\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"deepspeed-testing\"",
+        "protobuf; extra == \"deepspeed-testing\"",
+        "ruff==0.11.2; extra == \"ruff\"",
+        "datasets>=2.15.0; extra == \"quality\"",
+        "ruff==0.11.2; extra == \"quality\"",
+        "GitPython<3.1.19; extra == \"quality\"",
+        "urllib3<2.0.0; extra == \"quality\"",
+        "libcst; extra == \"quality\"",
+        "rich; extra == \"quality\"",
+        "pandas<2.3.0; extra == \"quality\"",
+        "tensorflow<2.16,>2.9; extra == \"all\"",
+        "onnxconverter-common; extra == \"all\"",
+        "tf2onnx; extra == \"all\"",
+        "tensorflow-text<2.16; extra == \"all\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"all\"",
+        "torch>=2.1; extra == \"all\"",
+        "accelerate>=0.26.0; extra == \"all\"",
+        "jax<=0.4.13,>=0.4.1; extra == \"all\"",
+        "jaxlib<=0.4.13,>=0.4.1; extra == \"all\"",
+        "flax<=0.7.0,>=0.4.1; extra == \"all\"",
+        "optax<=0.1.4,>=0.0.8; extra == \"all\"",
+        "scipy<1.13.0; extra == \"all\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"all\"",
+        "protobuf; extra == \"all\"",
+        "tokenizers<0.22,>=0.21; extra == \"all\"",
+        "torchaudio; extra == \"all\"",
+        "librosa; extra == \"all\"",
+        "pyctcdecode>=0.4.0; extra == \"all\"",
+        "phonemizer; extra == \"all\"",
+        "kenlm; extra == \"all\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"all\"",
+        "kernels<0.7,>=0.6.1; extra == \"all\"",
+        "optuna; extra == \"all\"",
+        "ray[tune]>=2.7.0; extra == \"all\"",
+        "sigopt; extra == \"all\"",
+        "timm!=1.0.18,<=1.0.19; extra == \"all\"",
+        "torchvision; extra == \"all\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"all\"",
+        "codecarbon>=2.8.1; extra == \"all\"",
+        "accelerate>=0.26.0; extra == \"all\"",
+        "av; extra == \"all\"",
+        "num2words; extra == \"all\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"all\"",
+        "pytest>=7.2.0; extra == \"dev-torch\"",
+        "pytest-asyncio; extra == \"dev-torch\"",
+        "pytest-rich; extra == \"dev-torch\"",
+        "pytest-xdist; extra == \"dev-torch\"",
+        "pytest-order; extra == \"dev-torch\"",
+        "pytest-rerunfailures; extra == \"dev-torch\"",
+        "timeout-decorator; extra == \"dev-torch\"",
+        "parameterized; extra == \"dev-torch\"",
+        "psutil; extra == \"dev-torch\"",
+        "datasets>=2.15.0; extra == \"dev-torch\"",
+        "dill<0.3.5; extra == \"dev-torch\"",
+        "evaluate>=0.2.0; extra == \"dev-torch\"",
+        "pytest-timeout; extra == \"dev-torch\"",
+        "ruff==0.11.2; extra == \"dev-torch\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"dev-torch\"",
+        "nltk<=3.8.1; extra == \"dev-torch\"",
+        "GitPython<3.1.19; extra == \"dev-torch\"",
+        "sacremoses; extra == \"dev-torch\"",
+        "rjieba; extra == \"dev-torch\"",
+        "beautifulsoup4; extra == \"dev-torch\"",
+        "tensorboard; extra == \"dev-torch\"",
+        "pydantic>=2; extra == \"dev-torch\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev-torch\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"dev-torch\"",
+        "faiss-cpu; extra == \"dev-torch\"",
+        "datasets>=2.15.0; extra == \"dev-torch\"",
+        "cookiecutter==1.7.3; extra == \"dev-torch\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"dev-torch\"",
+        "torch>=2.1; extra == \"dev-torch\"",
+        "accelerate>=0.26.0; extra == \"dev-torch\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev-torch\"",
+        "protobuf; extra == \"dev-torch\"",
+        "tokenizers<0.22,>=0.21; extra == \"dev-torch\"",
+        "torchaudio; extra == \"dev-torch\"",
+        "librosa; extra == \"dev-torch\"",
+        "pyctcdecode>=0.4.0; extra == \"dev-torch\"",
+        "phonemizer; extra == \"dev-torch\"",
+        "kenlm; extra == \"dev-torch\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev-torch\"",
+        "kernels<0.7,>=0.6.1; extra == \"dev-torch\"",
+        "optuna; extra == \"dev-torch\"",
+        "ray[tune]>=2.7.0; extra == \"dev-torch\"",
+        "sigopt; extra == \"dev-torch\"",
+        "timm!=1.0.18,<=1.0.19; extra == \"dev-torch\"",
+        "torchvision; extra == \"dev-torch\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev-torch\"",
+        "codecarbon>=2.8.1; extra == \"dev-torch\"",
+        "datasets>=2.15.0; extra == \"dev-torch\"",
+        "ruff==0.11.2; extra == \"dev-torch\"",
+        "GitPython<3.1.19; extra == \"dev-torch\"",
+        "urllib3<2.0.0; extra == \"dev-torch\"",
+        "libcst; extra == \"dev-torch\"",
+        "rich; extra == \"dev-torch\"",
+        "pandas<2.3.0; extra == \"dev-torch\"",
+        "fugashi>=1.0; extra == \"dev-torch\"",
+        "ipadic<2.0,>=1.0.0; extra == \"dev-torch\"",
+        "unidic_lite>=1.0.7; extra == \"dev-torch\"",
+        "unidic>=1.0.2; extra == \"dev-torch\"",
+        "sudachipy>=0.6.6; extra == \"dev-torch\"",
+        "sudachidict_core>=20220729; extra == \"dev-torch\"",
+        "rhoknp<1.3.1,>=1.1.0; extra == \"dev-torch\"",
+        "scikit-learn; extra == \"dev-torch\"",
+        "cookiecutter==1.7.3; extra == \"dev-torch\"",
+        "onnxruntime>=1.4.0; extra == \"dev-torch\"",
+        "onnxruntime-tools>=1.4.2; extra == \"dev-torch\"",
+        "num2words; extra == \"dev-torch\"",
+        "pytest>=7.2.0; extra == \"dev-tensorflow\"",
+        "pytest-asyncio; extra == \"dev-tensorflow\"",
+        "pytest-rich; extra == \"dev-tensorflow\"",
+        "pytest-xdist; extra == \"dev-tensorflow\"",
+        "pytest-order; extra == \"dev-tensorflow\"",
+        "pytest-rerunfailures; extra == \"dev-tensorflow\"",
+        "timeout-decorator; extra == \"dev-tensorflow\"",
+        "parameterized; extra == \"dev-tensorflow\"",
+        "psutil; extra == \"dev-tensorflow\"",
+        "datasets>=2.15.0; extra == \"dev-tensorflow\"",
+        "dill<0.3.5; extra == \"dev-tensorflow\"",
+        "evaluate>=0.2.0; extra == \"dev-tensorflow\"",
+        "pytest-timeout; extra == \"dev-tensorflow\"",
+        "ruff==0.11.2; extra == \"dev-tensorflow\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"dev-tensorflow\"",
+        "nltk<=3.8.1; extra == \"dev-tensorflow\"",
+        "GitPython<3.1.19; extra == \"dev-tensorflow\"",
+        "sacremoses; extra == \"dev-tensorflow\"",
+        "rjieba; extra == \"dev-tensorflow\"",
+        "beautifulsoup4; extra == \"dev-tensorflow\"",
+        "tensorboard; extra == \"dev-tensorflow\"",
+        "pydantic>=2; extra == \"dev-tensorflow\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev-tensorflow\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"dev-tensorflow\"",
+        "faiss-cpu; extra == \"dev-tensorflow\"",
+        "datasets>=2.15.0; extra == \"dev-tensorflow\"",
+        "cookiecutter==1.7.3; extra == \"dev-tensorflow\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"dev-tensorflow\"",
+        "tensorflow<2.16,>2.9; extra == \"dev-tensorflow\"",
+        "onnxconverter-common; extra == \"dev-tensorflow\"",
+        "tf2onnx; extra == \"dev-tensorflow\"",
+        "tensorflow-text<2.16; extra == \"dev-tensorflow\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"dev-tensorflow\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev-tensorflow\"",
+        "protobuf; extra == \"dev-tensorflow\"",
+        "tokenizers<0.22,>=0.21; extra == \"dev-tensorflow\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev-tensorflow\"",
+        "datasets>=2.15.0; extra == \"dev-tensorflow\"",
+        "ruff==0.11.2; extra == \"dev-tensorflow\"",
+        "GitPython<3.1.19; extra == \"dev-tensorflow\"",
+        "urllib3<2.0.0; extra == \"dev-tensorflow\"",
+        "libcst; extra == \"dev-tensorflow\"",
+        "rich; extra == \"dev-tensorflow\"",
+        "pandas<2.3.0; extra == \"dev-tensorflow\"",
+        "scikit-learn; extra == \"dev-tensorflow\"",
+        "cookiecutter==1.7.3; extra == \"dev-tensorflow\"",
+        "onnxconverter-common; extra == \"dev-tensorflow\"",
+        "tf2onnx; extra == \"dev-tensorflow\"",
+        "onnxruntime>=1.4.0; extra == \"dev-tensorflow\"",
+        "onnxruntime-tools>=1.4.2; extra == \"dev-tensorflow\"",
+        "librosa; extra == \"dev-tensorflow\"",
+        "pyctcdecode>=0.4.0; extra == \"dev-tensorflow\"",
+        "phonemizer; extra == \"dev-tensorflow\"",
+        "kenlm; extra == \"dev-tensorflow\"",
+        "tensorflow<2.16,>2.9; extra == \"dev\"",
+        "onnxconverter-common; extra == \"dev\"",
+        "tf2onnx; extra == \"dev\"",
+        "tensorflow-text<2.16; extra == \"dev\"",
+        "keras-nlp<0.14.0,>=0.3.1; extra == \"dev\"",
+        "torch>=2.1; extra == \"dev\"",
+        "accelerate>=0.26.0; extra == \"dev\"",
+        "jax<=0.4.13,>=0.4.1; extra == \"dev\"",
+        "jaxlib<=0.4.13,>=0.4.1; extra == \"dev\"",
+        "flax<=0.7.0,>=0.4.1; extra == \"dev\"",
+        "optax<=0.1.4,>=0.0.8; extra == \"dev\"",
+        "scipy<1.13.0; extra == \"dev\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev\"",
+        "protobuf; extra == \"dev\"",
+        "tokenizers<0.22,>=0.21; extra == \"dev\"",
+        "torchaudio; extra == \"dev\"",
+        "librosa; extra == \"dev\"",
+        "pyctcdecode>=0.4.0; extra == \"dev\"",
+        "phonemizer; extra == \"dev\"",
+        "kenlm; extra == \"dev\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev\"",
+        "kernels<0.7,>=0.6.1; extra == \"dev\"",
+        "optuna; extra == \"dev\"",
+        "ray[tune]>=2.7.0; extra == \"dev\"",
+        "sigopt; extra == \"dev\"",
+        "timm!=1.0.18,<=1.0.19; extra == \"dev\"",
+        "torchvision; extra == \"dev\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev\"",
+        "codecarbon>=2.8.1; extra == \"dev\"",
+        "accelerate>=0.26.0; extra == \"dev\"",
+        "av; extra == \"dev\"",
+        "num2words; extra == \"dev\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"dev\"",
+        "pytest>=7.2.0; extra == \"dev\"",
+        "pytest-asyncio; extra == \"dev\"",
+        "pytest-rich; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-order; extra == \"dev\"",
+        "pytest-rerunfailures; extra == \"dev\"",
+        "timeout-decorator; extra == \"dev\"",
+        "parameterized; extra == \"dev\"",
+        "psutil; extra == \"dev\"",
+        "datasets>=2.15.0; extra == \"dev\"",
+        "dill<0.3.5; extra == \"dev\"",
+        "evaluate>=0.2.0; extra == \"dev\"",
+        "pytest-timeout; extra == \"dev\"",
+        "ruff==0.11.2; extra == \"dev\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"dev\"",
+        "nltk<=3.8.1; extra == \"dev\"",
+        "GitPython<3.1.19; extra == \"dev\"",
+        "sacremoses; extra == \"dev\"",
+        "rjieba; extra == \"dev\"",
+        "beautifulsoup4; extra == \"dev\"",
+        "tensorboard; extra == \"dev\"",
+        "pydantic>=2; extra == \"dev\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"dev\"",
+        "faiss-cpu; extra == \"dev\"",
+        "datasets>=2.15.0; extra == \"dev\"",
+        "cookiecutter==1.7.3; extra == \"dev\"",
+        "mistral-common[opencv]>=1.6.3; extra == \"dev\"",
+        "datasets>=2.15.0; extra == \"dev\"",
+        "ruff==0.11.2; extra == \"dev\"",
+        "GitPython<3.1.19; extra == \"dev\"",
+        "urllib3<2.0.0; extra == \"dev\"",
+        "libcst; extra == \"dev\"",
+        "rich; extra == \"dev\"",
+        "pandas<2.3.0; extra == \"dev\"",
+        "fugashi>=1.0; extra == \"dev\"",
+        "ipadic<2.0,>=1.0.0; extra == \"dev\"",
+        "unidic_lite>=1.0.7; extra == \"dev\"",
+        "unidic>=1.0.2; extra == \"dev\"",
+        "sudachipy>=0.6.6; extra == \"dev\"",
+        "sudachidict_core>=20220729; extra == \"dev\"",
+        "rhoknp<1.3.1,>=1.1.0; extra == \"dev\"",
+        "scikit-learn; extra == \"dev\"",
+        "cookiecutter==1.7.3; extra == \"dev\"",
+        "filelock; extra == \"torchhub\"",
+        "huggingface-hub<1.0,>=0.34.0; extra == \"torchhub\"",
+        "importlib_metadata; extra == \"torchhub\"",
+        "numpy>=1.17; extra == \"torchhub\"",
+        "packaging>=20.0; extra == \"torchhub\"",
+        "protobuf; extra == \"torchhub\"",
+        "regex!=2019.12.17; extra == \"torchhub\"",
+        "requests; extra == \"torchhub\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"torchhub\"",
+        "torch>=2.1; extra == \"torchhub\"",
+        "tokenizers<0.22,>=0.21; extra == \"torchhub\"",
+        "tqdm>=4.27; extra == \"torchhub\"",
+        "optimum-benchmark>=0.3.0; extra == \"benchmark\"",
+        "opentelemetry-api; extra == \"open-telemetry\"",
+        "opentelemetry-exporter-otlp; extra == \"open-telemetry\"",
+        "opentelemetry-sdk; extra == \"open-telemetry\""
+      ],
+      "installer": "pip"
+    },
+    "pandas": {
+      "version": "2.3.2",
+      "requires_dist": [
+        "numpy>=1.22.4; python_version < \"3.11\"",
+        "numpy>=1.23.2; python_version == \"3.11\"",
+        "numpy>=1.26.0; python_version >= \"3.12\"",
+        "python-dateutil>=2.8.2",
+        "pytz>=2020.1",
+        "tzdata>=2022.7",
+        "hypothesis>=6.46.1; extra == \"test\"",
+        "pytest>=7.3.2; extra == \"test\"",
+        "pytest-xdist>=2.2.0; extra == \"test\"",
+        "pyarrow>=10.0.1; extra == \"pyarrow\"",
+        "bottleneck>=1.3.6; extra == \"performance\"",
+        "numba>=0.56.4; extra == \"performance\"",
+        "numexpr>=2.8.4; extra == \"performance\"",
+        "scipy>=1.10.0; extra == \"computation\"",
+        "xarray>=2022.12.0; extra == \"computation\"",
+        "fsspec>=2022.11.0; extra == \"fss\"",
+        "s3fs>=2022.11.0; extra == \"aws\"",
+        "gcsfs>=2022.11.0; extra == \"gcp\"",
+        "pandas-gbq>=0.19.0; extra == \"gcp\"",
+        "odfpy>=1.4.1; extra == \"excel\"",
+        "openpyxl>=3.1.0; extra == \"excel\"",
+        "python-calamine>=0.1.7; extra == \"excel\"",
+        "pyxlsb>=1.0.10; extra == \"excel\"",
+        "xlrd>=2.0.1; extra == \"excel\"",
+        "xlsxwriter>=3.0.5; extra == \"excel\"",
+        "pyarrow>=10.0.1; extra == \"parquet\"",
+        "pyarrow>=10.0.1; extra == \"feather\"",
+        "tables>=3.8.0; extra == \"hdf5\"",
+        "pyreadstat>=1.2.0; extra == \"spss\"",
+        "SQLAlchemy>=2.0.0; extra == \"postgresql\"",
+        "psycopg2>=2.9.6; extra == \"postgresql\"",
+        "adbc-driver-postgresql>=0.8.0; extra == \"postgresql\"",
+        "SQLAlchemy>=2.0.0; extra == \"mysql\"",
+        "pymysql>=1.0.2; extra == \"mysql\"",
+        "SQLAlchemy>=2.0.0; extra == \"sql-other\"",
+        "adbc-driver-postgresql>=0.8.0; extra == \"sql-other\"",
+        "adbc-driver-sqlite>=0.8.0; extra == \"sql-other\"",
+        "beautifulsoup4>=4.11.2; extra == \"html\"",
+        "html5lib>=1.1; extra == \"html\"",
+        "lxml>=4.9.2; extra == \"html\"",
+        "lxml>=4.9.2; extra == \"xml\"",
+        "matplotlib>=3.6.3; extra == \"plot\"",
+        "jinja2>=3.1.2; extra == \"output-formatting\"",
+        "tabulate>=0.9.0; extra == \"output-formatting\"",
+        "PyQt5>=5.15.9; extra == \"clipboard\"",
+        "qtpy>=2.3.0; extra == \"clipboard\"",
+        "zstandard>=0.19.0; extra == \"compression\"",
+        "dataframe-api-compat>=0.1.7; extra == \"consortium-standard\"",
+        "adbc-driver-postgresql>=0.8.0; extra == \"all\"",
+        "adbc-driver-sqlite>=0.8.0; extra == \"all\"",
+        "beautifulsoup4>=4.11.2; extra == \"all\"",
+        "bottleneck>=1.3.6; extra == \"all\"",
+        "dataframe-api-compat>=0.1.7; extra == \"all\"",
+        "fastparquet>=2022.12.0; extra == \"all\"",
+        "fsspec>=2022.11.0; extra == \"all\"",
+        "gcsfs>=2022.11.0; extra == \"all\"",
+        "html5lib>=1.1; extra == \"all\"",
+        "hypothesis>=6.46.1; extra == \"all\"",
+        "jinja2>=3.1.2; extra == \"all\"",
+        "lxml>=4.9.2; extra == \"all\"",
+        "matplotlib>=3.6.3; extra == \"all\"",
+        "numba>=0.56.4; extra == \"all\"",
+        "numexpr>=2.8.4; extra == \"all\"",
+        "odfpy>=1.4.1; extra == \"all\"",
+        "openpyxl>=3.1.0; extra == \"all\"",
+        "pandas-gbq>=0.19.0; extra == \"all\"",
+        "psycopg2>=2.9.6; extra == \"all\"",
+        "pyarrow>=10.0.1; extra == \"all\"",
+        "pymysql>=1.0.2; extra == \"all\"",
+        "PyQt5>=5.15.9; extra == \"all\"",
+        "pyreadstat>=1.2.0; extra == \"all\"",
+        "pytest>=7.3.2; extra == \"all\"",
+        "pytest-xdist>=2.2.0; extra == \"all\"",
+        "python-calamine>=0.1.7; extra == \"all\"",
+        "pyxlsb>=1.0.10; extra == \"all\"",
+        "qtpy>=2.3.0; extra == \"all\"",
+        "scipy>=1.10.0; extra == \"all\"",
+        "s3fs>=2022.11.0; extra == \"all\"",
+        "SQLAlchemy>=2.0.0; extra == \"all\"",
+        "tables>=3.8.0; extra == \"all\"",
+        "tabulate>=0.9.0; extra == \"all\"",
+        "xarray>=2022.12.0; extra == \"all\"",
+        "xlrd>=2.0.1; extra == \"all\"",
+        "xlsxwriter>=3.0.5; extra == \"all\"",
+        "zstandard>=0.19.0; extra == \"all\""
+      ],
+      "installer": "pip"
+    }
+  },
+  "module_presence": {
+    "torch": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "3caf7f40140ede2465bde40b9003af10cbbc8f7bcf436fa1de026471daa1b288"
+    },
+    "torchvision": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "ee2c9f4110cf1203db48c42601607329ac1f19709fa91c152f8d95eb53437a73"
+    },
+    "torchaudio": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "f0e07710f18256217b2e0056cbf6d4c99505e872542296d323ca2e446c27ea55"
+    },
+    "transformers": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "45c88ca57ca9c5f65beaf91f5b0e6e47836d2f5bbbd340c35fb27b498ae57db7"
+    },
+    "tokenizers": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "644e596a052fa1b05272b1c141d1286e1c78c2a3346ecabd17d68b62404d8d84"
+    },
+    "huggingface_hub": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "e586b64612482128fd24119364bc4dc6dca938407134e454a5e876341b364e35"
+    },
+    "diffusers": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "a438cb7ca43fedd9829104f472b4af7b1eeb05f51176a8746678d95a22ce8411"
+    },
+    "diffsynth": {
+      "found": false
+    },
+    "accelerate": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "79eabeeaf005b8b0f98dc833610e8c073f33e9653f3630c4ed22c30808052d27"
+    },
+    "peft": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "69de31a0c28be2a6e22d48a0cff469c8e98a67eece5a0bfccc37e52cee004751"
+    },
+    "swift": {
+      "found": false
+    },
+    "verl": {
+      "found": false
+    },
+    "vllm": {
+      "found": false
+    },
+    "fla": {
+      "found": false
+    },
+    "tilelang": {
+      "found": false
+    },
+    "triton": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "4ec1df745c6bfd03257370a2aab1534e7a31eaa62110cad2ff6c3b56a67cfc50"
+    },
+    "cv2": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "936bd94c5a5debf0212fc751af79d3a163652f3e850259df2159db6aa3ed8ad8"
+    },
+    "numpy": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "6ae17b070c0f70a8e3cad89a510a256942e5a1f37ea5feb120cec167ed2a6236"
+    }
+  },
+  "pip_check": {
+    "exit_code": 1,
+    "output": "decord 0.6.0 is not supported on this platform\n"
+  },
+  "GPU": {
+    "checked": true,
+    "basic_tensor_check": {
+      "exit_code": 0,
+      "output": "{\"torch\": \"2.9.1+cu130\", \"cuda_runtime\": \"13.0\", \"cuda_available\": true, \"arch_list\": [\"sm_75\", \"sm_80\", \"sm_86\", \"sm_90\", \"sm_100\", \"sm_120\", \"compute_120\"], \"gpu\": \"NVIDIA RPBZZZ6\", \"capability\": [12, 0], \"tiny_bfloat16_result\": 4.0}\n"
+    },
+    "training_or_model_inference_run": false
+  },
+  "original_environment_modified": false,
+  "credentials_included": false
+}

egoexo/generator/import_validation.json

@@ -0,0 +1,54 @@
+{
+  "torch": {
+    "ok": true,
+    "version": "2.9.1+cu130"
+  },
+  "torchvision": {
+    "ok": true,
+    "version": "0.24.1+cu130"
+  },
+  "torchaudio": {
+    "ok": true,
+    "version": "2.9.1+cu130"
+  },
+  "transformers": {
+    "ok": true,
+    "version": "4.54.0"
+  },
+  "peft": {
+    "ok": true,
+    "version": "0.19.1"
+  },
+  "accelerate": {
+    "ok": true,
+    "version": "1.14.0"
+  },
+  "safetensors": {
+    "ok": true,
+    "version": "0.6.2"
+  },
+  "decord": {
+    "ok": true,
+    "version": "0.6.0"
+  },
+  "vllm": {
+    "ok": false,
+    "error_type": "ModuleNotFoundError",
+    "error": "No module named 'vllm'"
+  },
+  "verl": {
+    "ok": false,
+    "error_type": "ModuleNotFoundError",
+    "error": "No module named 'verl'"
+  },
+  "fla": {
+    "ok": false,
+    "error_type": "ModuleNotFoundError",
+    "error": "No module named 'fla'"
+  },
+  "tilelang": {
+    "ok": false,
+    "error_type": "ModuleNotFoundError",
+    "error": "No module named 'tilelang'"
+  }
+}

egoexo/generator/requirements.core.in

@@ -0,0 +1,27 @@
+# Observed role-specific root requirements; transitive dependencies resolved/locked separately.
+torch==2.9.1+cu130
+torchvision==0.24.1+cu130
+torchaudio==2.9.1+cu130
+triton==3.5.1
+transformers==4.54.0
+tokenizers==0.21.4
+huggingface-hub==0.36.2
+accelerate==1.14.0
+diffusers==0.34.0
+peft==0.19.1
+safetensors==0.6.2
+einops==0.8.1
+numpy==2.2.6
+pillow==11.0.0
+opencv-python-headless==4.12.0.88
+pandas==2.3.2
+scipy==1.17.1
+sentencepiece==0.2.0
+ftfy==6.3.1
+omegaconf==2.3.1
+hydra-core==1.3.4
+imageio==2.37.0
+imageio-ffmpeg==0.6.0
+lpips==0.1.4
+torchmetrics==1.9.0
+tqdm==4.68.3

egoexo/generator/requirements.observed-full.lock.txt

@@ -0,0 +1,119 @@
+absl-py==2.5.0
+accelerate==1.14.0
+aiohappyeyeballs==2.7.1
+aiohttp==3.14.1
+aiosignal==1.4.0
+antlr4-python3-runtime==4.9.3
+attrs==26.1.0
+cd-fvd==0.1.1
+certifi==2026.6.17
+charset-normalizer==3.4.8
+datasets==4.0.0
+decord==0.6.0
+diffusers==0.34.0
+dill==0.3.8
+einops==0.8.1
+filelock==3.29.5
+frozenlist==1.8.0
+fsspec==2025.3.0
+ftfy==6.3.1
+grpcio==1.83.0
+hf-xet==1.5.1
+huggingface-hub==0.36.2
+hydra-core==1.3.4
+idna==3.18
+imageio==2.37.0
+imageio-ffmpeg==0.6.0
+importlib-metadata==9.0.0
+iopath==0.1.10
+jinja2==3.1.6
+lazy-loader==0.5
+lightning-utilities==0.15.3
+loguru==0.7.3
+lpips==0.1.4
+markdown==3.10.3
+markupsafe==3.0.3
+modelscope==1.36.3
+mpmath==1.3.0
+multidict==6.7.1
+multiprocess==0.70.16
+networkx==3.6.1
+numpy==2.2.6
+nvidia-cublas==13.0.0.19
+nvidia-cublas-cu12==12.6.4.1
+nvidia-cuda-cupti==13.0.48
+nvidia-cuda-cupti-cu12==12.6.80
+nvidia-cuda-nvrtc==13.0.48
+nvidia-cuda-nvrtc-cu12==12.6.77
+nvidia-cuda-runtime==13.0.48
+nvidia-cuda-runtime-cu12==12.6.77
+nvidia-cudnn-cu12==9.5.1.17
+nvidia-cudnn-cu13==9.13.0.50
+nvidia-cufft==12.0.0.15
+nvidia-cufft-cu12==11.3.0.4
+nvidia-cufile==1.15.0.42
+nvidia-cufile-cu12==1.11.1.6
+nvidia-curand==10.4.0.35
+nvidia-curand-cu12==10.3.7.77
+nvidia-cusolver==12.0.3.29
+nvidia-cusolver-cu12==11.7.1.2
+nvidia-cusparse==12.6.2.49
+nvidia-cusparse-cu12==12.5.4.2
+nvidia-cusparselt-cu12==0.6.3
+nvidia-cusparselt-cu13==0.8.0
+nvidia-nccl-cu12==2.26.2
+nvidia-nccl-cu13==2.27.7
+nvidia-nvjitlink==13.0.39
+nvidia-nvjitlink-cu12==12.6.85
+nvidia-nvshmem-cu13==3.3.24
+nvidia-nvtx==13.0.39
+nvidia-nvtx-cu12==12.6.77
+omegaconf==2.3.1
+opencv-python==4.12.0.88
+opencv-python-headless==4.12.0.88
+packaging==26.0
+pandas==2.3.2
+peft==0.19.1
+pillow==11.0.0
+pip==26.1.2
+portalocker==3.2.0
+propcache==0.5.2
+protobuf==6.32.0
+psutil==7.2.2
+pyarrow==24.0.0
+pymupdf==1.28.2
+pypdf==6.14.2
+python-dateutil==2.9.0.post0
+pytz==2026.2
+pyyaml==6.0.3
+regex==2026.6.28
+reportlab==5.0.1
+requests==2.34.2
+safetensors==0.6.2
+sam-2==1.0
+scikit-image==0.26.0
+scipy==1.17.1
+sentencepiece==0.2.0
+setuptools==82.0.1
+six==1.17.0
+sympy==1.14.0
+tensorboard==2.21.0
+tensorboard-data-server==0.7.2
+tifffile==2026.3.3
+tokenizers==0.21.4
+torch==2.9.1+cu130
+torchaudio==2.9.1+cu130
+torchmetrics==1.9.0
+torchvision==0.24.1+cu130
+tqdm==4.68.3
+transformers==4.54.0
+triton==3.5.1
+typing-extensions==4.16.0
+tzdata==2026.2
+urllib3==2.7.0
+wcwidth==0.8.2
+werkzeug==3.1.8
+wheel==0.47.0
+xxhash==3.8.1
+yarl==1.24.2
+zipp==4.1.0

robot/data_cpu/environment.snapshot.json

@@ -0,0 +1,399 @@
+{
+  "profile": "robot_data_cpu",
+  "capture_scope": "observed current retained runtime; not proof of original launch-time environment",
+  "python": "3.12.13",
+  "implementation": "CPython",
+  "SOABI": "cpython-312-x86_64-linux-gnu",
+  "machine": "x86_64",
+  "platform": "Linux",
+  "libc": [
+    "glibc",
+    "2.35"
+  ],
+  "packages": {
+    "anyio": "4.15.1",
+    "certifi": "2026.7.22",
+    "charset-normalizer": "3.5.1",
+    "click": "8.5.0",
+    "filelock": "4.0.5",
+    "fsspec": "2026.6.0",
+    "h11": "0.16.0",
+    "h5py": "3.16.0",
+    "hf-xet": "1.6.0",
+    "httpcore": "1.0.9",
+    "httpx": "0.28.1",
+    "huggingface-hub": "1.33.0",
+    "idna": "3.20",
+    "iniconfig": "2.3.0",
+    "numpy": "2.3.5",
+    "opencv-python-headless": "5.0.0.93",
+    "packaging": "25.0",
+    "pandas": "3.0.6",
+    "pillow": "12.3.0",
+    "pip": "25.0.1",
+    "pluggy": "1.6.0",
+    "pyarrow": "25.0.1",
+    "pygments": "2.21.0",
+    "pytest": "9.1.1",
+    "python-dateutil": "2.9.0.post0",
+    "pyyaml": "6.0.3",
+    "requests": "2.34.2",
+    "six": "1.17.0",
+    "tokenizers": "0.22.2",
+    "tqdm": "4.70.1",
+    "typing-extensions": "4.16.0",
+    "urllib3": "2.8.0"
+  },
+  "core_package_provenance": {
+    "pillow": {
+      "version": "12.3.0",
+      "requires_dist": [
+        "furo; extra == \"docs\"",
+        "olefile; extra == \"docs\"",
+        "sphinx>=8.2; extra == \"docs\"",
+        "sphinx-autobuild; extra == \"docs\"",
+        "sphinx-copybutton; extra == \"docs\"",
+        "sphinx-inline-tabs; extra == \"docs\"",
+        "sphinxext-opengraph; extra == \"docs\"",
+        "olefile; extra == \"fpx\"",
+        "olefile; extra == \"mic\"",
+        "arro3-compute; extra == \"test-arrow\"",
+        "arro3-core; extra == \"test-arrow\"",
+        "nanoarrow; extra == \"test-arrow\"",
+        "pyarrow; extra == \"test-arrow\"",
+        "coverage>=7.4.2; extra == \"tests\"",
+        "defusedxml; extra == \"tests\"",
+        "markdown2; extra == \"tests\"",
+        "olefile; extra == \"tests\"",
+        "packaging; extra == \"tests\"",
+        "pytest; extra == \"tests\"",
+        "pytest-cov; extra == \"tests\"",
+        "pytest-timeout; extra == \"tests\"",
+        "pytest-xdist; extra == \"tests\"",
+        "setuptools; extra == \"tests\"",
+        "trove-classifiers>=2024.10.12; extra == \"tests\"",
+        "defusedxml; extra == \"xmp\""
+      ],
+      "installer": "pip"
+    },
+    "pyarrow": {
+      "version": "25.0.1",
+      "requires_dist": [],
+      "installer": "pip"
+    },
+    "pandas": {
+      "version": "3.0.6",
+      "requires_dist": [
+        "numpy>=1.26.0; python_version < \"3.14\"",
+        "numpy>=2.3.3; python_version >= \"3.14\"",
+        "python-dateutil>=2.8.2",
+        "tzdata; sys_platform == \"win32\"",
+        "tzdata; sys_platform == \"emscripten\"",
+        "hypothesis>=6.116.0; extra == \"test\"",
+        "pytest<9.1,>=8.3.4; extra == \"test\"",
+        "pytest-xdist>=3.6.1; extra == \"test\"",
+        "pyarrow>=13.0.0; extra == \"pyarrow\"",
+        "bottleneck>=1.4.2; extra == \"performance\"",
+        "numba>=0.60.0; extra == \"performance\"",
+        "numexpr>=2.10.2; extra == \"performance\"",
+        "scipy>=1.14.1; extra == \"computation\"",
+        "xarray>=2024.10.0; extra == \"computation\"",
+        "fsspec>=2024.10.0; extra == \"fss\"",
+        "s3fs>=2024.10.0; extra == \"aws\"",
+        "gcsfs>=2024.10.0; extra == \"gcp\"",
+        "odfpy>=1.4.1; extra == \"excel\"",
+        "openpyxl>=3.1.5; extra == \"excel\"",
+        "python-calamine>=0.3.0; extra == \"excel\"",
+        "pyxlsb>=1.0.10; extra == \"excel\"",
+        "xlrd>=2.0.1; extra == \"excel\"",
+        "xlsxwriter>=3.2.0; extra == \"excel\"",
+        "pyarrow>=13.0.0; extra == \"parquet\"",
+        "pyarrow>=13.0.0; extra == \"feather\"",
+        "pyiceberg>=0.8.1; extra == \"iceberg\"",
+        "tables>=3.10.1; extra == \"hdf5\"",
+        "pyreadstat>=1.2.8; extra == \"spss\"",
+        "SQLAlchemy>=2.0.36; extra == \"postgresql\"",
+        "psycopg2>=2.9.10; extra == \"postgresql\"",
+        "adbc-driver-postgresql>=1.2.0; extra == \"postgresql\"",
+        "SQLAlchemy>=2.0.36; extra == \"mysql\"",
+        "pymysql>=1.1.1; extra == \"mysql\"",
+        "SQLAlchemy>=2.0.36; extra == \"sql-other\"",
+        "adbc-driver-postgresql>=1.2.0; extra == \"sql-other\"",
+        "adbc-driver-sqlite>=1.2.0; extra == \"sql-other\"",
+        "beautifulsoup4>=4.12.3; extra == \"html\"",
+        "html5lib>=1.1; extra == \"html\"",
+        "lxml>=5.3.0; extra == \"html\"",
+        "lxml>=5.3.0; extra == \"xml\"",
+        "matplotlib>=3.9.3; extra == \"plot\"",
+        "jinja2>=3.1.5; extra == \"output-formatting\"",
+        "tabulate>=0.9.0; extra == \"output-formatting\"",
+        "PyQt5>=5.15.9; extra == \"clipboard\"",
+        "qtpy>=2.4.2; extra == \"clipboard\"",
+        "zstandard>=0.23.0; extra == \"compression\"",
+        "pytz>=2020.1; extra == \"timezone\"",
+        "adbc-driver-postgresql>=1.2.0; extra == \"all\"",
+        "adbc-driver-sqlite>=1.2.0; extra == \"all\"",
+        "beautifulsoup4>=4.12.3; extra == \"all\"",
+        "bottleneck>=1.4.2; extra == \"all\"",
+        "fastparquet>=2024.11.0; extra == \"all\"",
+        "fsspec>=2024.10.0; extra == \"all\"",
+        "gcsfs>=2024.10.0; extra == \"all\"",
+        "html5lib>=1.1; extra == \"all\"",
+        "hypothesis>=6.116.0; extra == \"all\"",
+        "jinja2>=3.1.5; extra == \"all\"",
+        "lxml>=5.3.0; extra == \"all\"",
+        "matplotlib>=3.9.3; extra == \"all\"",
+        "numba>=0.60.0; extra == \"all\"",
+        "numexpr>=2.10.2; extra == \"all\"",
+        "odfpy>=1.4.1; extra == \"all\"",
+        "openpyxl>=3.1.5; extra == \"all\"",
+        "psycopg2>=2.9.10; extra == \"all\"",
+        "pyarrow>=13.0.0; extra == \"all\"",
+        "pyiceberg>=0.8.1; extra == \"all\"",
+        "pymysql>=1.1.1; extra == \"all\"",
+        "PyQt5>=5.15.9; extra == \"all\"",
+        "pyreadstat>=1.2.8; extra == \"all\"",
+        "pytest>=8.3.4; extra == \"all\"",
+        "pytest-xdist>=3.6.1; extra == \"all\"",
+        "python-calamine>=0.3.0; extra == \"all\"",
+        "pytz>=2020.1; extra == \"all\"",
+        "pyxlsb>=1.0.10; extra == \"all\"",
+        "qtpy>=2.4.2; extra == \"all\"",
+        "scipy>=1.14.1; extra == \"all\"",
+        "s3fs>=2024.10.0; extra == \"all\"",
+        "SQLAlchemy>=2.0.36; extra == \"all\"",
+        "tables>=3.10.1; extra == \"all\"",
+        "tabulate>=0.9.0; extra == \"all\"",
+        "xarray>=2024.10.0; extra == \"all\"",
+        "xlrd>=2.0.1; extra == \"all\"",
+        "xlsxwriter>=3.2.0; extra == \"all\"",
+        "zstandard>=0.23.0; extra == \"all\""
+      ],
+      "installer": "pip"
+    },
+    "tokenizers": {
+      "version": "0.22.2",
+      "requires_dist": [
+        "huggingface-hub>=0.16.4,<2.0",
+        "pytest ; extra == 'testing'",
+        "pytest-asyncio ; extra == 'testing'",
+        "requests ; extra == 'testing'",
+        "numpy ; extra == 'testing'",
+        "datasets ; extra == 'testing'",
+        "ruff ; extra == 'testing'",
+        "ty ; extra == 'testing'",
+        "sphinx ; extra == 'docs'",
+        "sphinx-rtd-theme ; extra == 'docs'",
+        "setuptools-rust ; extra == 'docs'",
+        "tokenizers[testing] ; extra == 'dev'"
+      ],
+      "installer": "pip"
+    },
+    "opencv-python-headless": {
+      "version": "5.0.0.93",
+      "requires_dist": [
+        "numpy<2.0; python_version < \"3.9\"",
+        "numpy>=2; python_version >= \"3.9\""
+      ],
+      "installer": "pip"
+    },
+    "huggingface-hub": {
+      "version": "1.33.0",
+      "requires_dist": [
+        "click<9.0.0,>=8.4.2",
+        "filelock>=3.10.0",
+        "fsspec>=2023.5.0",
+        "hf-xet<2.0.0,>=1.6.0; platform_machine == \"x86_64\" or platform_machine == \"amd64\" or platform_machine == \"AMD64\" or platform_machine == \"arm64\" or platform_machine == \"aarch64\"",
+        "httpx<1,>=0.23.0",
+        "packaging>=20.9",
+        "pyyaml>=5.1",
+        "tomli>=1.1.0; python_version < \"3.11\"",
+        "tqdm>=4.42.1",
+        "typing-extensions>=4.1.0",
+        "authlib>=1.3.2; extra == \"oauth\"",
+        "fastapi; extra == \"oauth\"",
+        "httpx; extra == \"oauth\"",
+        "itsdangerous; extra == \"oauth\"",
+        "torch; extra == \"torch\"",
+        "safetensors[torch]; extra == \"torch\"",
+        "toml; extra == \"fastai\"",
+        "fastai>=2.4; extra == \"fastai\"",
+        "fastcore>=1.3.27; extra == \"fastai\"",
+        "hf-xet<2.0.0,>=1.6.0; extra == \"hf-xet\"",
+        "mcp<2.0.0,>=1.9.4; extra == \"mcp\"",
+        "authlib>=1.3.2; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "httpx; extra == \"testing\"",
+        "itsdangerous; extra == \"testing\"",
+        "jedi; extra == \"testing\"",
+        "Jinja2; extra == \"testing\"",
+        "pytest>=8.4.2; extra == \"testing\"",
+        "pytest-cov; extra == \"testing\"",
+        "pytest-env; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-vcr; extra == \"testing\"",
+        "pytest-asyncio; extra == \"testing\"",
+        "pytest-rerunfailures>=16.2; extra == \"testing\"",
+        "pytest-mock; extra == \"testing\"",
+        "urllib3<2.0; extra == \"testing\"",
+        "soundfile; extra == \"testing\"",
+        "Pillow; extra == \"testing\"",
+        "numpy; extra == \"testing\"",
+        "duckdb; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "gradio>=5.0.0; extra == \"gradio\"",
+        "requests; extra == \"gradio\"",
+        "typing-extensions>=4.8.0; extra == \"typing\"",
+        "types-PyYAML; extra == \"typing\"",
+        "types-simplejson; extra == \"typing\"",
+        "types-toml; extra == \"typing\"",
+        "types-tqdm; extra == \"typing\"",
+        "types-urllib3; extra == \"typing\"",
+        "ruff>=0.9.0; extra == \"quality\"",
+        "mypy==1.15.0; extra == \"quality\"",
+        "libcst>=1.4.0; extra == \"quality\"",
+        "ty; extra == \"quality\"",
+        "authlib>=1.3.2; extra == \"all\"",
+        "fastapi; extra == \"all\"",
+        "httpx; extra == \"all\"",
+        "itsdangerous; extra == \"all\"",
+        "jedi; extra == \"all\"",
+        "Jinja2; extra == \"all\"",
+        "pytest>=8.4.2; extra == \"all\"",
+        "pytest-cov; extra == \"all\"",
+        "pytest-env; extra == \"all\"",
+        "pytest-xdist; extra == \"all\"",
+        "pytest-vcr; extra == \"all\"",
+        "pytest-asyncio; extra == \"all\"",
+        "pytest-rerunfailures>=16.2; extra == \"all\"",
+        "pytest-mock; extra == \"all\"",
+        "urllib3<2.0; extra == \"all\"",
+        "soundfile; extra == \"all\"",
+        "Pillow; extra == \"all\"",
+        "numpy; extra == \"all\"",
+        "duckdb; extra == \"all\"",
+        "fastapi; extra == \"all\"",
+        "ruff>=0.9.0; extra == \"all\"",
+        "mypy==1.15.0; extra == \"all\"",
+        "libcst>=1.4.0; extra == \"all\"",
+        "ty; extra == \"all\"",
+        "typing-extensions>=4.8.0; extra == \"all\"",
+        "types-PyYAML; extra == \"all\"",
+        "types-simplejson; extra == \"all\"",
+        "types-toml; extra == \"all\"",
+        "types-tqdm; extra == \"all\"",
+        "types-urllib3; extra == \"all\"",
+        "authlib>=1.3.2; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "httpx; extra == \"dev\"",
+        "itsdangerous; extra == \"dev\"",
+        "jedi; extra == \"dev\"",
+        "Jinja2; extra == \"dev\"",
+        "pytest>=8.4.2; extra == \"dev\"",
+        "pytest-cov; extra == \"dev\"",
+        "pytest-env; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-vcr; extra == \"dev\"",
+        "pytest-asyncio; extra == \"dev\"",
+        "pytest-rerunfailures>=16.2; extra == \"dev\"",
+        "pytest-mock; extra == \"dev\"",
+        "urllib3<2.0; extra == \"dev\"",
+        "soundfile; extra == \"dev\"",
+        "Pillow; extra == \"dev\"",
+        "numpy; extra == \"dev\"",
+        "duckdb; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "ruff>=0.9.0; extra == \"dev\"",
+        "mypy==1.15.0; extra == \"dev\"",
+        "libcst>=1.4.0; extra == \"dev\"",
+        "ty; extra == \"dev\"",
+        "typing-extensions>=4.8.0; extra == \"dev\"",
+        "types-PyYAML; extra == \"dev\"",
+        "types-simplejson; extra == \"dev\"",
+        "types-toml; extra == \"dev\"",
+        "types-tqdm; extra == \"dev\"",
+        "types-urllib3; extra == \"dev\""
+      ],
+      "installer": "pip"
+    },
+    "numpy": {
+      "version": "2.3.5",
+      "requires_dist": [],
+      "installer": "pip"
+    }
+  },
+  "module_presence": {
+    "torch": {
+      "found": false
+    },
+    "torchvision": {
+      "found": false
+    },
+    "torchaudio": {
+      "found": false
+    },
+    "transformers": {
+      "found": false
+    },
+    "tokenizers": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "148ecb122f3fee03be9abb1fe213dd85fbc691bf2ad2b437ff531533781a6a3a"
+    },
+    "huggingface_hub": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "284be6e001a015f9933e756c6f6552b7266d811aff7b5124c960b5f1fc66b27e"
+    },
+    "diffusers": {
+      "found": false
+    },
+    "diffsynth": {
+      "found": false
+    },
+    "accelerate": {
+      "found": false
+    },
+    "peft": {
+      "found": false
+    },
+    "swift": {
+      "found": false
+    },
+    "verl": {
+      "found": false
+    },
+    "vllm": {
+      "found": false
+    },
+    "fla": {
+      "found": false
+    },
+    "tilelang": {
+      "found": false
+    },
+    "triton": {
+      "found": false
+    },
+    "cv2": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "69192904dead3ada37c8b305ee850ac7474924730881a9d5e0b2390944eb5aae"
+    },
+    "numpy": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "93924ac4b793328947dfd9eb9355e54eccdfd26a92c8dd52188aed2e52c7eb38"
+    }
+  },
+  "pip_check": {
+    "exit_code": 0,
+    "output": "No broken requirements found.\n"
+  },
+  "GPU": {
+    "checked": false,
+    "training_or_model_inference_run": false
+  },
+  "original_environment_modified": false,
+  "credentials_included": false
+}

robot/data_cpu/requirements.core.in

@@ -0,0 +1,11 @@
+# Observed role-specific root requirements; transitive dependencies resolved/locked separately.
+numpy==2.3.5
+pillow==12.3.0
+requests==2.34.2
+opencv-python-headless==5.0.0.93
+h5py==3.16.0
+pandas==3.0.6
+pyarrow==25.0.1
+pytest==9.1.1
+tokenizers==0.22.2
+huggingface-hub==1.33.0

robot/data_cpu/requirements.install.lock.txt

@@ -0,0 +1,727 @@
+# This file was autogenerated by uv via the following command:
+#    uv pip compile /pfs/zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in --python-version 3.12 --generate-hashes -o /pfs/zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.install.lock.txt
+anyio==4.16.0 \
+    --hash=sha256:59bb03251907983d1603d27c3087ddb1e93299ea2163adcdba37899a4fb12ffb \
+    --hash=sha256:c66ae634e0b8228396b801741beb3294326d4b6102f7bba4041c0dfb58c29026
+    # via httpx
+certifi==2026.10.23 \
+    --hash=sha256:a9895e6e2758b371e0545e321a3a8fd1fcfd9829ada8824a936dfb16aaaf36cd \
+    --hash=sha256:e9b54de76e530bb6cf0ccd036787103acb58fd42ba930ea8bf948ddbaa2f61d1
+    # via
+    #   httpcore
+    #   httpx
+    #   requests
+charset-normalizer==3.5.2 \
+    --hash=sha256:01077390b03f7988f11d700a2194e69b119741a86b1a638b1db88891e3eced8e \
+    --hash=sha256:01b0c0d2262a9e28e8484a278c7e1b5d650e3ac8cf2683d2967e25899f208bdf \
+    --hash=sha256:04851f73ae72b8413dddadb16a49dfee95263553741fd42d546f7d66907e6be5 \
+    --hash=sha256:0521c5665880b33d603717defa76c094048900010897909952397feb3039da56 \
+    --hash=sha256:0774bf9bf620249fee3e0b8b9fd3065de213be30f3aa94ce2494b3b638949e26 \
+    --hash=sha256:0891b9d3903c5571c03771ca669a4b0ec5618ca722a5c957d3d29cd4e5062848 \
+    --hash=sha256:0c951d5e6dd9c2ff60609476752bee49da4206adde960ebc247766937f72e718 \
+    --hash=sha256:0fed1d06615f022ee3b13caf5e8b180cfea32bb2c5aded8a9d44277afc040f93 \
+    --hash=sha256:114e4d0c92d618409ed82a99e22b5c5e768fe995f2973f78265f4524f49d4640 \
+    --hash=sha256:11912e4bb14baae7c5d8791aa55ba0a3a03ec6729073307b0f57270abaa713d3 \
+    --hash=sha256:11a4d68a6ecda3292cb1e50239e111543ba5d709bb62a6b4ea1afcfa729d8875 \
+    --hash=sha256:124fbf1a8ff966d87ae05bb8bd45a71f966055ed8bba320d0c7cf450bc5f4d0e \
+    --hash=sha256:1461ac396c4fdb983a675f20aa555624f0ee18ac83d832b9244ffff3d8055275 \
+    --hash=sha256:1503bccbeb36d5527790c3930327704c39af22de3112f1b1666a9f3ce15ee204 \
+    --hash=sha256:15bb4005af6320d259dc7593ca84a38d7fe06a421dbcf7b910ae23979101e787 \
+    --hash=sha256:15c44f7edfd477b06f517a5cc317fc1707edb9de2c865f43d4b6513907473234 \
+    --hash=sha256:16fa0eccf81304b79c5cd87f9271c3b85dd9dd99245e4422ae9c0dd45e0f99d3 \
+    --hash=sha256:183b88127acdb4fabe59d951ab424faf1af7b63cdbb5f776186c1ea2ffcaed98 \
+    --hash=sha256:195c26fb65950f8fce54e26349852b7bdd7c5f120aeefbcc440b8a20faaed4a3 \
+    --hash=sha256:1afb975bd5d68d5ce9f6b6d44fdf2f7e34b895a35e95708a7a91b20a3b51d187 \
+    --hash=sha256:1b4cbc7c3491ccb4aa17fcd8165649d01cf39f76de1696da8631b5f71b85401d \
+    --hash=sha256:1bc0baf5ef96b6ede57d47f4b8fe4d9d84019c3bfcbeb20a41edc6a6ee341f1f \
+    --hash=sha256:1c50fe28bbc2ced33386f298650d91218076c05420e6cbd790b913adc41659e7 \
+    --hash=sha256:1db38f4c5496827c1a501846d64d14c3b80c7e6714e406cd7dc36a9899fa1011 \
+    --hash=sha256:211d5a3eb6af8f513b8d4ca19a8c1b7accab1b5f0d3175f9826b03c1a920dc1f \
+    --hash=sha256:23851fb4e1b85ed3f6c2a27b777cdfe2e19fb5b38429a8faf38c7542b7665869 \
+    --hash=sha256:254eb48b9fa5ee9898a3c445825a1f340fe53712a098904b39b0bddba8ea3cb1 \
+    --hash=sha256:2625388c6c754520c37abaf3b41eb34d1cc4a373f457898f08606c8e362b891d \
+    --hash=sha256:281cb91036248400f4cc957495cccd44c275c2e0c5854f7e45ac5cf7dc193847 \
+    --hash=sha256:28a15fdad492a99b6eccfaaed66ef3f74050680545ea61ec8b2f4c538f1f1320 \
+    --hash=sha256:28b4f0d66fb834ff90f28209ac7bce77868c45d8c93e26f906709d9b7c2e1af9 \
+    --hash=sha256:2a925889534b3748302dae5dead07cc13480de1dac3aea80a941b729b471ef93 \
+    --hash=sha256:2b7b3bbfb4fe8ef40600792d762fbaa9057559f9d3fad209525b7a22b99e91fd \
+    --hash=sha256:2c9ad19a6cfcd5ea5c0d41161d22f9df1dcc277e9bef2751391334546a314c00 \
+    --hash=sha256:2cc961b171b3f3440f410489ab3573e86aea8736134ebbb40ea1338b7f0831bc \
+    --hash=sha256:2ce45c6627b22c47e390bc91a41c3d13032192e699fa0bea96e9671b373d69b0 \
+    --hash=sha256:2e06a3a98f916dd41d27f3105e02e7a40181c98c94b9158733d03a6f80506c09 \
+    --hash=sha256:304d5463e65a35d7bb0850550e0780395395f6fcf452f04db7d5ca7cecc425ac \
+    --hash=sha256:304d8e4d493af723536393eee0c689eb7813f4a474c8b479dee63f1fdd98f621 \
+    --hash=sha256:30fcd120b732aa79317f08dee04d7de0847822e4cf7ee0e9f445bb958832252c \
+    --hash=sha256:31f3930700408d211f13378ccbe1c40845d8da54bd0681fac3a9b5aae81c7aa8 \
+    --hash=sha256:34276fd796040bf0993ab33a369aa572e6979c7aab225a88893667ad8eac8f7a \
+    --hash=sha256:355ad8011081dec5412240c087a9a0c9d4d5039f3ed11a3f13e18c2b29b56c51 \
+    --hash=sha256:38a873987f3be698494da8b2e3085e29da02da7b633dce73e79c699a113d7bf0 \
+    --hash=sha256:39de2a259fc954455c57274dc94c79d5842774e1247a016aff30bc0efed0f4ef \
+    --hash=sha256:3d14b50de6bf4d0edf857a9386836846f982b8f524e188e2e68b96d702bcf4aa \
+    --hash=sha256:3d21b8b13c7592db2ac5e544a6d83187b995257472b0c9e8351b6d507ae37ed6 \
+    --hash=sha256:3d31298449090ab8d47b7b1b2a555ff73cac7ed438a08b7ac160980c7ebed649 \
+    --hash=sha256:3ddacd27458c45bdacd6bd6db644bfb730efbf9e830310186e3045c9c5be8fb2 \
+    --hash=sha256:3df041de8887954562c9b261cba85ca0e9ded74048daf125f45edcfaa4832229 \
+    --hash=sha256:40ab6bffa02ae10a0581e6c198be7d2d8ca5c2a0c64e4ed3465d766df457573e \
+    --hash=sha256:4275811936e2f06feff5e598fb42a1b7ae852da8e39605211892b56b81a34efd \
+    --hash=sha256:443eae2bf318abeaf6f15d785138f71fd6de770e99a92158b8b814265e079115 \
+    --hash=sha256:447441e76ec720b15e64418d32e092297340387053047c7c694f579efb0ee1d9 \
+    --hash=sha256:4495c5002a7b28557e7e222e77e0b661183e432b7d6d2e788101e3f240e05b8c \
+    --hash=sha256:44bd4fbb29dfbeba60e7d2bd000c59e4b21ddb3cc53912b14048d37092706d7c \
+    --hash=sha256:4685902cf26edf013ed7a3da0f426ebba7a00ebb9541386d835afbf002c11cab \
+    --hash=sha256:498dc3188ca05a68231ac3fdbfc7f57eb67e1343c30e0fea17f8218c1599b253 \
+    --hash=sha256:4c2b5031f63e331e3839b40aed2dd6f191e9c07edbde303e7876846ea1946995 \
+    --hash=sha256:4d48f2d08b9de5864e2c8744d4461b862fb149a18274abc8b698c45975573438 \
+    --hash=sha256:4f87960d57feabfb618e4e0af6e7371645fa26a277860739d6e5d6e0012c92f0 \
+    --hash=sha256:50e3adfb96fc189eb27b1cf62d3b598b89b4bb0420d93a3d3e42e137409011be \
+    --hash=sha256:51cf45226a9b588d0d2b4880c62d686934b63ab0bd79ca23ab0e9762eb27441b \
+    --hash=sha256:52aa6992700996af31f375de0c6bacd402b0097fe40b53c426b9f51a90ebabc7 \
+    --hash=sha256:55ea99acb17b9325618de155a0cd6a2e8f5d10be008113e1d433bbb58db543b2 \
+    --hash=sha256:56bc200a365efb37383b7852e4cc5898d3b2da5987289b543956cf8cad71018a \
+    --hash=sha256:588461c2e8384d309bd63e5826019b6977bc66d629b99ac8737bb795d7b2cb5a \
+    --hash=sha256:58ca3755ee7ff7f59b57789ec9833c9de9ea275405cdd240eda1f193112e398a \
+    --hash=sha256:58f361dcbab699cf8f42db3f47c8e7fd1036f138c23a5d08de9fde5f425a730c \
+    --hash=sha256:598a11a2c7ebaa5334bf698bf29568c9c390abac6a154d8170fedecd1cea38c5 \
+    --hash=sha256:59f63901b0031c3136cf64704dcb21de0bbae62ce2c9529bc39d27665463de37 \
+    --hash=sha256:5cde776b7cc66e4f6c99612cea4aa7269aa65863f7a15841b2c264f103822f4e \
+    --hash=sha256:5e2b6b57e9733d39f0c9fd3185efa6b8e29652c4cd8fe94180272cf6ed9a78c4 \
+    --hash=sha256:5fb29fb8cd1a46c27a1bf9613ad5ec2599310d46b4025d9556404a6b6a292800 \
+    --hash=sha256:6045373d5a89a5ec71afde535db987ca28e76dfa276c2d4c818265b375d4b055 \
+    --hash=sha256:619799369eeef6366ed3e8755a5670f4f2f0fb6b30a0fd7264dc0fdc2357058e \
+    --hash=sha256:62588a277bfb59def052abd940703fa35107152bf479781a878617d60faf8fb5 \
+    --hash=sha256:62603db9a7caa0802eaa28c1c46fecd7b3a263a774069c24c3c28c302448721c \
+    --hash=sha256:65cd72beeeca9d3aaea1201e5923859f308f952f9c71de93f06063c79f0f7a3b \
+    --hash=sha256:68eb192d85ab8e5f6ec69c2bc6ac0179fbf04a5ac1569d12fbef74883fe102d0 \
+    --hash=sha256:6bd128f206a7752ae1f2ab6c61bf8a24ba28913a10df8b14c2637b973ff97a80 \
+    --hash=sha256:6be488a102b8cf28d0391d8c4ba7748938ae28b78ad901f8585520fca33ead1a \
+    --hash=sha256:7218e8f32b0956cfcd048fd42d9d5779809745ca1d86113ca56f66e7ae1549c4 \
+    --hash=sha256:7441d755b7ab94f8d4eb3e43ec05482d760842fd263d003a99102d742cd835e2 \
+    --hash=sha256:749e97e1b32313717a565abbe321bc2190bc8b35f1a67e4cdbc7c56c8d8ffe58 \
+    --hash=sha256:75a3ceed0724d625d64b86ca20aba182e4df462e04c2414fc941c0f523f06aac \
+    --hash=sha256:780fbe7cab297b81dad9fb8dc5eb003c0468ffb0d9e5f65068c53a34661a96bc \
+    --hash=sha256:78456a747de8dc58360ffa581f30a002baf5aa28cb262536545e91f113ed7639 \
+    --hash=sha256:7967d08cf06dee78443b874f98c98036f624f3a4e73e11f9f64f5be4d25393cf \
+    --hash=sha256:7a881931aa470808df94a8c380eed2bbbc76cd9dc622310f99665658c821eb6d \
+    --hash=sha256:7dcd882da75ef9adf94903b1e3b9419e8aa8fb4c7396822b834b9ef7fb96954f \
+    --hash=sha256:7e841fb9010836c992c9f12fcbd43a831de93a5f726fc1ccd8ca1d0268c5014c \
+    --hash=sha256:7fdde2c9fd9e3eca40631e024664cf2584272cc8f96308cbe5fdfc930f51d8bc \
+    --hash=sha256:8024d00c3faf3fc0c16e07a69f4405e8eac7cc0ab15f65fe6cf43827c4cf72b4 \
+    --hash=sha256:80d02b6f04e92601a081dd97b23d3128033098bff5d35d392ddcc0476ea11253 \
+    --hash=sha256:838dcc90063569a0448120554591a1d6c4a4ffe11babf048908793154ab86ade \
+    --hash=sha256:849df64e889b2e17230d58410a03dba311a65b163508fd33679b2b737d4b7858 \
+    --hash=sha256:87475fabc8d9996fd9c27debb395e642e8c838d78a00b6e932227a0e06b81e26 \
+    --hash=sha256:87e50a3e7cb90af586b6c5faf23e302a970415ac73bd7bd90a515a04b427ef96 \
+    --hash=sha256:89b53f3cda69831909888e0494f4fa0bcd3537e3e138dabeb620bd6ad946bae8 \
+    --hash=sha256:8a893cc101149f80a653f82062ebc95b34525a2614382e1da5458fe7c6997249 \
+    --hash=sha256:8b2bfab86aa71ae13aa41a6a26aab338e0db2b8bc75434b05aea89e011ff35a4 \
+    --hash=sha256:8d86d6fc60743dc916eb79e2eb1ec4818e21e427731543af40a3021851174a13 \
+    --hash=sha256:915563965d418f986e7e145accc592eae9e1a1be3566ff98a05d7a9ec42a76e1 \
+    --hash=sha256:92888bb3187c5ba50500b00b3b310c9f2c651709d28036077680cb5255450a03 \
+    --hash=sha256:93223adc95033dd47133a46ccfc316a0139176fd79085762e27202ec56018f03 \
+    --hash=sha256:9373ad13ef0d2c0fb761e04e55bfdee5a08b52cef2c882c8fbe9935b1517152e \
+    --hash=sha256:9409a8bf35cf78353942504b24a57de3d75b708997a1e4bd8db71ac8633ce364 \
+    --hash=sha256:9b7f416ff0978e2f2249330527f0ad6fa02f4932e6199692d3b52da2048c19e4 \
+    --hash=sha256:9bde855991b7e362c146535e3136a50bfaffc0487d38b33ca7e5edefc6e23849 \
+    --hash=sha256:9cae88599c7219005d879f98e5ed53341e9a122af585e1091200358a3003d2a0 \
+    --hash=sha256:9cf9b1a857e25c4baceeb3624e92a56df3668f398c4acba74e174d81fb4d1d3a \
+    --hash=sha256:9f56f72050826f63dcee7a7f55b0a77168cb3bfc553fd405e7f8f9ece75a4036 \
+    --hash=sha256:a090bb2c68df85450502e3e20d665e3a5af9c65a84d6508ed477badd49166fd3 \
+    --hash=sha256:a192e2c40070d92c3ccf777e3a5c4ff515573cd2bb7ed0c537fdadbbec5bbf21 \
+    --hash=sha256:a19a731138fc27d5682277d3b9df22855cea1239bce7fcec5f78f42ef2d1f3c3 \
+    --hash=sha256:a66c3bc5ab1f0ff2164fc9965ddd611ff0802173f4b9d24554c563f6ab7e1d6e \
+    --hash=sha256:a815775b6c38d4e0ff7bcffbeba67feded90202bb6a226b8dd35f1c855217413 \
+    --hash=sha256:a89012d6d5476ee112d20d998570ed58df2260a852afb1758809cd6900411d21 \
+    --hash=sha256:ae4f5fea5b8b8ccff88238cc8569303e5ee95efae67fa62922a311397a71f346 \
+    --hash=sha256:b6856554c4f44d79fc2307d5768854310a8f0096e501c75637542c82292b0429 \
+    --hash=sha256:b6b751274acb69d77b3323d6b7dbaa3c7fdfc1eb829b7eb61d262f32e1af9685 \
+    --hash=sha256:b736353c0a625bbd5fcec108576e2385db3496f4f771f785ff32e108d3c3bc45 \
+    --hash=sha256:b7fd005a73d9e657273b7a10dc71a9e03c8fb9ee6999798d6918ce095b81ac7f \
+    --hash=sha256:b91363207bd9dc966a691e959bb47f64b30f7ac4b072be9968b366982f7db77c \
+    --hash=sha256:ba0b1d2620edf869789c3879223f52bf2afc5d31b3cb47cc57b3a12c05e2aa9d \
+    --hash=sha256:bbbfc8e28816f19d7c0f1816664980c0a9875d01b27cdf8eedddb639d9e108ad \
+    --hash=sha256:bd16aabe4a02a297c23417aa17ac6299dbd8c49f673bcd645b4929b11f5a4400 \
+    --hash=sha256:c0afc6800ba57ccc350374c5bd6150419915d95ce93cdbab2d783d75eaf30ecb \
+    --hash=sha256:c6708715abcf3c73b99508253e961a9967f02fe536532834149574eda6de0d1c \
+    --hash=sha256:c7c9ab723cde841fefb34efbad91e87f00a674b1fe1cd0784fde742bf2c154dc \
+    --hash=sha256:c8f3d67aeaf55f017982b73683f0e7342ba2f6635a78f69ce89ebb26aa411e5c \
+    --hash=sha256:c9790464842f85f437dbbb54417eda1e0e6bfc52dd8d22d6fd1c994b73b2dc74 \
+    --hash=sha256:ca403d7e4798f525fdfc78e258820419cbbd0f0ecbab9de7840e3c017cf6b8cf \
+    --hash=sha256:d008d90a7f2471519aef0c90dfbe73b3e6e4d5e66ac48e19154c17e89e98b604 \
+    --hash=sha256:d19fbd981a488e22cd04883659ca6b08f50b5974f9fd7c95655ef6a043e5893f \
+    --hash=sha256:d1befeed746d247c81127bb14de9dc3d30edb6e5976d34f83f86ed262b1d9105 \
+    --hash=sha256:d2374b62878abb00cd8309b32af6c0b715cd02dec0ca74ef12e5069bdc64144a \
+    --hash=sha256:d376bbd28b3a8999db1a103b3b388aee6f1ddeb3e51bc2172993efdcd86e064d \
+    --hash=sha256:d4a7319f304a774bed22115bc891618e45f85065ab44ea6acd07d274e750519a \
+    --hash=sha256:d6734d2ef8a50fbf8445c139477da401f50d62a0606bf00e20ec6d87773fefb1 \
+    --hash=sha256:d760fe2a4d7c3b226cb9026d6a842868d52a7901bd98420e1baf14e80da85cf5 \
+    --hash=sha256:d913de495d90407cd859d263bee2e5d1a4ed3eb6573c04e70d9ec619a7cbed7f \
+    --hash=sha256:db19d07e2e0129e974a0e65d0064fc222a446cd5122c2fd4184d2af9fc734a9e \
+    --hash=sha256:dca9ab98072a5a54ebacebdc45f53e645336b320c667410b061be1ca588ae709 \
+    --hash=sha256:ddc7dacc8ece3a182e7f15cb862d1fd616b46d076cb1ae9dd232b2c38b655874 \
+    --hash=sha256:ddf19c062bea7a0cc80f519243d2c01dd091be0cf952a0750d4ad576709559f5 \
+    --hash=sha256:def79fa35ef0cef8d2accec024f4fdc7ead3012ff02f5215c783f39f03ef8cfc \
+    --hash=sha256:df29a0a7107f7011e77f4eebdddec4c7331e24d787a0b21a46d63bdf7445da95 \
+    --hash=sha256:e09a3942ecbdee5cce73ea9d42da82b81b72ac1bf031ce069b93b5adf4eac8cd \
+    --hash=sha256:e242bb1c5e76e97dfa9e7f209a71e93a01d7f19ffdd5cfbb2e2d55b4f08f8ab0 \
+    --hash=sha256:e243bd13217235fc7290c621941c3f5cc8b66e4872495be821d7436ba2fb838d \
+    --hash=sha256:e2af3aad578aa6bd1384bcf4750fc285e5a9de53f40b7d41e5a0bf748edeb2b3 \
+    --hash=sha256:e4e81e09c1578b8df602e3db08b0b3ea0a6947ad612f52bf8dc5ea8d47691f0c \
+    --hash=sha256:e54da4baf05720032d527874d40b65fa4d7e5c6c6a43d0c3adbeffcaf275a2b3 \
+    --hash=sha256:e80e6c2f55656b4824d72065abb4ddd6a525c74bd78a0aab5d9fc2cf4fb5af50 \
+    --hash=sha256:ed2a239c0ea213acc1908150a3037257083c7c083128f1a4cec2ec4b97dca491 \
+    --hash=sha256:ed905975ab14056a2e5eb1c376cb2e1ebc5396baf84163939c518556fccde9f5 \
+    --hash=sha256:ee21e28f0430bd6dc9086c6e525d5e818a44a5ad19720c8a0ef766792f3eb5e5 \
+    --hash=sha256:ee43c17b173d46a3212baa6ead3ae258eeabdae48c263a01ccf0218c366dd655 \
+    --hash=sha256:ef4fcbf3327382cd4c9f540babd61248208af7b93eec4de397b4d5f58a09e288 \
+    --hash=sha256:eff0ac9dbe711a4aee69bf04a83896aa9b85f19641264053a9f6d48573abb7dd \
+    --hash=sha256:f0aa869112ef88429ae17820d99c3dd9504c9e9c671d3c246f3d7442cb051084 \
+    --hash=sha256:f3c96f633825733f735c5a9cf21d21a257d8e1edf0b1cee0a064b9c424ca0f7d \
+    --hash=sha256:f5833ad231be5eb6553de524a70f48d71b2c8563101750531e0b80184e175cd4 \
+    --hash=sha256:f5ec61164adcec446f8969a3358ec3f9b26bbda3b9213e5586d219afa8df2915 \
+    --hash=sha256:f7d486c83842422badd511868fd8a9a20e9407ace71564b6af47ce7e60a336c1 \
+    --hash=sha256:fb9e68df06293761f9fe66ade60a9bc6d0f5e42b8acf2939a9158af86ab0e5bd \
+    --hash=sha256:fc14a032f813bf5fe624d991960ea83e9715adc27e4c1830a2361eb1d02ac341 \
+    --hash=sha256:fcff63213e8e6e47770541a4607175404f47cbb3ebea7b6058cc82d524a0e424 \
+    --hash=sha256:fd1fbe0f116b6e55da77aca2c6ddcddcfac2186cbf78bdebf40fc156efca389d \
+    --hash=sha256:fe9753dfee015c570d73df76f899f18444d41388bffcde097deba51c4fadbb9f
+    # via requests
+click==8.5.0 \
+    --hash=sha256:255bc9599cf7748b4b1a446ccc735421bd08a2ae529a8b88597d3de5664ee360 \
+    --hash=sha256:ba0d2089de75ea0310e2dde03160e6ca10009947fb95a182f9b54021bb272e34
+    # via huggingface-hub
+filelock==4.2.8 \
+    --hash=sha256:23df640ac0c3351f8d15c9739072d63be9078758b7d102a2183277ceb96605a3 \
+    --hash=sha256:c2be657e2c279cb032af8e4ea8ce2e20aa982e540be77fcaa34910d068ac2510
+    # via huggingface-hub
+fsspec==2026.9.0 \
+    --hash=sha256:0f08147951c8cb31d844c3547d631053b127863b60be04cf06e121333ee0e2fe \
+    --hash=sha256:8dd6e646e99ea382bd85f97a45e6b526a442d79423a7dc673f1e2756d05fcb5f
+    # via huggingface-hub
+h11==0.16.0 \
+    --hash=sha256:4e35b956cf45792e4caa5885e69fba00bdbc6ffafbfa020300e549b208ee5ff1 \
+    --hash=sha256:63cf8bbe7522de3bf65932fda1d9c2772064ffb3dae62d55932da54b31cb6c86
+    # via httpcore
+h5py==3.16.0 \
+    --hash=sha256:099f2525c9dcf28de366970a5fb34879aab20491589fa89ce2863a84218bb524 \
+    --hash=sha256:0f456f556e4e2cebeebd9d66adf8dc321770a42593494a0b6f0af54a7567b242 \
+    --hash=sha256:15922e485844f77c0b9d275396d435db3baa58292a9c2176a386e072e0cf2491 \
+    --hash=sha256:1677ad48b703f44efc9ea0c3ab284527f81bc4f318386aaaebc5fede6bbae56f \
+    --hash=sha256:171038f23bccddfc23f344cadabdfc9917ff554db6a0d417180d2747fe4c75a7 \
+    --hash=sha256:17d1f1630f92ad74494a9a7392ab25982ce2b469fc62da6074c0ce48366a2999 \
+    --hash=sha256:1897a771a7f40d05c262fc8f37376ec37873218544b70216872876c627640f63 \
+    --hash=sha256:18f2bbcd545e6991412253b98727374c356d67caa920e68dc79eab36bf5fedad \
+    --hash=sha256:2b2c02b0a160faed5fb33f1ba8a264a37ee240b22e049ecc827345d0d9043074 \
+    --hash=sha256:314b6054fe0b1051c2b0cb2df5cbdab15622fb05e80f202e3b6a5eee0d6fe365 \
+    --hash=sha256:346df559a0f7dcb31cf8e44805319e2ab24b8957c45e7708ce503b2ec79ba725 \
+    --hash=sha256:370a845f432c2c9619db8eed334d1e610c6015796122b0e57aa46312c22617d9 \
+    --hash=sha256:39c2838fb1e8d97bcf1755e60ad1f3dd76a7b2a475928dc321672752678b96db \
+    --hash=sha256:3e6cb3387c756de6a9492d601553dffea3fe11b5f22b443aac708c69f3f55e16 \
+    --hash=sha256:3fae9197390c325e62e0a1aa977f2f62d994aa87aab182abbea85479b791197c \
+    --hash=sha256:42108e93326c50c2810025aade9eac9d6827524cdccc7d4b75a546e5ab308edb \
+    --hash=sha256:42b012933a83e1a558c673176676a10ce2fd3759976a0fedee1e672d1e04fc9d \
+    --hash=sha256:43259303989ac8adacc9986695b31e35dba6fd1e297ff9c6a04b7da5542139cc \
+    --hash=sha256:4c6ab014ab704b4feaa719ae783b86522ed0bf1f82184704ed3c9e4e3228796e \
+    --hash=sha256:656f00e4d903199a1d58df06b711cf3ca632b874b4207b7dbec86185b5c8c7d4 \
+    --hash=sha256:698dd69291272642ffda44a0ecd6cd3bda5faf9621452d255f57ce91487b9794 \
+    --hash=sha256:719439d14b83f74eeb080e9650a6c7aa6d0d9ea0ca7f804347b05fac6fbf18af \
+    --hash=sha256:7c4dd4cf5f0a4e36083f73172f6cfc25a5710789269547f132a20975bfe2434c \
+    --hash=sha256:7e420b539fb6023a259a1b14d4c9f6df8cf50d7268f48e161169987a57b737ff \
+    --hash=sha256:8389e13a1fd745ad2856873e8187fd10268b2d9677877bb667b41aebd771d8b7 \
+    --hash=sha256:85b9c49dd58dc44cf70af944784e2c2038b6f799665d0dcbbc812a26e0faa859 \
+    --hash=sha256:86385ea895508220b8a7e45efa428aeafaa586bd737c7af9ee04661d8d84a10d \
+    --hash=sha256:8975273c2c5921c25700193b408e28d6bdd0111c37468b2d4e25dcec4cd1d84d \
+    --hash=sha256:8c1eff849cdd53cbc73c214c30ebdb6f1bb8b64790b4b4fc36acdb5e43570210 \
+    --hash=sha256:9300ad32dea9dfc5171f94d5f6948e159ed93e4701280b0f508773b3f582f402 \
+    --hash=sha256:96b422019a1c8975c2d5dadcf61d4ba6f01c31f92bbde6e4649607885fe502d6 \
+    --hash=sha256:9c9d307c0ef862d1cd5714f72ecfafe0a5d7529c44845afa8de9f46e5ba8bd65 \
+    --hash=sha256:a0dbaad796840ccaa67a4c144a0d0c8080073c34c76d5a6941d6818678ef2738 \
+    --hash=sha256:a6fbc5367d4046801f9b7db9191b31895f22f1c6df1f9987d667854cac493538 \
+    --hash=sha256:bdef06507725b455fccba9c16529121a5e1fbf56aa375f7d9713d9e8ff42454d \
+    --hash=sha256:c3f0a0e136f2e95dd0b67146abb6668af4f1a69c81ef8651a2d316e8e01de447 \
+    --hash=sha256:c5313566f4643121a78503a473f0fb1e6dcc541d5115c44f05e037609c565c4d \
+    --hash=sha256:df02dd29bd247f98674634dfe41f89fd7c16ba3d7de8695ec958f58404a4e618 \
+    --hash=sha256:dfc21898ff025f1e8e67e194965a95a8d4754f452f83454538f98f8a3fcb207e \
+    --hash=sha256:e06f864bedb2c8e7c1358e6c73af48519e317457c444d6f3d332bb4e8fa6d7d9 \
+    --hash=sha256:e2c04d129f180019e216ee5f9c40b78a418634091c8782e1f723a6ca3658b965 \
+    --hash=sha256:e4360f15875a532bc7b98196c7592ed4fc92672a57c0a621355961cafb17a6dd \
+    --hash=sha256:ec86d4fffd87a0f4cb3d5796ceb5a50123a2a6d99b43e616e5504e66a953eca3 \
+    --hash=sha256:fa48993a0b799737ba7fd21e2350fa0a60701e58180fae9f2de834bc39a147ab \
+    --hash=sha256:faca8fb4e4319c09d83337adc80b2ca7d5c5a343c2d6f1b6388f32cfecca13c1 \
+    --hash=sha256:fb1720028d99040792bb2fb31facb8da44a6f29df7697e0b84f0d79aff2e9bd3 \
+    --hash=sha256:ff24039e2573297787c3063df64b60aab0591980ac898329a08b0320e0cf2527 \
+    --hash=sha256:ffbab2fedd6581f6aa31cf1639ca2cb86e02779de525667892ebf4cc9fd26434
+    # via -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+hf-xet==1.8.0 \
+    --hash=sha256:07a06d7481702616726c2d8578cfb00052a5b3f21beed6a674592bc837a10330 \
+    --hash=sha256:10ed23b51dacf7917d64f335c25048deb470efa755e9989b218fc0531386138e \
+    --hash=sha256:1ebaf248c9bef8bb0c9d1d8d71db1fdf0cac14f922cc7a936a81adbf12eb5f01 \
+    --hash=sha256:23c262e9f5e137b18dadbce60becfeea84156b0d43119c320bffe260d32cd2fa \
+    --hash=sha256:29dab7c904cd1914a0439c6e1306c12c7497c8bd32290fbe563ebf51ca4b6d6d \
+    --hash=sha256:3e8934784425781f0f779b91044df441f3a7af4dab47375d7e3aeaecbb8eecb4 \
+    --hash=sha256:3ecfe2934448cca55a4c42732b60f985266aecb241905945219b3d388aceb7ba \
+    --hash=sha256:42b66fdc688eaa4ff3eb77dc9275e5a0f1258ecc9fdd083ddaf43bb49bec925b \
+    --hash=sha256:56735c2d4f416d7473fc986586b26a8756e067832ca4b7fdaad32a0d6f2ce740 \
+    --hash=sha256:62b056aedb89ac809f78940baf465fe553b70c8b25ee56890fbab3039fffec1b \
+    --hash=sha256:7edeb21c5d381903ee5f4857fc746fa7b52a4b374035e8811ecfb67ee27add0c \
+    --hash=sha256:8800e3444770f2c630b361bc6d5c6ebb770b7830646cf632fbb79d0e5bbc9e7f \
+    --hash=sha256:92188598dedc1e1f5609bd718d4db8687a6c89915553912571e2f587892eecec \
+    --hash=sha256:a584a7b3268da62f23c3187cc7201ca01f7b610985a36b1b6119c66bccb0383c \
+    --hash=sha256:c74e3751cbcdae9222c5170e8d788e83201a03dc5e5f4308a20f17a5b6023b56 \
+    --hash=sha256:e00d8f60170cd3dd55d98306e41dbaf28b27ff1bcfefbfa84279d197221be83e \
+    --hash=sha256:f42940869572dc2b05dfe06f4cc398c6291503e086ddc4eddfb7a6ab9409ad01
+    # via huggingface-hub
+httpcore==1.0.9 \
+    --hash=sha256:2d400746a40668fc9dec9810239072b40b4484b640a8c38fd654a024c7a1bf55 \
+    --hash=sha256:6e34463af53fd2ab5d807f399a9b45ea31c3dfa2276f15a2c3f00afff6e176e8
+    # via httpx
+httpx==0.28.1 \
+    --hash=sha256:75e98c5f16b0f35b567856f597f06ff2270a374470a5c2392242528e3e3e42fc \
+    --hash=sha256:d909fcccc110f8c7faf814ca82a9a4d816bc5a6dbfea25d6591d6985b8ba59ad
+    # via huggingface-hub
+huggingface-hub==1.33.0 \
+    --hash=sha256:04e434b06e100eddbce9a6e817d72693a7884b10a79bd67ab48080d5c07eb899 \
+    --hash=sha256:367be21a201db9523eddf8aeac7048f2602c1b308691c97640d5e72ed188007e
+    # via
+    #   -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+    #   tokenizers
+idna==3.20 \
+    --hash=sha256:a7db850025b95ded1eae8a46181a1a6c56c92c96f0e2b005d9ff8dc0210cab44 \
+    --hash=sha256:ab7ae7122974553370f0bdb919e1a960b2cd1bc1ef0276416d896db81c14582c
+    # via
+    #   anyio
+    #   httpx
+    #   requests
+iniconfig==2.3.1 \
+    --hash=sha256:67f4b9c50da0dedf52af349e7749a80a9057a5031199791b906c3bb3ae878960 \
+    --hash=sha256:9121e2c1fdb355232495be3194c8dfe87ccc2d5dee45947b78e68f499790d7a7
+    # via pytest
+numpy==2.3.5 \
+    --hash=sha256:00dc4e846108a382c5869e77c6ed514394bdeb3403461d25a829711041217d5b \
+    --hash=sha256:0472f11f6ec23a74a906a00b48a4dcf3849209696dff7c189714511268d103ae \
+    --hash=sha256:04822c00b5fd0323c8166d66c701dc31b7fbd252c100acd708c48f763968d6a3 \
+    --hash=sha256:052e8c42e0c49d2575621c158934920524f6c5da05a1d3b9bab5d8e259e045f0 \
+    --hash=sha256:09a1bea522b25109bf8e6f3027bd810f7c1085c64a0c7ce050c1676ad0ba010b \
+    --hash=sha256:0cd00b7b36e35398fa2d16af7b907b65304ef8bb4817a550e06e5012929830fa \
+    --hash=sha256:0d8163f43acde9a73c2a33605353a4f1bc4798745a8b1d73183b28e5b435ae28 \
+    --hash=sha256:1062fde1dcf469571705945b0f221b73928f34a20c904ffb45db101907c3454e \
+    --hash=sha256:11e06aa0af8c0f05104d56450d6093ee639e15f24ecf62d417329d06e522e017 \
+    --hash=sha256:17531366a2e3a9e30762c000f2c43a9aaa05728712e25c11ce1dbe700c53ad41 \
+    --hash=sha256:1978155dd49972084bd6ef388d66ab70f0c323ddee6f693d539376498720fb7e \
+    --hash=sha256:1ed1ec893cff7040a02c8aa1c8611b94d395590d553f6b53629a4461dc7f7b63 \
+    --hash=sha256:2dcd0808a421a482a080f89859a18beb0b3d1e905b81e617a188bd80422d62e9 \
+    --hash=sha256:2e2eb32ddb9ccb817d620ac1d8dae7c3f641c1e5f55f531a33e8ab97960a75b8 \
+    --hash=sha256:2feae0d2c91d46e59fcd62784a3a83b3fb677fead592ce51b5a6fbb4f95965ff \
+    --hash=sha256:3095bdb8dd297e5920b010e96134ed91d852d81d490e787beca7e35ae1d89cf7 \
+    --hash=sha256:30bc11310e8153ca664b14c5f1b73e94bd0503681fcf136a163de856f3a50139 \
+    --hash=sha256:3101e5177d114a593d79dd79658650fe28b5a0d8abeb8ce6f437c0e6df5be1a4 \
+    --hash=sha256:396084a36abdb603546b119d96528c2f6263921c50df3c8fd7cb28873a237748 \
+    --hash=sha256:3997b5b3c9a771e157f9aae01dd579ee35ad7109be18db0e85dbdbe1de06e952 \
+    --hash=sha256:414802f3b97f3c1eef41e530aaba3b3c1620649871d8cb38c6eaff034c2e16bd \
+    --hash=sha256:51c1e14eb1e154ebd80e860722f9e6ed6ec89714ad2db2d3aa33c31d7c12179b \
+    --hash=sha256:51c55fe3451421f3a6ef9a9c1439e82101c57a2c9eab9feb196a62b1a10b58ce \
+    --hash=sha256:5ee6609ac3604fa7780e30a03e5e241a7956f8e2fcfe547d51e3afa5247ac47f \
+    --hash=sha256:612a95a17655e213502f60cfb9bf9408efdc9eb1d5f50535cc6eb365d11b42b5 \
+    --hash=sha256:6203fdf9f3dc5bdaed7319ad8698e685c7a3be10819f41d32a0723e611733b42 \
+    --hash=sha256:63c0e9e7eea69588479ebf4a8a270d5ac22763cc5854e9a7eae952a3908103f7 \
+    --hash=sha256:66f85ce62c70b843bab1fb14a05d5737741e74e28c7b8b5a064de10142fad248 \
+    --hash=sha256:6cf9b429b21df6b99f4dee7a1218b8b7ffbbe7df8764dc0bd60ce8a0708fed1e \
+    --hash=sha256:70b37199913c1bd300ff6e2693316c6f869c7ee16378faf10e4f5e3275b299c3 \
+    --hash=sha256:727fd05b57df37dc0bcf1a27767a3d9a78cbbc92822445f32cc3436ba797337b \
+    --hash=sha256:74ae7b798248fe62021dbf3c914245ad45d1a6b0cb4a29ecb4b31d0bfbc4cc3e \
+    --hash=sha256:784db1dcdab56bf0517743e746dfb0f885fc68d948aba86eeec2cba234bdf1c0 \
+    --hash=sha256:86945f2ee6d10cdfd67bcb4069c1662dd711f7e2a4343db5cecec06b87cf31aa \
+    --hash=sha256:86d835afea1eaa143012a2d7a3f45a3adce2d7adc8b4961f0b362214d800846a \
+    --hash=sha256:872a5cf366aec6bb1147336480fef14c9164b154aeb6542327de4970282cd2f5 \
+    --hash=sha256:8b973c57ff8e184109db042c842423ff4f60446239bd585a5131cc47f06f789d \
+    --hash=sha256:8cba086a43d54ca804ce711b2a940b16e452807acebe7852ff327f1ecd49b0d4 \
+    --hash=sha256:8f7f0e05112916223d3f438f293abf0727e1181b5983f413dfa2fefc4098245c \
+    --hash=sha256:900218e456384ea676e24ea6a0417f030a3b07306d29d7ad843957b40a9d8d52 \
+    --hash=sha256:93eebbcf1aafdf7e2ddd44c2923e2672e1010bddc014138b229e49725b4d6be5 \
+    --hash=sha256:9c75442b2209b8470d6d5d8b1c25714270686f14c749028d2199c54e29f20b4d \
+    --hash=sha256:9ee2197ef8c4f0dfe405d835f3b6a14f5fee7782b5de51ba06fb65fc9b36e9f1 \
+    --hash=sha256:a414504bef8945eae5f2d7cb7be2d4af77c5d1cb5e20b296c2c25b61dff2900c \
+    --hash=sha256:a4b9159734b326535f4dd01d947f919c6eefd2d9827466a696c44ced82dfbc18 \
+    --hash=sha256:a80afd79f45f3c4a7d341f13acbe058d1ca8ac017c165d3fa0d3de6bc1a079d7 \
+    --hash=sha256:aa5bc7c5d59d831d9773d1170acac7893ce3a5e130540605770ade83280e7188 \
+    --hash=sha256:acfd89508504a19ed06ef963ad544ec6664518c863436306153e13e94605c218 \
+    --hash=sha256:aeffcab3d4b43712bb7a60b65f6044d444e75e563ff6180af8f98dd4b905dfd2 \
+    --hash=sha256:afaffc4393205524af9dfa400fa250143a6c3bc646c08c9f5e25a9f4b4d6a903 \
+    --hash=sha256:b0c7088a73aef3d687c4deef8452a3ac7c1be4e29ed8bf3b366c8111128ac60c \
+    --hash=sha256:b46b4ec24f7293f23adcd2d146960559aaf8020213de8ad1909dba6c013bf89c \
+    --hash=sha256:b501b5fa195cc9e24fe102f21ec0a44dffc231d2af79950b451e0d99cea02234 \
+    --hash=sha256:bf06bc2af43fa8d32d30fae16ad965663e966b1a3202ed407b84c989c3221e82 \
+    --hash=sha256:c804e3a5aba5460c73955c955bdbd5c08c354954e9270a2c1565f62e866bdc39 \
+    --hash=sha256:c8a9958e88b65c3b27e22ca2a076311636850b612d6bbfb76e8d156aacde2aaf \
+    --hash=sha256:cc0a57f895b96ec78969c34f682c602bf8da1a0270b09bc65673df2e7638ec20 \
+    --hash=sha256:cc8920d2ec5fa99875b670bb86ddeb21e295cb07aa331810d9e486e0b969d946 \
+    --hash=sha256:ccc933afd4d20aad3c00bcef049cb40049f7f196e0397f1109dba6fed63267b0 \
+    --hash=sha256:ce581db493ea1a96c0556360ede6607496e8bf9b3a8efa66e06477267bc831e9 \
+    --hash=sha256:d0f23b44f57077c1ede8c5f26b30f706498b4862d3ff0a7298b8411dd2f043ff \
+    --hash=sha256:d21644de1b609825ede2f48be98dfde4656aefc713654eeee280e37cadc4e0ad \
+    --hash=sha256:d6889ec4ec662a1a37eb4b4fb26b6100841804dac55bd9df579e326cdc146227 \
+    --hash=sha256:de5672f4a7b200c15a4127042170a694d4df43c992948f5e1af57f0174beed10 \
+    --hash=sha256:e6a0bc88393d65807d751a614207b7129a310ca4fe76a74e5c7da5fa5671417e \
+    --hash=sha256:ed89927b86296067b4f81f108a2271d8926467a8868e554eaf370fc27fa3ccaf \
+    --hash=sha256:ee3888d9ff7c14604052b2ca5535a30216aa0a58e948cdd3eeb8d3415f638769 \
+    --hash=sha256:f0963b55cdd70fad460fa4c1341f12f976bb26cb66021a5580329bd498988310 \
+    --hash=sha256:f16417ec91f12f814b10bafe79ef77e70113a2f5f7018640e7425ff979253425 \
+    --hash=sha256:f28620fe26bee16243be2b7b874da327312240a7cdc38b769a697578d2100013 \
+    --hash=sha256:f4255143f5160d0de972d28c8f9665d882b5f61309d8362fdd3e103cf7bf010c \
+    --hash=sha256:ffac52f28a7849ad7576293c0cb7b9f08304e8f7d738a8cb8a90ec4c55a998eb \
+    --hash=sha256:ffe22d2b05504f786c867c8395de703937f934272eb67586817b46188b4ded6d \
+    --hash=sha256:fffe29a1ef00883599d1dc2c51aa2e5d80afe49523c261a74933df395c15c520
+    # via
+    #   -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+    #   h5py
+    #   opencv-python-headless
+    #   pandas
+opencv-python-headless==5.0.0.93 \
+    --hash=sha256:030ca5e0837a2963ab36ef896baa9767eb8d2b83353fb28af5a521e40dd8756f \
+    --hash=sha256:09a872a157c1376ab922a69bbf22f9a95bcc7b658a9d8b436a60212b02b2eeb4 \
+    --hash=sha256:10818d91510e05c04568ae12b5cd120779c70c01bf897b001a6221fe430df80f \
+    --hash=sha256:1e55af3abfb462eeeabe5c775f12bdb36216d8a93a3583d69e6bd6e1d6ba7d00 \
+    --hash=sha256:829717b6a95554f273e49e357cee3b3a2a26b6f4842fbc1bed2b45bdd8f87e0e \
+    --hash=sha256:840bd717c21e5c11cadadc022a823315ea417f961213d06b4df010e019eb16f4 \
+    --hash=sha256:b82f9831daab90b725c7c1ee1b36cb5732c367096ac76d119e64e14eb70d5f3c \
+    --hash=sha256:c6bcd96b185975ea240d22cfdb15a1f6d080cc95264cfbe2621f21bb144d89b9 \
+    --hash=sha256:ed709fdf9aa0bd1f2ed8549e71d19449b03a675bb581eb292285f6861953be37
+    # via -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+packaging==26.3 \
+    --hash=sha256:94edc256424af38762eb31306eed28beb9f0efc50a8837492c9d6fd6004aed79 \
+    --hash=sha256:d7193f7c8e4e93f444fde0262bf90af30e16fa0ad0ad44cb553c87339b23cd1c
+    # via
+    #   huggingface-hub
+    #   pytest
+pandas==3.0.6 \
+    --hash=sha256:0704044b676496b8350e023b09f174a26772456c974a2b11c36bebb558c9490d \
+    --hash=sha256:085e3786ae6b2e82b406266bce36690f72b9dc1421903ba9296b2981a9fcf586 \
+    --hash=sha256:097090508a1dd335013d39106fc10b20f4fd4a171638e47b77d55798ed9dab6c \
+    --hash=sha256:1bcb3e9ed29e74a7439cedff9e2aefd3ea65de84d7de9ccb6c194192541bd60e \
+    --hash=sha256:1e7c0afdcaf6661d795fcefc2f647ddd1136f62cdc153fba177c685d97a87808 \
+    --hash=sha256:1e92d9fa834c7d877130027cddc0cad8dcff97c1f6cca26bd6310f847228b658 \
+    --hash=sha256:22172a92e7ee678ec0140c7af4fc9366b55413834a1cd86af78b3caa0b0574de \
+    --hash=sha256:253e12cb9081b0afbac607920f6142975966bc315135e09de275fdbaa415d2de \
+    --hash=sha256:265f562fdd1079f69f3de96dd425c3405224038c0af4f920c54bd240ee2c4640 \
+    --hash=sha256:2a8fc94be2ee5f1d86f97aacd8cc566f81680b6498e76f3007421bb5d98151bf \
+    --hash=sha256:2e5fa32ff162dfdbc280157d664f44d23049ae414725af9676df339c501d82cd \
+    --hash=sha256:3ef908d28590b3f42d7070e7ad8f9b34b442b260b7f3c1afb57e0040c58cdb1b \
+    --hash=sha256:429d9df32731ab01383ed98f2baa7a60368090d1a94fc06019a12062510e8630 \
+    --hash=sha256:47121f9571503f724c9b93e297ab6254ac99c77adf5e9ed085ea419fd585c258 \
+    --hash=sha256:4e25e2e1adee99ddfada6f7206a79ae8e9c8a8861b0e3eaaba165006d3eef18e \
+    --hash=sha256:4ff44b2cb51cbd691c91f92c4ea6c71e34003f239ebd67c2e857dc898466b49c \
+    --hash=sha256:50c44cbf5820b6b91a5f74aae04972472aefadd3cd9fbd1010409d85528bd570 \
+    --hash=sha256:569e114072b24fc4970c12e2b4bab252671668a40b324318903380cab0254c0c \
+    --hash=sha256:583be68728a31d0d750d5b8d9e00f02b153df0d4655f858bde93cb84cfc4227c \
+    --hash=sha256:5e75072773c1b2f7cb63faa3a6f562aede11f3976f68ed34cb538bc091a28171 \
+    --hash=sha256:5edd0a7abb0986ecce1ac81f56d99b6763f86aa6946dceb6c661224f90af5a19 \
+    --hash=sha256:60d81f9e1799b36f3739e7fff44d1fbb2e8fd5a271b3863e03de9715fccda0fa \
+    --hash=sha256:62f51d7f651c8054c5e82a69265c98082e795d1442df7ca6edc3a545d61214b1 \
+    --hash=sha256:654aae059295dbba6ecd2328ca12712a2cf1676214c8699f1c29213f7ccf9c34 \
+    --hash=sha256:66b07ef7315a31bfe1089cd3d71a7de781c9dca986762d0b4fe7c0ef17465d10 \
+    --hash=sha256:6ff482fa91fa2bafd92e8fe66ce3645c851824310f295c1f0a2f96e928fc4541 \
+    --hash=sha256:77ccbe5057aece6fc172b9b77f19c04335af6882bc2e10c8f3ee4e6bfb3da553 \
+    --hash=sha256:7dac2d65e9087e8e7b5a45fe15c4920911a221df061ab629943ce016489145c7 \
+    --hash=sha256:83e91d15738d7783c050197cef2f2cf82fc6353dae9865aa87ed1fa16aa4d55a \
+    --hash=sha256:86fa853a12e0b70927e2b1ee00d56d2224ec9cbb4b9d58348b5ad52d2f21150e \
+    --hash=sha256:8fe77b408d82e2615674dfed62533b95e18a03610573877422aada4f625d4947 \
+    --hash=sha256:963ca21199097a84c7827c4678b04e30833084fbf8ef44fde3fa7180a29f8fa0 \
+    --hash=sha256:97274c9adf6255bb48c620cd6959805efa7f09ea2167f0e0ae006a448cd2fca7 \
+    --hash=sha256:994a79608263fe1c14cc48ffa7300e2b834b7d1cb406ffe96a08828cb0cdd79b \
+    --hash=sha256:9ae8073aed8e21d1a7fe263dcdc6840743549722a6738198a0a46000fa9476f2 \
+    --hash=sha256:9dab635a549e58a053c7b0fa054dc0bd7be22f0ed9a720f4a85d5fb993276172 \
+    --hash=sha256:9e492cd4bdba6778de4fe0df7f4590c012161ebcf9902dce01b01dc683105514 \
+    --hash=sha256:a3a22e07fe75347eaacc75b0e85297947af4fba6b4aae23916bd8b6828d0bba3 \
+    --hash=sha256:a4dbd4dc65cbe645b92b8785d0f96dd7311010dc6606cf620e51b07b8788a12a \
+    --hash=sha256:a77a1a44e4d88f1c6a2a64d3eb12efec8420875722e14279800b173a7c7c2804 \
+    --hash=sha256:b27c8d890e4aa2171437ae2a39de1d215e674158e4865c4023a8b31c932513b2 \
+    --hash=sha256:bd75ed0c840f709fc2ae26ddd9534ac77ca1a48ac0cce521a74acaa85f3340a7 \
+    --hash=sha256:c6e4aae3e9bea26c6c9a20d88d96c86ec4a99b4db5fd516bcb4e829ab2c0ee36 \
+    --hash=sha256:c826e9babb7790142c399f58599d8de679bea059d7b39c5b6efa2096fac37266 \
+    --hash=sha256:cc39303913e2ea129915670de5d1c9fbd647f543bb72e5543bac8baa94e9e42f \
+    --hash=sha256:d7564d86a94c2eb8ab290b07f63ddaae5c032fa53897c29a2ff2197d43aee8af \
+    --hash=sha256:d7dcd21238cbb4828ff148481ba01cac8946dc5121457b5aeba28636f8f99a60 \
+    --hash=sha256:db7ec631f26223beee8e5c9e0b8f23c24d8197bbd1d982421d4e3188bea51965 \
+    --hash=sha256:e3dccb584123b399c07562ac4d62543e90ede49ddf8ce3c13ffc64cbe828c281 \
+    --hash=sha256:e7c1905ef02c3d6d43d9dbd5b6ccb4da4870a0b0c821bbc103fbdb6f3ad2707b \
+    --hash=sha256:eb6900de08ac85f93ac4948aa6b80842eba555875337b8359035ac9c43e92d34 \
+    --hash=sha256:ee913a91669056c1de1a6b733fbfeab711de9e54e3bee2dfa5fe79d9457247d1 \
+    --hash=sha256:ef738d71d1059245b6bb03e312be06d8b3821326a83486c1ad03b9aba3710e44 \
+    --hash=sha256:f3ce8a6968045481e91a3990e797e348ce13db45ee164a7095bbc824e26c09dd \
+    --hash=sha256:f4e7c52eb108d752e7592268108fd3e98efd76d83a3125cdd06c621c2e44359b \
+    --hash=sha256:f8029ec0f1f89e4f985929ce1f6626dabf3140d61a4e9c1215afdab34eaf9a5d \
+    --hash=sha256:fb625f426b375bcc96e3a04c5d5d266cd7be6ae5d6866e0e703382ab5164068c \
+    --hash=sha256:ff51a4459ed036e93d1eb1bb5e6e7b28685d3cb6b7c12b91c05b31024e234729
+    # via -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+pillow==12.3.0 \
+    --hash=sha256:00808c5e14ef63ac5161091d242999076604ff74b883423a11e5d7bbb38bf756 \
+    --hash=sha256:04f01d28a6aaff387bf842a13be313df23ba0597a44f1a976c9feb3c6ff4711a \
+    --hash=sha256:06ff022112bc9cbf83b60f8e028d94ad87b60621706487e65f673de61610ab59 \
+    --hash=sha256:0740a512dc522224c77d9aa5a8d70d8b7d73fb91f2c21125d8d025d3b8990e45 \
+    --hash=sha256:0847a763afefb695bc912d7c131e7e0632d4edc1d8698f58ddabec8e46b8b6d3 \
+    --hash=sha256:0dd2064cbc55aaec028ef5fbb60fa47bb6c3e7918e07ff17935284b227a9d2df \
+    --hash=sha256:0feb2e9d6ad6c9e3c06effe9d00f3f1e618a6643273576b016f591e9315a7139 \
+    --hash=sha256:10e41f0fbf1eec8cfd234b8fe17a4caac7c9d0db4c204d3c173a8f9f6ef3232b \
+    --hash=sha256:1182d52bc2d5e5d7d0949503aa7e36d12f42205dc287e4883f407b1988820d39 \
+    --hash=sha256:164b31cd1a0490ab6efae01aa5df49da7061be0af1b30e035b6e9a1bfe34ee6e \
+    --hash=sha256:1657923d2d45afb66526e5b933e5b3052e6bdea196c90d3abb2424e18c77dae8 \
+    --hash=sha256:186941b6aef820ad110fb01fb06eb925374dc3a21b17e37ec9a53b250c6fe2d1 \
+    --hash=sha256:1cca606cd25738df4ed873d5ad46bbdb3d83b5cbca291f6b4ff13a4df6b0bbe8 \
+    --hash=sha256:21900ce7ba264168cd50defae43cd75d25c833ad4ad6e73ffc5596d12e25ac89 \
+    --hash=sha256:236ff70b9312fb68943c703aa842ca6a758abfa45ac187a5e7c1452e96ef72b5 \
+    --hash=sha256:23aceaa007d6172b02c277f0cd359c79492bbb14f7072b4ede9fbcaf20648130 \
+    --hash=sha256:23d27a3e0307ec2244cc51e7287b919aa68d097504ebe19df4e76a98a3eea5bd \
+    --hash=sha256:24870b09b224f7ae3c39ed07d10e819d06f8720bc551847b1d623832b5b0e28d \
+    --hash=sha256:251bf95b67017e27b13d82f5b326234ca62d70f9cf4c2b9032de2358a3b12c7b \
+    --hash=sha256:25b9b82bb22e6e2b3cd07b39c68b7b862001226cb3dff7130d1cb914121b39ed \
+    --hash=sha256:28ce87c5ab450a9dd970b52e5aca5fe63ed432d18a2eaddd1979a00a1ba24ace \
+    --hash=sha256:300557495eb45ebb8aec96c2da9c4be642fbf7cd937278b4013ba894ea8eb0eb \
+    --hash=sha256:30f2aa603c41533cc25c05acd0da21636e84a315768feb631c937177db558931 \
+    --hash=sha256:331b624368d4f1d069149002f25f44bc61c8919ce8ddb3c45bdad8f6e2d89510 \
+    --hash=sha256:37d6d0a00072fd2948eb22bce7e1475f34569d90c87c59f7a2ec59541b77f7a6 \
+    --hash=sha256:37dc8f7bbb66efe481bb60defacef820c950c24713fb44962ed6aa2a50966de1 \
+    --hash=sha256:3b8182a766685eaa002637e28b4ec8d6b18819a0c71f579bf0dbaa5830297cce \
+    --hash=sha256:3edce1d53195db527e0191f84b71d02022de0540bf43a16ed734ed7537b07385 \
+    --hash=sha256:446c34dcc4324b084a53b705127dc15717b22c5e140ae0a3c38349d4efec071e \
+    --hash=sha256:4998562bf62a445225f22e07c896bb04b35b1b1f2eb6d760584c9c51d7a5f78c \
+    --hash=sha256:4b0a7fe987b14c31ebda6083f74f22b561fd3739bc0ac51e019622e3d72668c7 \
+    --hash=sha256:4e8c2a84d977f50b9daed6eeaf3baef67d00d5d74d932288f02cb94518ee3ace \
+    --hash=sha256:4f883547d4b7f0495ebe7056b0cc2aea76094e7a4abc8e933540f3271df27d9c \
+    --hash=sha256:514435a37670e3e5e08f3945b68718b6ed329bb84367777e16f9f4dfe1e61a0f \
+    --hash=sha256:53aa02d20d10c3d814d536aa4e5ac9b84ca0ff5a88377963b085ad6822f93e64 \
+    --hash=sha256:5594fc43d548a7ed94949d139aa1341b270f1863f11cfd37f5a6c8b778a6b67f \
+    --hash=sha256:571b9fcb07b97ef3a492028fb3d2dc0993ca23a06138b0315286566d29ef718a \
+    --hash=sha256:57b3d78c95ba9059768b10e28b813002261d3f3dfc55cc48b0c988f625175827 \
+    --hash=sha256:5afb51d599ea772b8365ae807ae557f18bccfe46ab261fd1c2a9ed700fc6eb17 \
+    --hash=sha256:6b02afb9b97f65fbca5f31db6a2a3ba21aa93030225f150fa3f249717e938fb4 \
+    --hash=sha256:6c0016e7b354317c4e9e525b937ac8596c38d2d232b419529b9cd7a1cd46e39a \
+    --hash=sha256:71d6097b330eea8fd15097780c8e89cb1a8ce7838669f48c5bacd6f663dd4701 \
+    --hash=sha256:756c768d0c9c2955feb7a56c37ea24aea2e369f8d36a88da270b6a9f19e62b5e \
+    --hash=sha256:78cb2c6865a35ab8ff8b75fd122f6033b92a62c82801110e48ddd6c936a45d91 \
+    --hash=sha256:7a743ff716f746fc19a9557f60dab1600d4613255f8a7aeb3cdde4db7eb15a66 \
+    --hash=sha256:85f998ea1848bc6757289e739cfbdda3a04adfd58b02fc018ce54d754a5ce468 \
+    --hash=sha256:8728f216dcdb6e6d555cf971cb34076139ad74b31fc2c14da4fafc741c5f6217 \
+    --hash=sha256:877c3f311ff35410f690861c4409e7ccbf0cd2f878e50628a28e5a0bb689e658 \
+    --hash=sha256:8cd2f7bdda092d99c9fc2fb7391354f306d01443d22785d0cbfafa2e2c8bb418 \
+    --hash=sha256:8e95e1385e4998ae9694eeaa4730ba5457ff61185b3a55e2e7bea0880aef452a \
+    --hash=sha256:962864dc93511324d51ddbb5b9f8731bf71675b93ca612a07441896f4688fb8c \
+    --hash=sha256:9cf95fe4d0f84c82d282745d9bb08ad9f926efa00be4697e767b814ce40d4330 \
+    --hash=sha256:9e881fca225083806662a5c43d627d215f258ff43c890f831966c7d7ba9c7402 \
+    --hash=sha256:a2b55dd6b2a4c4b7d87ffa56bdb33fdc5fdb9a462173861a7bc097f17d91cb09 \
+    --hash=sha256:a45650e8ce7fafffd731db8550230db6b0d306d181a90b67d3e6bca2f1990930 \
+    --hash=sha256:a876864214e136f0eb367788dbd7df045f4806801518e2cfe9e13229cfe06d8f \
+    --hash=sha256:ae26d61dfa7a47befdc7572b521024e8745f3d809bd95ca9505a7bba9ef849ec \
+    --hash=sha256:af8d94b0db561cf68b88a267c5c44b49e134f525d0dc2cb7ed413a66bc23559a \
+    --hash=sha256:b343699e8308bdc51978310e1c959c584e7869cc8c40780058c87da7781a1e94 \
+    --hash=sha256:b3c777e849237620b022f7f297dd67705f9f5cf1685f09f02e46f93e92725468 \
+    --hash=sha256:b629de27fda84b42cde7edef0d85f13b958b47f6e9bbcbba9b673c562a89bd8b \
+    --hash=sha256:ba09209fbe443b4acccebe845d8a138b89a8f4fbaeedd44953490b5315d5e965 \
+    --hash=sha256:ba54cfebe86920a559a7c4d6b9050791c20513650a1952ebe3368c7dc70306f8 \
+    --hash=sha256:bcb46e2f9feff8d06323983bd83ed00c201fdcab3d74973e7072a889b3979fcd \
+    --hash=sha256:bcc33feacfaefce60c12fd500a277533bdc02b10a19f7f6d348763d8140bbba7 \
+    --hash=sha256:bf16ba1b4d0b6b7c8e534936632270cf70eb00dbe09005bc345b2677b726855c \
+    --hash=sha256:cf1845d02ad822a369a49f2bb9345b1614744267682e7a03527dc3bf6eea1777 \
+    --hash=sha256:d69141514cc30b774ceea5e3ed3a6635c8d8a96edf664689b890f4089111fb35 \
+    --hash=sha256:d9c7f76c0673154f044e9d78c8655fb4213f6ca31a836df48b40fe5d187717b9 \
+    --hash=sha256:dbce0b29841537a2fa4a214c2bbf14de3587c9680caa9b4e217568472490b28f \
+    --hash=sha256:dc624f6bc473dacdf7ef7eb8678d0d08edf15cd94fad6ae5c7d6cc67a4e4902f \
+    --hash=sha256:e158cb00350dc278f3b91551101aa7d12415a66ebf2c91d8d5ac14e56ddd3ad0 \
+    --hash=sha256:e491916b378fba47242221bb9ead245211b70d504f495d105d17b14a24b4907c \
+    --hash=sha256:e795b7eb908249c4e43c7c99fac7c2c75dab0c43566e37db472a355f63693d71 \
+    --hash=sha256:e7e480451b9fa137494bccd3a7d69adbe8ac65a87d97be61e11f1b1050a5bac3 \
+    --hash=sha256:e91206ee562682b51b98ef4b26a6ef48fd84e15fd4c4bc5ec768eb641d206838 \
+    --hash=sha256:e9871b1ffbfa9656b60aeee92ed5136a5742696006fa322b29ea3d8da0ecc9cf \
+    --hash=sha256:e9aeb04d6aef139de265b29683e119b638208f88cf73cdd1658aa07221165321 \
+    --hash=sha256:ebaea975e03d3141d9d3a507df75c9b3ec90fa9d2ffd07567b3a978d9d790b26 \
+    --hash=sha256:f0606c8bf2cdefea14a43530f7657cbbb7ecf1c4222512492ef4a4434a9501ec \
+    --hash=sha256:f13c32a3abd6079a66d9526e18dad9b6d280384d49d7c54040cd57b6424041d9 \
+    --hash=sha256:f7401aebd7f581d7f83a439d87d474999317ee099218e5ad25d125290990ba65 \
+    --hash=sha256:fa4ecea169a355be7a3ade2c783e2ed12f0e40d2c5621cda8b3297faf7fbb9f5 \
+    --hash=sha256:fbd139c8447d25dd750ab79ee274cc5e1fe80fc56340ab10b18a195e1b6eca3e \
+    --hash=sha256:fdafc9cce40277e0f7a0feabce0ee50dd2fa1800f3b38015e51296b5e814048d \
+    --hash=sha256:fe3cca2e4e8a592be0f269a1ca4835c25199d9f3ce815c8491048f785b0a0198 \
+    --hash=sha256:ffd0c5368496f41b0944be820fcb7a838aa6e623d250b01acf2643939c3f99d7
+    # via -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+pluggy==1.7.0 \
+    --hash=sha256:7dd7b0d8832ba3cb632c306926ded123429211b83641b35dc5c41ad2d34f9bec \
+    --hash=sha256:d1eaa46ebb595891b860ab086b4d09c8588af65ebd4361b8e8f4bb8920b90ba8
+    # via pytest
+pyarrow==25.0.1 \
+    --hash=sha256:0b1edbb2f385a6a65e9711b62ba86ac54a7816a3f8d17bb3e8a5929d65fb2485 \
+    --hash=sha256:0b726ad7e7b669be982b0c71c07fe4b037d654354130da79a7902a669e93a66b \
+    --hash=sha256:0befcf816e45a1af33ac775a9970b749e4868a230c7372f0ae5e932bee27039f \
+    --hash=sha256:0fe7c8b6c03969b49c8c66182e4a18e3819ab92d07cfab5d8370c531b9369ef0 \
+    --hash=sha256:119297a6dc197e45d9c6d4415f7814a67ffa36c180d26f68c154c58067ae782d \
+    --hash=sha256:169d3429d5be7c752125890620f75a60776d38b0035eddae939651640822332e \
+    --hash=sha256:25f8720bf6387d5dc2ebd2622112de630760419e4b66134405dd24110d15f37e \
+    --hash=sha256:31e49a7888fcdf3a835da33ae777f6bb9a866334e5a789282fc26dcf426f7f15 \
+    --hash=sha256:35935cd5de130aa5cf4dea052a63e6bf2e17006c35c3a468194242b9b2bf5956 \
+    --hash=sha256:38a9a4b4b9613380e200641891495a56c3d5a98a092db4a870af9975e220471d \
+    --hash=sha256:3f89685964f46e4216103c75483aac0c0692a5f72212d7ca835adba5ede56ce3 \
+    --hash=sha256:4288f27577352d608ca08553b0865e4a9b3aa14820c5d95b53337218d609835b \
+    --hash=sha256:4340f0ba6c1d2e13f21658de1d7c662ca2545018568d0030a1e9afca159d87e3 \
+    --hash=sha256:44a9120ce5bd81936b8ab9a88076e3fd47c2c6838e0e43630fed83626aca81d9 \
+    --hash=sha256:4facd65742a024a4a366328a1d2292062d72d6e023c1b7dda8d4c37544933a25 \
+    --hash=sha256:51093dd9e10325fbdb3c10a2ae7c4806e5c822d94e74ae4938b26524a3323fee \
+    --hash=sha256:514ddb60285631af068875550c90eddc181db3e8e63a032b1559be189e82f056 \
+    --hash=sha256:5389cdf79447ed1515c9e31620e6e1e2302249564d603f2ad727d4f6d313e4c3 \
+    --hash=sha256:59a2de54c0cbd954da861eee4d1d330f8e909c45b53455baef696380f2c55033 \
+    --hash=sha256:60e89d8f13861a1f7f8d950fa54aebb8023b30734d0ac51ffa80beabe2df4bba \
+    --hash=sha256:6109c94d8b9f3b17a041daca16cacb2f651ad8f1ef70a4232c2c0f37a23da2a8 \
+    --hash=sha256:62cd0d785b8aa6675ee355f9fc02252a340f4441257c42674937826fd7594325 \
+    --hash=sha256:6943e2fe7954d29d84de45d29d34c8dc36ce96570e67d89aa9976e650a4a9138 \
+    --hash=sha256:6a1fdfc6659b6b19022f2e50627fb5cf7156a66c46bf4299379955cbe742382a \
+    --hash=sha256:880523be3d29efcf83d3998835d206118ccf35e3871dbd2fb60408cf6b007a80 \
+    --hash=sha256:8858d7bfc22e3f51529aeaa4077225029724623e4595dc9eff8c793935c34140 \
+    --hash=sha256:9150a83248bfed9813ea3c3af74c3856c1984d444aa28e58bf7733b9750ddf6a \
+    --hash=sha256:9171748cdf796972d85a4b60157c279913e242992e350c90c7450182a9838b2a \
+    --hash=sha256:a4d6d5e9a3d1879a97c08ded0c797579b7965eafd0f0c26c30b45ccc06db939b \
+    --hash=sha256:a4dd8bf99a8fac133efc0ed6a92f5fddbe2adba0d0f6dd720e39ba9855cea85c \
+    --hash=sha256:aa0559502e1cd6254d6814614085dd9c5a3dd0419362978a936a3f68a9e5c3df \
+    --hash=sha256:b7a296aac7a71fa0886c08e155ddb6c636a50013f801f6178daafa0f9e726188 \
+    --hash=sha256:bddd0c4f7630c2a3ddf6347c1bdaa79d97bcf6bd445f9e60c816b7d77c85a5ae \
+    --hash=sha256:bf0b672390cdcb640d7288f96b826d71ff4e9abb254a86c89890baf51a29cee6 \
+    --hash=sha256:c7c534ec03c358a76ea3e505e74c1b6aef290af90c444dfd092dbfe23e755b85 \
+    --hash=sha256:cab40b1edfef0262e0e5251aa2c58d75630f24d06dd7794480243acc001a1d7d \
+    --hash=sha256:cc4aa407fde9fc660be3939e49ea31f50f3e9fec17c0ec63159f7711edd3efc9 \
+    --hash=sha256:d51592cb7561e87877c506113e7adbf1342ab579e6c21f0ef44b8ba41cb74c80 \
+    --hash=sha256:dda9470024204d7bbf2042b47c6e8a0e47a3eeb8e34405882dfaea6577e0c153 \
+    --hash=sha256:df961f2e7ae9cf496459259d798652c70625f6c080650d6952f8c04053c58ee9 \
+    --hash=sha256:eb6203482ff3746a5632303a7279ae0b5a304c46985b49ed1378cb350ea6728d \
+    --hash=sha256:f3831aaa25c67a99f99dc8b05873cb9d64560390372e2aa197ce9dd4a3f06a44 \
+    --hash=sha256:f729cfdbd36fd99d543b67a914d2de044c84ebe45be8b34902b299b608c15c8f
+    # via -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+pygments==2.21.0 \
+    --hash=sha256:2363c69b61c4a97c838da3b130dcd6468f4848992b21a82f2a63ec34377137d9 \
+    --hash=sha256:610ca751c9bc2492b38eb9a38a7fbc93edbbb2d7182edaf34e66ae493dee5c8c
+    # via pytest
+pytest==9.1.1 \
+    --hash=sha256:1088fbde8f2b49d95a549a195707afa7a76a3ce9bcadc26b6d71f0ffda5fe313 \
+    --hash=sha256:37a86b45efb9a47a61a36449063e8e18d0cab3161329fc099eb21783169c4f0c
+    # via -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+python-dateutil==4.0.0 \
+    --hash=sha256:a5e246f79eb80475e691bf93862cead98decd8a498cf53ba0516fcdc0f50dc1c \
+    --hash=sha256:d7295ff14fdc0f92f662100bfb6c551d160ca090f2939f1ec5b9794265c03f06
+    # via pandas
+pyyaml==6.0.3 \
+    --hash=sha256:00c4bdeba853cc34e7dd471f16b4114f4162dc03e6b7afcc2128711f0eca823c \
+    --hash=sha256:0150219816b6a1fa26fb4699fb7daa9caf09eb1999f3b70fb6e786805e80375a \
+    --hash=sha256:02893d100e99e03eda1c8fd5c441d8c60103fd175728e23e431db1b589cf5ab3 \
+    --hash=sha256:02ea2dfa234451bbb8772601d7b8e426c2bfa197136796224e50e35a78777956 \
+    --hash=sha256:0f29edc409a6392443abf94b9cf89ce99889a1dd5376d94316ae5145dfedd5d6 \
+    --hash=sha256:10892704fc220243f5305762e276552a0395f7beb4dbf9b14ec8fd43b57f126c \
+    --hash=sha256:16249ee61e95f858e83976573de0f5b2893b3677ba71c9dd36b9cf8be9ac6d65 \
+    --hash=sha256:1d37d57ad971609cf3c53ba6a7e365e40660e3be0e5175fa9f2365a379d6095a \
+    --hash=sha256:1ebe39cb5fc479422b83de611d14e2c0d3bb2a18bbcb01f229ab3cfbd8fee7a0 \
+    --hash=sha256:214ed4befebe12df36bcc8bc2b64b396ca31be9304b8f59e25c11cf94a4c033b \
+    --hash=sha256:2283a07e2c21a2aa78d9c4442724ec1eb15f5e42a723b99cb3d822d48f5f7ad1 \
+    --hash=sha256:22ba7cfcad58ef3ecddc7ed1db3409af68d023b7f940da23c6c2a1890976eda6 \
+    --hash=sha256:27c0abcb4a5dac13684a37f76e701e054692a9b2d3064b70f5e4eb54810553d7 \
+    --hash=sha256:28c8d926f98f432f88adc23edf2e6d4921ac26fb084b028c733d01868d19007e \
+    --hash=sha256:2e71d11abed7344e42a8849600193d15b6def118602c4c176f748e4583246007 \
+    --hash=sha256:34d5fcd24b8445fadc33f9cf348c1047101756fd760b4dacb5c3e99755703310 \
+    --hash=sha256:37503bfbfc9d2c40b344d06b2199cf0e96e97957ab1c1b546fd4f87e53e5d3e4 \
+    --hash=sha256:3c5677e12444c15717b902a5798264fa7909e41153cdf9ef7ad571b704a63dd9 \
+    --hash=sha256:3ff07ec89bae51176c0549bc4c63aa6202991da2d9a6129d7aef7f1407d3f295 \
+    --hash=sha256:41715c910c881bc081f1e8872880d3c650acf13dfa8214bad49ed4cede7c34ea \
+    --hash=sha256:418cf3f2111bc80e0933b2cd8cd04f286338bb88bdc7bc8e6dd775ebde60b5e0 \
+    --hash=sha256:44edc647873928551a01e7a563d7452ccdebee747728c1080d881d68af7b997e \
+    --hash=sha256:4a2e8cebe2ff6ab7d1050ecd59c25d4c8bd7e6f400f5f82b96557ac0abafd0ac \
+    --hash=sha256:4ad1906908f2f5ae4e5a8ddfce73c320c2a1429ec52eafd27138b7f1cbe341c9 \
+    --hash=sha256:501a031947e3a9025ed4405a168e6ef5ae3126c59f90ce0cd6f2bfc477be31b7 \
+    --hash=sha256:5190d403f121660ce8d1d2c1bb2ef1bd05b5f68533fc5c2ea899bd15f4399b35 \
+    --hash=sha256:5498cd1645aa724a7c71c8f378eb29ebe23da2fc0d7a08071d89469bf1d2defb \
+    --hash=sha256:5cf4e27da7e3fbed4d6c3d8e797387aaad68102272f8f9752883bc32d61cb87b \
+    --hash=sha256:5e0b74767e5f8c593e8c9b5912019159ed0533c70051e9cce3e8b6aa699fcd69 \
+    --hash=sha256:5ed875a24292240029e4483f9d4a4b8a1ae08843b9c54f43fcc11e404532a8a5 \
+    --hash=sha256:5fcd34e47f6e0b794d17de1b4ff496c00986e1c83f7ab2fb8fcfe9616ff7477b \
+    --hash=sha256:5fdec68f91a0c6739b380c83b951e2c72ac0197ace422360e6d5a959d8d97b2c \
+    --hash=sha256:6344df0d5755a2c9a276d4473ae6b90647e216ab4757f8426893b5dd2ac3f369 \
+    --hash=sha256:64386e5e707d03a7e172c0701abfb7e10f0fb753ee1d773128192742712a98fd \
+    --hash=sha256:652cb6edd41e718550aad172851962662ff2681490a8a711af6a4d288dd96824 \
+    --hash=sha256:66291b10affd76d76f54fad28e22e51719ef9ba22b29e1d7d03d6777a9174198 \
+    --hash=sha256:66e1674c3ef6f541c35191caae2d429b967b99e02040f5ba928632d9a7f0f065 \
+    --hash=sha256:6adc77889b628398debc7b65c073bcb99c4a0237b248cacaf3fe8a557563ef6c \
+    --hash=sha256:79005a0d97d5ddabfeeea4cf676af11e647e41d81c9a7722a193022accdb6b7c \
+    --hash=sha256:7c6610def4f163542a622a73fb39f534f8c101d690126992300bf3207eab9764 \
+    --hash=sha256:7f047e29dcae44602496db43be01ad42fc6f1cc0d8cd6c83d342306c32270196 \
+    --hash=sha256:8098f252adfa6c80ab48096053f512f2321f0b998f98150cea9bd23d83e1467b \
+    --hash=sha256:850774a7879607d3a6f50d36d04f00ee69e7fc816450e5f7e58d7f17f1ae5c00 \
+    --hash=sha256:8d1fab6bb153a416f9aeb4b8763bc0f22a5586065f86f7664fc23339fc1c1fac \
+    --hash=sha256:8da9669d359f02c0b91ccc01cac4a67f16afec0dac22c2ad09f46bee0697eba8 \
+    --hash=sha256:8dc52c23056b9ddd46818a57b78404882310fb473d63f17b07d5c40421e47f8e \
+    --hash=sha256:9149cad251584d5fb4981be1ecde53a1ca46c891a79788c0df828d2f166bda28 \
+    --hash=sha256:93dda82c9c22deb0a405ea4dc5f2d0cda384168e466364dec6255b293923b2f3 \
+    --hash=sha256:96b533f0e99f6579b3d4d4995707cf36df9100d67e0c8303a0c55b27b5f99bc5 \
+    --hash=sha256:9c57bb8c96f6d1808c030b1687b9b5fb476abaa47f0db9c0101f5e9f394e97f4 \
+    --hash=sha256:9c7708761fccb9397fe64bbc0395abcae8c4bf7b0eac081e12b809bf47700d0b \
+    --hash=sha256:9f3bfb4965eb874431221a3ff3fdcddc7e74e3b07799e0e84ca4a0f867d449bf \
+    --hash=sha256:a33284e20b78bd4a18c8c2282d549d10bc8408a2a7ff57653c0cf0b9be0afce5 \
+    --hash=sha256:a80cb027f6b349846a3bf6d73b5e95e782175e52f22108cfa17876aaeff93702 \
+    --hash=sha256:b30236e45cf30d2b8e7b3e85881719e98507abed1011bf463a8fa23e9c3e98a8 \
+    --hash=sha256:b3bc83488de33889877a0f2543ade9f70c67d66d9ebb4ac959502e12de895788 \
+    --hash=sha256:b865addae83924361678b652338317d1bd7e79b1f4596f96b96c77a5a34b34da \
+    --hash=sha256:b8bb0864c5a28024fac8a632c443c87c5aa6f215c0b126c449ae1a150412f31d \
+    --hash=sha256:ba1cc08a7ccde2d2ec775841541641e4548226580ab850948cbfda66a1befcdc \
+    --hash=sha256:bdb2c67c6c1390b63c6ff89f210c8fd09d9a1217a465701eac7316313c915e4c \
+    --hash=sha256:c1ff362665ae507275af2853520967820d9124984e0f7466736aea23d8611fba \
+    --hash=sha256:c2514fceb77bc5e7a2f7adfaa1feb2fb311607c9cb518dbc378688ec73d8292f \
+    --hash=sha256:c3355370a2c156cffb25e876646f149d5d68f5e0a3ce86a5084dd0b64a994917 \
+    --hash=sha256:c458b6d084f9b935061bc36216e8a69a7e293a2f1e68bf956dcd9e6cbcd143f5 \
+    --hash=sha256:d0eae10f8159e8fdad514efdc92d74fd8d682c933a6dd088030f3834bc8e6b26 \
+    --hash=sha256:d76623373421df22fb4cf8817020cbb7ef15c725b9d5e45f17e189bfc384190f \
+    --hash=sha256:ebc55a14a21cb14062aa4162f906cd962b28e2e9ea38f9b4391244cd8de4ae0b \
+    --hash=sha256:eda16858a3cab07b80edaf74336ece1f986ba330fdb8ee0d6c0d68fe82bc96be \
+    --hash=sha256:ee2922902c45ae8ccada2c5b501ab86c36525b883eff4255313a253a3160861c \
+    --hash=sha256:efd7b85f94a6f21e4932043973a7ba2613b059c4a000551892ac9f1d11f5baf3 \
+    --hash=sha256:f7057c9a337546edc7973c0d3ba84ddcdf0daa14533c2065749c9075001090e6 \
+    --hash=sha256:fa160448684b4e94d80416c0fa4aac48967a969efe22931448d853ada8baf926 \
+    --hash=sha256:fc09d0aa354569bc501d4e787133afc08552722d3ab34836a80547331bb5d4a0
+    # via huggingface-hub
+requests==2.34.2 \
+    --hash=sha256:2a0d60c172f83ac6ab31e4554906c0f3b3588d37b5cb939b1c061f4907e278e0 \
+    --hash=sha256:f288924cae4e29463698d6d60bc6a4da69c89185ad1e0bcc4104f584e960b9ed
+    # via -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+tokenizers==0.22.2 \
+    --hash=sha256:143b999bdc46d10febb15cbffb4207ddd1f410e2c755857b5a0797961bbdc113 \
+    --hash=sha256:1a62ba2c5faa2dd175aaeed7b15abf18d20266189fb3406c5d0550dd34dd5f37 \
+    --hash=sha256:1c774b1276f71e1ef716e5486f21e76333464f47bece56bbd554485982a9e03e \
+    --hash=sha256:1e418a55456beedca4621dbab65a318981467a2b188e982a23e117f115ce5001 \
+    --hash=sha256:1e50f8554d504f617d9e9d6e4c2c2884a12b388a97c5c77f0bc6cf4cd032feee \
+    --hash=sha256:2249487018adec45d6e3554c71d46eb39fa8ea67156c640f7513eb26f318cec7 \
+    --hash=sha256:25b85325d0815e86e0bac263506dd114578953b7b53d7de09a6485e4a160a7dd \
+    --hash=sha256:29c30b83d8dcd061078b05ae0cb94d3c710555fbb44861139f9f83dcca3dc3e4 \
+    --hash=sha256:319f659ee992222f04e58f84cbf407cfa66a65fe3a8de44e8ad2bc53e7d99012 \
+    --hash=sha256:369cc9fc8cc10cb24143873a0d95438bb8ee257bb80c71989e3ee290e8d72c67 \
+    --hash=sha256:37ae80a28c1d3265bb1f22464c856bd23c02a05bb211e56d0c5301a435be6c1a \
+    --hash=sha256:38337540fbbddff8e999d59970f3c6f35a82de10053206a7562f1ea02d046fa5 \
+    --hash=sha256:473b83b915e547aa366d1eee11806deaf419e17be16310ac0a14077f1e28f917 \
+    --hash=sha256:544dd704ae7238755d790de45ba8da072e9af3eea688f698b137915ae959281c \
+    --hash=sha256:64d94e84f6660764e64e7e0b22baa72f6cd942279fdbb21d46abd70d179f0195 \
+    --hash=sha256:753d47ebd4542742ef9261d9da92cd545b2cacbb48349a1225466745bb866ec4 \
+    --hash=sha256:791135ee325f2336f498590eb2f11dc5c295232f288e75c99a36c5dbce63088a \
+    --hash=sha256:9ce725d22864a1e965217204946f830c37876eee3b2ba6fc6255e8e903d5fcbc \
+    --hash=sha256:a6bf3f88c554a2b653af81f3204491c818ae2ac6fbc09e76ef4773351292bc92 \
+    --hash=sha256:bfb88f22a209ff7b40a576d5324bf8286b519d7358663db21d6246fb17eea2d5 \
+    --hash=sha256:c9ea31edff2968b44a88f97d784c2f16dc0729b8b143ed004699ebca91f05c48 \
+    --hash=sha256:df6c4265b289083bf710dff49bc51ef252f9d5be33a45ee2bed151114a56207b \
+    --hash=sha256:e10bf9113d209be7cd046d40fbabbaf3278ff6d18eb4da4c500443185dc1896c \
+    --hash=sha256:f01a9c019878532f98927d2bacb79bbb404b43d3437455522a00a30718cdedb5
+    # via -r zhangyuxuan/training_env_repro_20261105/robot/data_cpu/requirements.core.in
+tqdm==4.70.1 \
+    --hash=sha256:c293e525e6fef9c20e8728fd4612df02a0aa31bb5fe91ecd93e123b1b7bffa73 \
+    --hash=sha256:cefd0eca11b2a37a3aee776544d4f4ae913f02688135b5556b8788dfa474afc4
+    # via huggingface-hub
+typing-extensions==4.16.0 \
+    --hash=sha256:481caa481374e813c1b176ada14e97f1f67a4539ce9cfeb3f350d78d6370c2e8 \
+    --hash=sha256:dc983d19a509c94dba722ee6abd33940f7c05a89e243c47e907eb4db6f1a43e5
+    # via
+    #   anyio
+    #   huggingface-hub
+tzdata==2026.6 \
+    --hash=sha256:0a10eec6e13ef16b627fabed9a5f2ba61a62b1c67f327305f4e7e412a70b4aec \
+    --hash=sha256:85d2cd9c09f81f0bb29db27759fc83af4f1e620efb814f8c5bf8b4cc4feece6d
+    # via python-dateutil
+urllib3==2.8.0 \
+    --hash=sha256:0cf3cae568d36aa9576b28dfb35f11328f1cb974ca7647d9475ebb86c75ac6e3 \
+    --hash=sha256:63bf2ead4c879426ebf22ef2a781eeb4aa3b4ae798a0435506f8687fd5bb9b63
+    # via requests

robot/data_cpu/requirements.observed-full.lock.txt

@@ -0,0 +1,32 @@
+anyio==4.15.1
+certifi==2026.7.22
+charset-normalizer==3.5.1
+click==8.5.0
+filelock==4.0.5
+fsspec==2026.6.0
+h11==0.16.0
+h5py==3.16.0
+hf-xet==1.6.0
+httpcore==1.0.9
+httpx==0.28.1
+huggingface-hub==1.33.0
+idna==3.20
+iniconfig==2.3.0
+numpy==2.3.5
+opencv-python-headless==5.0.0.93
+packaging==25.0
+pandas==3.0.6
+pillow==12.3.0
+pip==25.0.1
+pluggy==1.6.0
+pyarrow==25.0.1
+pygments==2.21.0
+pytest==9.1.1
+python-dateutil==2.9.0.post0
+pyyaml==6.0.3
+requests==2.34.2
+six==1.17.0
+tokenizers==0.22.2
+tqdm==4.70.1
+typing-extensions==4.16.0
+urllib3==2.8.0

robot/legacy_sft_grpo/environment.snapshot.json

@@ -0,0 +1,1733 @@
+{
+  "profile": "robot_legacy_sft_grpo",
+  "capture_scope": "observed current retained runtime; not proof of original launch-time environment",
+  "python": "3.12.13",
+  "implementation": "CPython",
+  "SOABI": "cpython-312-x86_64-linux-gnu",
+  "machine": "x86_64",
+  "platform": "Linux",
+  "libc": [
+    "glibc",
+    "2.35"
+  ],
+  "packages": {
+    "absl-py": "2.5.0",
+    "accelerate": "1.15.0",
+    "agent-detector": "2.0.0",
+    "aiohappyeyeballs": "2.7.1",
+    "aiohttp": "3.14.3",
+    "aiohttp-cors": "0.8.1",
+    "aiosignal": "1.4.0",
+    "annotated-doc": "0.0.5",
+    "annotated-types": "0.8.0",
+    "anthropic": "1.8.0",
+    "antlr4-python3-runtime": "4.9.3",
+    "anyio": "4.15.1",
+    "apache-tvm-ffi": "0.1.9",
+    "astor": "0.8.1",
+    "attrs": "26.1.0",
+    "av": "18.1.0",
+    "blake3": "1.0.9",
+    "cachetools": "7.2.0",
+    "cbor2": "6.1.4",
+    "certifi": "2026.7.22",
+    "cffi": "2.1.1",
+    "charset-normalizer": "3.5.1",
+    "click": "8.5.0",
+    "cloudpickle": "3.1.2",
+    "codetiming": "1.4.0",
+    "colorful": "0.5.8",
+    "compressed-tensors": "0.17.0",
+    "cryptography": "50.0.1",
+    "cuda-bindings": "13.4.3",
+    "cuda-core": "1.2.1",
+    "cuda-pathfinder": "1.8.2",
+    "cuda-python": "13.4.1",
+    "cuda-tile": "1.3.0",
+    "cuda-toolkit": "13.0.2",
+    "datasets": "5.0.1",
+    "depyf": "0.20.0",
+    "detect-installer": "0.2.1",
+    "dill": "0.4.1",
+    "diskcache": "5.6.3",
+    "distlib": "0.4.3",
+    "dnspython": "2.8.0",
+    "docstring-parser": "0.18.0",
+    "einops": "0.8.2",
+    "email-validator": "2.3.0",
+    "fastapi": "0.136.3",
+    "fastapi-cli": "0.0.32",
+    "fastapi-cloud-cli": "0.26.0",
+    "fastar": "0.12.0",
+    "fastsafetensors": "0.4.0",
+    "filelock": "4.0.5",
+    "fla-core": "0.5.2",
+    "flash-attn": "2.8.3",
+    "flash-linear-attention": "0.5.2",
+    "flashinfer-cubin": "0.6.12",
+    "flashinfer-python": "0.6.12",
+    "frozenlist": "1.8.0",
+    "fsspec": "2026.6.0",
+    "google-api-core": "2.39.0",
+    "google-auth": "2.58.1",
+    "googleapis-common-protos": "1.75.4",
+    "grpcio": "1.84.0",
+    "h11": "0.16.0",
+    "h5py": "3.16.0",
+    "hf-xet": "1.6.0",
+    "httpcore": "1.0.9",
+    "httpcore2": "2.13.1",
+    "httptools": "0.8.0",
+    "httpx": "0.28.1",
+    "httpx2": "2.13.1",
+    "huggingface-hub": "1.33.0",
+    "humming-kernels": "0.1.6",
+    "hydra-core": "1.3.7",
+    "idna": "3.20",
+    "ijson": "3.5.1",
+    "importlib-metadata": "9.0.1",
+    "iniconfig": "2.3.0",
+    "interegular": "0.3.3",
+    "jinja2": "3.1.6",
+    "jiter": "0.17.0",
+    "jmespath": "1.1.0",
+    "jsonschema": "4.26.0",
+    "jsonschema-specifications": "2025.9.1",
+    "lark": "1.2.2",
+    "llguidance": "1.7.6",
+    "llvmlite": "0.47.0",
+    "lm-format-enforcer": "0.11.3",
+    "loguru": "0.7.3",
+    "markdown": "3.11",
+    "markdown-it-py": "4.2.0",
+    "markupsafe": "3.0.3",
+    "mathruler": "0.1.0",
+    "mbridge": "0.15.1",
+    "mcp": "2.2.0",
+    "mcp-types": "2.2.0",
+    "mdurl": "0.1.2",
+    "mistral-common": "1.12.0",
+    "ml-dtypes": "0.6.0",
+    "model-hosting-container-standards": "0.1.16",
+    "mpmath": "1.3.0",
+    "msgpack": "1.2.2",
+    "msgspec": "0.21.1",
+    "multidict": "6.9.1",
+    "multiprocess": "0.70.19",
+    "networkx": "3.7",
+    "ninja": "1.13.2",
+    "numba": "0.65.0",
+    "numpy": "2.3.5",
+    "nvidia-cublas": "13.1.0.3",
+    "nvidia-cuda-cccl": "13.3.4.3.1",
+    "nvidia-cuda-crt": "13.4.92",
+    "nvidia-cuda-cupti": "13.0.85",
+    "nvidia-cuda-nvcc": "13.2.86",
+    "nvidia-cuda-nvrtc": "13.0.88",
+    "nvidia-cuda-runtime": "13.0.96",
+    "nvidia-cuda-tileiras": "13.2.86",
+    "nvidia-cudnn-cu13": "9.19.0.56",
+    "nvidia-cudnn-frontend": "1.28.0",
+    "nvidia-cufft": "12.0.0.61",
+    "nvidia-cufile": "1.15.1.6",
+    "nvidia-curand": "10.4.0.35",
+    "nvidia-cusolver": "12.0.4.66",
+    "nvidia-cusparse": "12.6.3.3",
+    "nvidia-cusparselt-cu13": "0.8.0",
+    "nvidia-cutlass-dsl": "4.5.2",
+    "nvidia-cutlass-dsl-libs-base": "4.5.2",
+    "nvidia-cutlass-dsl-libs-cu13": "4.5.2",
+    "nvidia-ml-py": "13.615.71",
+    "nvidia-nccl-cu13": "2.28.9",
+    "nvidia-nvjitlink": "13.0.88",
+    "nvidia-nvshmem-cu13": "3.4.5",
+    "nvidia-nvtx": "13.0.85",
+    "nvidia-nvvm": "13.2.86",
+    "nvtx": "0.2.16",
+    "omegaconf": "2.3.1",
+    "openai": "3.19.2",
+    "openai-harmony": "0.0.8",
+    "opencensus": "0.11.4",
+    "opencensus-context": "0.1.3",
+    "opencv-python-headless": "5.0.0.93",
+    "opentelemetry-api": "1.45.0",
+    "opentelemetry-exporter-http-transport": "0.66b0",
+    "opentelemetry-exporter-otlp": "1.45.0",
+    "opentelemetry-exporter-otlp-common": "0.66b0",
+    "opentelemetry-exporter-otlp-proto-common": "1.45.0",
+    "opentelemetry-exporter-otlp-proto-grpc": "1.45.0",
+    "opentelemetry-exporter-otlp-proto-http": "1.45.0",
+    "opentelemetry-exporter-prometheus": "0.66b0",
+    "opentelemetry-proto": "1.45.0",
+    "opentelemetry-sdk": "1.45.0",
+    "opentelemetry-semantic-conventions": "0.66b0",
+    "opentelemetry-semantic-conventions-ai": "0.5.1",
+    "orjson": "3.12.0",
+    "outlines-core": "0.2.14",
+    "packaging": "25.0",
+    "pandas": "3.0.6",
+    "partial-json-parser": "0.2.1.1.post7",
+    "peft": "0.21.0",
+    "pillow": "12.3.0",
+    "platformdirs": "4.12.1",
+    "pluggy": "1.6.0",
+    "prometheus-client": "0.26.0",
+    "prometheus-fastapi-instrumentator": "8.1.0",
+    "propcache": "0.5.4",
+    "proto-plus": "1.28.4",
+    "protobuf": "7.36.2",
+    "psutil": "7.2.2",
+    "py-cpuinfo": "9.0.0",
+    "py-spy": "0.4.2",
+    "pyarrow": "25.0.1",
+    "pyasn1": "0.6.4",
+    "pyasn1-modules": "0.4.2",
+    "pybase64": "1.5.0",
+    "pybind11": "3.1.0",
+    "pycountry": "26.2.16",
+    "pycparser": "3.0",
+    "pydantic": "2.13.5",
+    "pydantic-core": "2.46.5",
+    "pydantic-extra-types": "2.11.1",
+    "pydantic-settings": "2.15.0",
+    "pyelftools": "0.33",
+    "pygments": "2.21.0",
+    "pyjwt": "2.15.0",
+    "pylatexenc": "2.11",
+    "pytest": "9.1.1",
+    "pytest-asyncio": "1.4.0",
+    "pytest-rerunfailures": "16.7",
+    "python-dateutil": "2.9.0.post0",
+    "python-discovery": "1.6.1",
+    "python-dotenv": "1.2.3",
+    "python-json-logger": "4.2.0",
+    "python-multipart": "0.0.32",
+    "pyvers": "0.1.0",
+    "pyyaml": "6.0.3",
+    "pyzmq": "27.2.0",
+    "quack-kernels": "0.5.0",
+    "qwen-vl-utils": "0.0.14",
+    "ray": "2.58.0",
+    "referencing": "0.37.0",
+    "regex": "2026.9.10",
+    "requests": "2.34.2",
+    "rich": "15.0.0",
+    "rich-toolkit": "0.20.5",
+    "rignore": "0.8.1",
+    "rpds-py": "2026.6.3",
+    "safetensors": "0.8.0",
+    "sentencepiece": "0.2.2",
+    "sentry-sdk": "2.70.0",
+    "setproctitle": "1.3.7",
+    "setuptools": "80.10.2",
+    "shellingham": "1.5.4",
+    "six": "1.17.0",
+    "smart-open": "8.0.1",
+    "sniffio": "1.3.1",
+    "sse-starlette": "3.4.11",
+    "starlette": "1.7.0",
+    "supervisor": "4.3.0",
+    "sympy": "1.14.0",
+    "tabulate": "0.10.0",
+    "tensorboard": "2.21.0",
+    "tensorboard-data-server": "0.7.2",
+    "tensordict": "0.10.0",
+    "tiktoken": "0.14.0",
+    "tilelang": "0.1.9",
+    "tokenizers": "0.22.2",
+    "tokenspeed-mla": "0.1.2",
+    "tokenspeed-triton": "3.8.10.post20260920",
+    "torch": "2.11.0",
+    "torch-c-dlpack-ext": "0.1.5",
+    "torchaudio": "2.11.0",
+    "torchcodec": "0.16.0",
+    "torchdata": "0.11.0",
+    "torchvision": "0.26.0",
+    "tqdm": "4.70.1",
+    "transferqueue": "0.1.11",
+    "transformers": "5.9.0",
+    "triton": "3.6.0",
+    "truststore": "0.10.4",
+    "typer": "0.27.2",
+    "typing-extensions": "4.16.0",
+    "typing-inspection": "0.4.4",
+    "urllib3": "2.8.0",
+    "uvicorn": "0.54.0",
+    "uvloop": "0.22.1",
+    "verl": "0.9.1",
+    "virtualenv": "21.13.0",
+    "vllm": "0.24.0",
+    "wandb": "0.30.0",
+    "watchfiles": "1.3.0",
+    "websockets": "17.1",
+    "werkzeug": "3.1.9",
+    "wrapt": "2.5.0",
+    "xgrammar": "0.2.3",
+    "xxhash": "4.0.1",
+    "yarl": "1.25.1",
+    "z3-solver": "4.15.4.0",
+    "zipp": "4.1.0"
+  },
+  "core_package_provenance": {
+    "triton": {
+      "version": "3.6.0",
+      "requires_dist": [
+        "importlib-metadata; python_version < \"3.10\"",
+        "cmake<4.0,>=3.20; extra == \"build\"",
+        "lit; extra == \"build\"",
+        "autopep8; extra == \"tests\"",
+        "isort; extra == \"tests\"",
+        "numpy; extra == \"tests\"",
+        "pytest; extra == \"tests\"",
+        "pytest-forked; extra == \"tests\"",
+        "pytest-xdist; extra == \"tests\"",
+        "scipy>=1.7.1; extra == \"tests\"",
+        "llnl-hatchet; extra == \"tests\"",
+        "matplotlib; extra == \"tutorials\"",
+        "pandas; extra == \"tutorials\"",
+        "tabulate; extra == \"tutorials\""
+      ],
+      "installer": "uv"
+    },
+    "torchvision": {
+      "version": "0.26.0",
+      "requires_dist": [
+        "numpy",
+        "torch (==2.11.0)",
+        "pillow (!=8.3.*,>=5.3.0)",
+        "gdown (>=4.7.3) ; extra == 'gdown'",
+        "scipy ; extra == 'scipy'"
+      ],
+      "installer": "uv"
+    },
+    "torchaudio": {
+      "version": "2.11.0",
+      "requires_dist": [],
+      "installer": "uv"
+    },
+    "pillow": {
+      "version": "12.3.0",
+      "requires_dist": [
+        "furo; extra == \"docs\"",
+        "olefile; extra == \"docs\"",
+        "sphinx>=8.2; extra == \"docs\"",
+        "sphinx-autobuild; extra == \"docs\"",
+        "sphinx-copybutton; extra == \"docs\"",
+        "sphinx-inline-tabs; extra == \"docs\"",
+        "sphinxext-opengraph; extra == \"docs\"",
+        "olefile; extra == \"fpx\"",
+        "olefile; extra == \"mic\"",
+        "arro3-compute; extra == \"test-arrow\"",
+        "arro3-core; extra == \"test-arrow\"",
+        "nanoarrow; extra == \"test-arrow\"",
+        "pyarrow; extra == \"test-arrow\"",
+        "coverage>=7.4.2; extra == \"tests\"",
+        "defusedxml; extra == \"tests\"",
+        "markdown2; extra == \"tests\"",
+        "olefile; extra == \"tests\"",
+        "packaging; extra == \"tests\"",
+        "pytest; extra == \"tests\"",
+        "pytest-cov; extra == \"tests\"",
+        "pytest-timeout; extra == \"tests\"",
+        "pytest-xdist; extra == \"tests\"",
+        "setuptools; extra == \"tests\"",
+        "trove-classifiers>=2024.10.12; extra == \"tests\"",
+        "defusedxml; extra == \"xmp\""
+      ],
+      "installer": "uv"
+    },
+    "pyarrow": {
+      "version": "25.0.1",
+      "requires_dist": [],
+      "installer": "uv"
+    },
+    "peft": {
+      "version": "0.21.0",
+      "requires_dist": [
+        "numpy>=1.17",
+        "packaging>=20.0",
+        "psutil",
+        "pyyaml",
+        "torch>=1.13.0",
+        "transformers",
+        "tqdm",
+        "accelerate>=0.21.0",
+        "safetensors",
+        "huggingface_hub>=0.25.0",
+        "black; extra == \"quality\"",
+        "hf-doc-builder; extra == \"quality\"",
+        "ruff==0.16.4; extra == \"quality\"",
+        "griffe; extra == \"quality\"",
+        "black; extra == \"docs-specific\"",
+        "requests; extra == \"docs-specific\"",
+        "hf-doc-builder; extra == \"docs-specific\"",
+        "black; extra == \"dev\"",
+        "hf-doc-builder; extra == \"dev\"",
+        "ruff==0.16.4; extra == \"dev\"",
+        "griffe; extra == \"dev\"",
+        "black; extra == \"dev\"",
+        "requests; extra == \"dev\"",
+        "hf-doc-builder; extra == \"dev\"",
+        "black; extra == \"test\"",
+        "hf-doc-builder; extra == \"test\"",
+        "ruff==0.16.4; extra == \"test\"",
+        "griffe; extra == \"test\"",
+        "black; extra == \"test\"",
+        "requests; extra == \"test\"",
+        "hf-doc-builder; extra == \"test\"",
+        "pytest; extra == \"test\"",
+        "pytest-cov; extra == \"test\"",
+        "pytest-xdist; extra == \"test\"",
+        "parameterized; extra == \"test\"",
+        "datasets; extra == \"test\"",
+        "diffusers; extra == \"test\"",
+        "scipy; extra == \"test\"",
+        "scikit-learn; extra == \"test\"",
+        "protobuf; extra == \"test\"",
+        "sentencepiece; extra == \"test\"",
+        "torchvision; extra == \"test\""
+      ],
+      "installer": "uv"
+    },
+    "transformers": {
+      "version": "5.9.0",
+      "requires_dist": [
+        "huggingface-hub<2.0,>=1.5.0",
+        "numpy>=1.17",
+        "packaging>=20.0",
+        "pyyaml>=5.1",
+        "regex>=2025.10.22",
+        "tokenizers<=0.23.0,>=0.22.0",
+        "typer",
+        "safetensors>=0.4.3",
+        "tqdm>=4.27",
+        "torch>=2.4; extra == \"torch\"",
+        "accelerate>=1.1.0; extra == \"torch\"",
+        "torchvision; extra == \"vision\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"vision\"",
+        "torchaudio; extra == \"audio\"",
+        "librosa; extra == \"audio\"",
+        "pyctcdecode>=0.4.0; extra == \"audio\"",
+        "phonemizer; extra == \"audio\"",
+        "av; extra == \"video\"",
+        "timm>=1.0.23; extra == \"timm\"",
+        "datasets>=2.15.0; extra == \"quality\"",
+        "ruff==0.14.10; extra == \"quality\"",
+        "GitPython<3.1.19; extra == \"quality\"",
+        "urllib3<2.0.0; extra == \"quality\"",
+        "libcst; extra == \"quality\"",
+        "rich; extra == \"quality\"",
+        "ty==0.0.20; extra == \"quality\"",
+        "tomli; extra == \"quality\"",
+        "transformers-mlinter==0.1.1; extra == \"quality\"",
+        "hf-doc-builder; extra == \"docs\"",
+        "kernels<0.13,>=0.12.0; extra == \"kernels\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"sentencepiece\"",
+        "protobuf; extra == \"sentencepiece\"",
+        "tiktoken; extra == \"tiktoken\"",
+        "blobfile; extra == \"tiktoken\"",
+        "mistral-common[image]>=1.10.0; extra == \"mistral-common\"",
+        "jinja2>=3.1.0; extra == \"chat-template\"",
+        "jmespath>=1.0.1; extra == \"chat-template\"",
+        "scikit-learn; extra == \"sklearn\"",
+        "accelerate>=1.1.0; extra == \"accelerate\"",
+        "faiss-cpu; extra == \"retrieval\"",
+        "datasets>=2.15.0; extra == \"retrieval\"",
+        "sagemaker>=2.31.0; extra == \"sagemaker\"",
+        "deepspeed>=0.9.3; extra == \"deepspeed\"",
+        "accelerate>=1.1.0; extra == \"deepspeed\"",
+        "optuna; extra == \"optuna\"",
+        "kernels<0.13,>=0.12.0; extra == \"integrations\"",
+        "optuna; extra == \"integrations\"",
+        "codecarbon>=2.8.1; extra == \"integrations\"",
+        "ray[tune]>=2.7.0; extra == \"integrations\"",
+        "ray[tune]>=2.7.0; extra == \"ray\"",
+        "codecarbon>=2.8.1; extra == \"codecarbon\"",
+        "openai>=1.98.0; extra == \"serving\"",
+        "pydantic>=2; extra == \"serving\"",
+        "uvicorn; extra == \"serving\"",
+        "fastapi; extra == \"serving\"",
+        "starlette; extra == \"serving\"",
+        "rich; extra == \"serving\"",
+        "torch>=2.4; extra == \"serving\"",
+        "accelerate>=1.1.0; extra == \"serving\"",
+        "num2words; extra == \"num2words\"",
+        "optimum-benchmark>=0.3.0; extra == \"benchmark\"",
+        "fugashi>=1.0; extra == \"ja\"",
+        "ipadic<2.0,>=1.0.0; extra == \"ja\"",
+        "unidic_lite>=1.0.7; extra == \"ja\"",
+        "unidic>=1.0.2; extra == \"ja\"",
+        "rhoknp<1.3.1,>=1.1.0; extra == \"ja\"",
+        "sudachipy>=0.6.6; extra == \"ja\"",
+        "sudachidict_core>=20220729; extra == \"ja\"",
+        "pytest<9.0.0,>=7.2.0; extra == \"testing\"",
+        "pytest-asyncio>=1.2.0; extra == \"testing\"",
+        "pytest-random-order; extra == \"testing\"",
+        "pytest-rich; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-order; extra == \"testing\"",
+        "pytest-rerunfailures<16.0; extra == \"testing\"",
+        "pytest-timeout; extra == \"testing\"",
+        "pytest-env; extra == \"testing\"",
+        "timeout-decorator; extra == \"testing\"",
+        "parameterized>=0.9; extra == \"testing\"",
+        "psutil; extra == \"testing\"",
+        "dill<0.3.5; extra == \"testing\"",
+        "evaluate>=0.4.6; extra == \"testing\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"testing\"",
+        "nltk<=3.8.1; extra == \"testing\"",
+        "sacremoses; extra == \"testing\"",
+        "rjieba; extra == \"testing\"",
+        "beautifulsoup4; extra == \"testing\"",
+        "tensorboard; extra == \"testing\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"testing\"",
+        "filelock; extra == \"testing\"",
+        "hf-doc-builder; extra == \"testing\"",
+        "datasets>=2.15.0; extra == \"testing\"",
+        "ruff==0.14.10; extra == \"testing\"",
+        "GitPython<3.1.19; extra == \"testing\"",
+        "urllib3<2.0.0; extra == \"testing\"",
+        "libcst; extra == \"testing\"",
+        "rich; extra == \"testing\"",
+        "ty==0.0.20; extra == \"testing\"",
+        "tomli; extra == \"testing\"",
+        "transformers-mlinter==0.1.1; extra == \"testing\"",
+        "faiss-cpu; extra == \"testing\"",
+        "datasets>=2.15.0; extra == \"testing\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"testing\"",
+        "protobuf; extra == \"testing\"",
+        "openai>=1.98.0; extra == \"testing\"",
+        "pydantic>=2; extra == \"testing\"",
+        "uvicorn; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "starlette; extra == \"testing\"",
+        "rich; extra == \"testing\"",
+        "torch>=2.4; extra == \"testing\"",
+        "accelerate>=1.1.0; extra == \"testing\"",
+        "mistral-common[image]>=1.10.0; extra == \"testing\"",
+        "deepspeed>=0.9.3; extra == \"deepspeed-testing\"",
+        "accelerate>=1.1.0; extra == \"deepspeed-testing\"",
+        "pytest<9.0.0,>=7.2.0; extra == \"deepspeed-testing\"",
+        "pytest-asyncio>=1.2.0; extra == \"deepspeed-testing\"",
+        "pytest-random-order; extra == \"deepspeed-testing\"",
+        "pytest-rich; extra == \"deepspeed-testing\"",
+        "pytest-xdist; extra == \"deepspeed-testing\"",
+        "pytest-order; extra == \"deepspeed-testing\"",
+        "pytest-rerunfailures<16.0; extra == \"deepspeed-testing\"",
+        "pytest-timeout; extra == \"deepspeed-testing\"",
+        "pytest-env; extra == \"deepspeed-testing\"",
+        "timeout-decorator; extra == \"deepspeed-testing\"",
+        "parameterized>=0.9; extra == \"deepspeed-testing\"",
+        "psutil; extra == \"deepspeed-testing\"",
+        "dill<0.3.5; extra == \"deepspeed-testing\"",
+        "evaluate>=0.4.6; extra == \"deepspeed-testing\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"deepspeed-testing\"",
+        "nltk<=3.8.1; extra == \"deepspeed-testing\"",
+        "sacremoses; extra == \"deepspeed-testing\"",
+        "rjieba; extra == \"deepspeed-testing\"",
+        "beautifulsoup4; extra == \"deepspeed-testing\"",
+        "tensorboard; extra == \"deepspeed-testing\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"deepspeed-testing\"",
+        "filelock; extra == \"deepspeed-testing\"",
+        "hf-doc-builder; extra == \"deepspeed-testing\"",
+        "datasets>=2.15.0; extra == \"deepspeed-testing\"",
+        "ruff==0.14.10; extra == \"deepspeed-testing\"",
+        "GitPython<3.1.19; extra == \"deepspeed-testing\"",
+        "urllib3<2.0.0; extra == \"deepspeed-testing\"",
+        "libcst; extra == \"deepspeed-testing\"",
+        "rich; extra == \"deepspeed-testing\"",
+        "ty==0.0.20; extra == \"deepspeed-testing\"",
+        "tomli; extra == \"deepspeed-testing\"",
+        "transformers-mlinter==0.1.1; extra == \"deepspeed-testing\"",
+        "faiss-cpu; extra == \"deepspeed-testing\"",
+        "datasets>=2.15.0; extra == \"deepspeed-testing\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"deepspeed-testing\"",
+        "protobuf; extra == \"deepspeed-testing\"",
+        "openai>=1.98.0; extra == \"deepspeed-testing\"",
+        "pydantic>=2; extra == \"deepspeed-testing\"",
+        "uvicorn; extra == \"deepspeed-testing\"",
+        "fastapi; extra == \"deepspeed-testing\"",
+        "starlette; extra == \"deepspeed-testing\"",
+        "rich; extra == \"deepspeed-testing\"",
+        "torch>=2.4; extra == \"deepspeed-testing\"",
+        "accelerate>=1.1.0; extra == \"deepspeed-testing\"",
+        "mistral-common[image]>=1.10.0; extra == \"deepspeed-testing\"",
+        "optuna; extra == \"deepspeed-testing\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"deepspeed-testing\"",
+        "protobuf; extra == \"deepspeed-testing\"",
+        "torch>=2.4; extra == \"all\"",
+        "accelerate>=1.1.0; extra == \"all\"",
+        "torchvision; extra == \"all\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"all\"",
+        "torchaudio; extra == \"all\"",
+        "librosa; extra == \"all\"",
+        "pyctcdecode>=0.4.0; extra == \"all\"",
+        "phonemizer; extra == \"all\"",
+        "av; extra == \"all\"",
+        "kernels<0.13,>=0.12.0; extra == \"all\"",
+        "timm>=1.0.23; extra == \"all\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"all\"",
+        "protobuf; extra == \"all\"",
+        "tiktoken; extra == \"all\"",
+        "blobfile; extra == \"all\"",
+        "jinja2>=3.1.0; extra == \"all\"",
+        "jmespath>=1.0.1; extra == \"all\"",
+        "num2words; extra == \"all\"",
+        "mistral-common[image]>=1.10.0; extra == \"all\"",
+        "torch>=2.4; extra == \"dev\"",
+        "accelerate>=1.1.0; extra == \"dev\"",
+        "torchvision; extra == \"dev\"",
+        "Pillow<=15.0,>=10.0.1; extra == \"dev\"",
+        "torchaudio; extra == \"dev\"",
+        "librosa; extra == \"dev\"",
+        "pyctcdecode>=0.4.0; extra == \"dev\"",
+        "phonemizer; extra == \"dev\"",
+        "av; extra == \"dev\"",
+        "kernels<0.13,>=0.12.0; extra == \"dev\"",
+        "timm>=1.0.23; extra == \"dev\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev\"",
+        "protobuf; extra == \"dev\"",
+        "tiktoken; extra == \"dev\"",
+        "blobfile; extra == \"dev\"",
+        "jinja2>=3.1.0; extra == \"dev\"",
+        "jmespath>=1.0.1; extra == \"dev\"",
+        "num2words; extra == \"dev\"",
+        "mistral-common[image]>=1.10.0; extra == \"dev\"",
+        "pytest<9.0.0,>=7.2.0; extra == \"dev\"",
+        "pytest-asyncio>=1.2.0; extra == \"dev\"",
+        "pytest-random-order; extra == \"dev\"",
+        "pytest-rich; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-order; extra == \"dev\"",
+        "pytest-rerunfailures<16.0; extra == \"dev\"",
+        "pytest-timeout; extra == \"dev\"",
+        "pytest-env; extra == \"dev\"",
+        "timeout-decorator; extra == \"dev\"",
+        "parameterized>=0.9; extra == \"dev\"",
+        "psutil; extra == \"dev\"",
+        "dill<0.3.5; extra == \"dev\"",
+        "evaluate>=0.4.6; extra == \"dev\"",
+        "rouge-score!=0.0.7,!=0.0.8,!=0.1,!=0.1.1; extra == \"dev\"",
+        "nltk<=3.8.1; extra == \"dev\"",
+        "sacremoses; extra == \"dev\"",
+        "rjieba; extra == \"dev\"",
+        "beautifulsoup4; extra == \"dev\"",
+        "tensorboard; extra == \"dev\"",
+        "sacrebleu<2.0.0,>=1.4.12; extra == \"dev\"",
+        "filelock; extra == \"dev\"",
+        "hf-doc-builder; extra == \"dev\"",
+        "datasets>=2.15.0; extra == \"dev\"",
+        "ruff==0.14.10; extra == \"dev\"",
+        "GitPython<3.1.19; extra == \"dev\"",
+        "urllib3<2.0.0; extra == \"dev\"",
+        "libcst; extra == \"dev\"",
+        "rich; extra == \"dev\"",
+        "ty==0.0.20; extra == \"dev\"",
+        "tomli; extra == \"dev\"",
+        "transformers-mlinter==0.1.1; extra == \"dev\"",
+        "faiss-cpu; extra == \"dev\"",
+        "datasets>=2.15.0; extra == \"dev\"",
+        "sentencepiece!=0.1.92,>=0.1.91; extra == \"dev\"",
+        "protobuf; extra == \"dev\"",
+        "openai>=1.98.0; extra == \"dev\"",
+        "pydantic>=2; extra == \"dev\"",
+        "uvicorn; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "starlette; extra == \"dev\"",
+        "rich; extra == \"dev\"",
+        "torch>=2.4; extra == \"dev\"",
+        "accelerate>=1.1.0; extra == \"dev\"",
+        "mistral-common[image]>=1.10.0; extra == \"dev\"",
+        "fugashi>=1.0; extra == \"dev\"",
+        "ipadic<2.0,>=1.0.0; extra == \"dev\"",
+        "unidic_lite>=1.0.7; extra == \"dev\"",
+        "unidic>=1.0.2; extra == \"dev\"",
+        "rhoknp<1.3.1,>=1.1.0; extra == \"dev\"",
+        "sudachipy>=0.6.6; extra == \"dev\"",
+        "sudachidict_core>=20220729; extra == \"dev\"",
+        "scikit-learn; extra == \"dev\""
+      ],
+      "installer": "uv"
+    },
+    "pandas": {
+      "version": "3.0.6",
+      "requires_dist": [
+        "numpy>=1.26.0; python_version < \"3.14\"",
+        "numpy>=2.3.3; python_version >= \"3.14\"",
+        "python-dateutil>=2.8.2",
+        "tzdata; sys_platform == \"win32\"",
+        "tzdata; sys_platform == \"emscripten\"",
+        "hypothesis>=6.116.0; extra == \"test\"",
+        "pytest<9.1,>=8.3.4; extra == \"test\"",
+        "pytest-xdist>=3.6.1; extra == \"test\"",
+        "pyarrow>=13.0.0; extra == \"pyarrow\"",
+        "bottleneck>=1.4.2; extra == \"performance\"",
+        "numba>=0.60.0; extra == \"performance\"",
+        "numexpr>=2.10.2; extra == \"performance\"",
+        "scipy>=1.14.1; extra == \"computation\"",
+        "xarray>=2024.10.0; extra == \"computation\"",
+        "fsspec>=2024.10.0; extra == \"fss\"",
+        "s3fs>=2024.10.0; extra == \"aws\"",
+        "gcsfs>=2024.10.0; extra == \"gcp\"",
+        "odfpy>=1.4.1; extra == \"excel\"",
+        "openpyxl>=3.1.5; extra == \"excel\"",
+        "python-calamine>=0.3.0; extra == \"excel\"",
+        "pyxlsb>=1.0.10; extra == \"excel\"",
+        "xlrd>=2.0.1; extra == \"excel\"",
+        "xlsxwriter>=3.2.0; extra == \"excel\"",
+        "pyarrow>=13.0.0; extra == \"parquet\"",
+        "pyarrow>=13.0.0; extra == \"feather\"",
+        "pyiceberg>=0.8.1; extra == \"iceberg\"",
+        "tables>=3.10.1; extra == \"hdf5\"",
+        "pyreadstat>=1.2.8; extra == \"spss\"",
+        "SQLAlchemy>=2.0.36; extra == \"postgresql\"",
+        "psycopg2>=2.9.10; extra == \"postgresql\"",
+        "adbc-driver-postgresql>=1.2.0; extra == \"postgresql\"",
+        "SQLAlchemy>=2.0.36; extra == \"mysql\"",
+        "pymysql>=1.1.1; extra == \"mysql\"",
+        "SQLAlchemy>=2.0.36; extra == \"sql-other\"",
+        "adbc-driver-postgresql>=1.2.0; extra == \"sql-other\"",
+        "adbc-driver-sqlite>=1.2.0; extra == \"sql-other\"",
+        "beautifulsoup4>=4.12.3; extra == \"html\"",
+        "html5lib>=1.1; extra == \"html\"",
+        "lxml>=5.3.0; extra == \"html\"",
+        "lxml>=5.3.0; extra == \"xml\"",
+        "matplotlib>=3.9.3; extra == \"plot\"",
+        "jinja2>=3.1.5; extra == \"output-formatting\"",
+        "tabulate>=0.9.0; extra == \"output-formatting\"",
+        "PyQt5>=5.15.9; extra == \"clipboard\"",
+        "qtpy>=2.4.2; extra == \"clipboard\"",
+        "zstandard>=0.23.0; extra == \"compression\"",
+        "pytz>=2020.1; extra == \"timezone\"",
+        "adbc-driver-postgresql>=1.2.0; extra == \"all\"",
+        "adbc-driver-sqlite>=1.2.0; extra == \"all\"",
+        "beautifulsoup4>=4.12.3; extra == \"all\"",
+        "bottleneck>=1.4.2; extra == \"all\"",
+        "fastparquet>=2024.11.0; extra == \"all\"",
+        "fsspec>=2024.10.0; extra == \"all\"",
+        "gcsfs>=2024.10.0; extra == \"all\"",
+        "html5lib>=1.1; extra == \"all\"",
+        "hypothesis>=6.116.0; extra == \"all\"",
+        "jinja2>=3.1.5; extra == \"all\"",
+        "lxml>=5.3.0; extra == \"all\"",
+        "matplotlib>=3.9.3; extra == \"all\"",
+        "numba>=0.60.0; extra == \"all\"",
+        "numexpr>=2.10.2; extra == \"all\"",
+        "odfpy>=1.4.1; extra == \"all\"",
+        "openpyxl>=3.1.5; extra == \"all\"",
+        "psycopg2>=2.9.10; extra == \"all\"",
+        "pyarrow>=13.0.0; extra == \"all\"",
+        "pyiceberg>=0.8.1; extra == \"all\"",
+        "pymysql>=1.1.1; extra == \"all\"",
+        "PyQt5>=5.15.9; extra == \"all\"",
+        "pyreadstat>=1.2.8; extra == \"all\"",
+        "pytest>=8.3.4; extra == \"all\"",
+        "pytest-xdist>=3.6.1; extra == \"all\"",
+        "python-calamine>=0.3.0; extra == \"all\"",
+        "pytz>=2020.1; extra == \"all\"",
+        "pyxlsb>=1.0.10; extra == \"all\"",
+        "qtpy>=2.4.2; extra == \"all\"",
+        "scipy>=1.14.1; extra == \"all\"",
+        "s3fs>=2024.10.0; extra == \"all\"",
+        "SQLAlchemy>=2.0.36; extra == \"all\"",
+        "tables>=3.10.1; extra == \"all\"",
+        "tabulate>=0.9.0; extra == \"all\"",
+        "xarray>=2024.10.0; extra == \"all\"",
+        "xlrd>=2.0.1; extra == \"all\"",
+        "xlsxwriter>=3.2.0; extra == \"all\"",
+        "zstandard>=0.23.0; extra == \"all\""
+      ],
+      "installer": "uv"
+    },
+    "accelerate": {
+      "version": "1.15.0",
+      "requires_dist": [
+        "numpy>=1.17",
+        "packaging>=20.0",
+        "psutil",
+        "pyyaml",
+        "torch>=2.0.0",
+        "huggingface_hub>=0.21.0",
+        "safetensors>=0.4.3",
+        "ruff==0.13.1; extra == \"quality\"",
+        "pytest>=7.2.0; extra == \"test-prod\"",
+        "pytest-xdist; extra == \"test-prod\"",
+        "pytest-subtests; extra == \"test-prod\"",
+        "parameterized; extra == \"test-prod\"",
+        "pytest-order; extra == \"test-prod\"",
+        "datasets; extra == \"test-dev\"",
+        "diffusers; extra == \"test-dev\"",
+        "evaluate; extra == \"test-dev\"",
+        "peft; extra == \"test-dev\"",
+        "torchdata>=0.8.0; extra == \"test-dev\"",
+        "torchpippy>=0.2.0; extra == \"test-dev\"",
+        "transformers; extra == \"test-dev\"",
+        "scipy; extra == \"test-dev\"",
+        "scikit-learn; extra == \"test-dev\"",
+        "tqdm; extra == \"test-dev\"",
+        "bitsandbytes; extra == \"test-dev\"",
+        "timm; extra == \"test-dev\"",
+        "pytest>=7.2.0; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-subtests; extra == \"testing\"",
+        "parameterized; extra == \"testing\"",
+        "pytest-order; extra == \"testing\"",
+        "datasets; extra == \"testing\"",
+        "diffusers; extra == \"testing\"",
+        "evaluate; extra == \"testing\"",
+        "peft; extra == \"testing\"",
+        "torchdata>=0.8.0; extra == \"testing\"",
+        "torchpippy>=0.2.0; extra == \"testing\"",
+        "transformers; extra == \"testing\"",
+        "scipy; extra == \"testing\"",
+        "scikit-learn; extra == \"testing\"",
+        "tqdm; extra == \"testing\"",
+        "bitsandbytes; extra == \"testing\"",
+        "timm; extra == \"testing\"",
+        "deepspeed; extra == \"deepspeed\"",
+        "rich; extra == \"rich\"",
+        "torchao; extra == \"test-fp8\"",
+        "wandb; extra == \"test-trackers\"",
+        "tensorboard; extra == \"test-trackers\"",
+        "dvclive; extra == \"test-trackers\"",
+        "matplotlib; extra == \"test-trackers\"",
+        "swanlab[dashboard]; extra == \"test-trackers\"",
+        "trackio; extra == \"test-trackers\"",
+        "ruff==0.13.1; extra == \"dev\"",
+        "pytest>=7.2.0; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-subtests; extra == \"dev\"",
+        "parameterized; extra == \"dev\"",
+        "pytest-order; extra == \"dev\"",
+        "datasets; extra == \"dev\"",
+        "diffusers; extra == \"dev\"",
+        "evaluate; extra == \"dev\"",
+        "peft; extra == \"dev\"",
+        "torchdata>=0.8.0; extra == \"dev\"",
+        "torchpippy>=0.2.0; extra == \"dev\"",
+        "transformers; extra == \"dev\"",
+        "scipy; extra == \"dev\"",
+        "scikit-learn; extra == \"dev\"",
+        "tqdm; extra == \"dev\"",
+        "bitsandbytes; extra == \"dev\"",
+        "timm; extra == \"dev\"",
+        "rich; extra == \"dev\"",
+        "sagemaker; extra == \"sagemaker\""
+      ],
+      "installer": "uv"
+    },
+    "verl": {
+      "version": "0.9.1",
+      "requires_dist": [
+        "accelerate",
+        "codetiming",
+        "datasets",
+        "dill",
+        "hydra-core",
+        "numpy>=2.0.0",
+        "pandas",
+        "peft",
+        "pyarrow>=19.0.0",
+        "pybind11",
+        "pylatexenc",
+        "ray[default]>=2.41.0",
+        "torchdata",
+        "tensordict!=0.9.0,<=0.10.0,>=0.8.0",
+        "transformers!=5.6.0,<5.11,>=5.5.3",
+        "wandb",
+        "packaging>=20.0",
+        "tensorboard",
+        "pytest; extra == \"test\"",
+        "pre-commit; extra == \"test\"",
+        "py-spy; extra == \"test\"",
+        "pytest-asyncio; extra == \"test\"",
+        "pytest-rerunfailures; extra == \"test\"",
+        "math-verify; extra == \"math\"",
+        "mbridge; extra == \"mcore\"",
+        "TransferQueue>=0.1.10; extra == \"verl-core\"",
+        "pyzmq; extra == \"verl-core\"",
+        "msgspec; extra == \"verl-core\"",
+        "accelerate; extra == \"verl-core\"",
+        "codetiming; extra == \"verl-core\"",
+        "datasets; extra == \"verl-core\"",
+        "dill; extra == \"verl-core\"",
+        "hydra-core; extra == \"verl-core\"",
+        "numpy>=2.0.0; extra == \"verl-core\"",
+        "pandas; extra == \"verl-core\"",
+        "peft; extra == \"verl-core\"",
+        "pyarrow>=19.0.0; extra == \"verl-core\"",
+        "pybind11; extra == \"verl-core\"",
+        "pylatexenc; extra == \"verl-core\"",
+        "transformers; extra == \"verl-core\"",
+        "ray[default]>=2.41.0; extra == \"verl-core\"",
+        "torchdata; extra == \"verl-core\"",
+        "tensordict!=0.9.0,<=0.10.0,>=0.8.0; extra == \"verl-core\"",
+        "wandb; extra == \"verl-core\"",
+        "packaging>=20.0; extra == \"verl-core\"",
+        "tensorboard; extra == \"verl-core\"",
+        "fastapi; extra == \"verl-core\"",
+        "uvicorn; extra == \"verl-core\"",
+        "mathruler; extra == \"verl-core\"",
+        "qwen_vl_utils; extra == \"verl-core\"",
+        "torchcodec; extra == \"verl-core\"",
+        "cachetools; extra == \"verl-core\"",
+        "nvtx; extra == \"verl-core\"",
+        "pytest; extra == \"verl-core\"",
+        "pytest-asyncio; extra == \"verl-core\"",
+        "pytest-rerunfailures; extra == \"verl-core\"",
+        "flash-attn==2.8.3; extra == \"flash-attn-cu130torch211\"",
+        "cupy-cuda13x==14.0.1; extra == \"cupy-cu130\"",
+        "verl[verl-core]; extra == \"vllm\"",
+        "torch==2.11.0; extra == \"vllm\"",
+        "torchvision==0.26.0; extra == \"vllm\"",
+        "torchaudio==2.11.0; extra == \"vllm\"",
+        "vllm==0.24.0; extra == \"vllm\"",
+        "mbridge; extra == \"vllm\"",
+        "transformers==5.9.0; extra == \"vllm\"",
+        "verl[verl-core]; extra == \"sglang\"",
+        "torch==2.11.0; extra == \"sglang\"",
+        "torchvision==0.26.0; extra == \"sglang\"",
+        "torchaudio==2.11.0; extra == \"sglang\"",
+        "sglang==0.5.12; extra == \"sglang\"",
+        "sglang-kernel; extra == \"sglang\"",
+        "mbridge; extra == \"sglang\"",
+        "transformers==5.9.0; extra == \"sglang\"",
+        "verl[verl-core]; extra == \"fsdp\"",
+        "verl[flash-attn-cu130torch211]; extra == \"fsdp\"",
+        "verl[cupy-cu130]; extra == \"fsdp\"",
+        "torch==2.11.0; extra == \"fsdp\"",
+        "torchvision==0.26.0; extra == \"fsdp\"",
+        "torchaudio==2.11.0; extra == \"fsdp\"",
+        "liger-kernel>=0.8.2; extra == \"fsdp\"",
+        "trl==0.27.0; extra == \"fsdp\"",
+        "flash-linear-attention==0.5.2; extra == \"fsdp\"",
+        "verl[verl-core]; extra == \"megatron\"",
+        "verl[flash-attn-cu130torch211]; extra == \"megatron\"",
+        "verl[cupy-cu130]; extra == \"megatron\"",
+        "torch==2.11.0; extra == \"megatron\"",
+        "torchvision==0.26.0; extra == \"megatron\"",
+        "torchaudio==2.11.0; extra == \"megatron\"",
+        "nvidia-mathdx; extra == \"megatron\"",
+        "apex; extra == \"megatron\"",
+        "transformer-engine; extra == \"megatron\"",
+        "onnxscript; extra == \"megatron\"",
+        "trl==0.27.0; extra == \"megatron\"",
+        "matplotlib; extra == \"megatron\"",
+        "liger-kernel>=0.8.2; extra == \"megatron\"",
+        "mbridge; extra == \"megatron\"",
+        "megatron-bridge==0.5.2; extra == \"megatron\"",
+        "megatron-core; extra == \"megatron\"",
+        "nvidia-modelopt==0.44.0; extra == \"megatron\"",
+        "hf-transfer; extra == \"ci\"",
+        "sglang-router==0.2.2; extra == \"ci\"",
+        "veomni==0.1.11; extra == \"veomni-sft\"",
+        "verl[verl-core]; extra == \"cpu\"",
+        "torch==2.11.0; extra == \"cpu\"",
+        "torchvision==0.26.0; extra == \"cpu\"",
+        "torchaudio==2.11.0; extra == \"cpu\"",
+        "transformers==5.9.0; extra == \"cpu\"",
+        "megatron-core; extra == \"cpu\"",
+        "triton; extra == \"cpu\"",
+        "pytest; extra == \"cpu\"",
+        "pre-commit; extra == \"cpu\"",
+        "py-spy; extra == \"cpu\"",
+        "pytest-asyncio; extra == \"cpu\"",
+        "pytest-rerunfailures; extra == \"cpu\""
+      ],
+      "installer": "uv"
+    },
+    "qwen-vl-utils": {
+      "version": "0.0.14",
+      "requires_dist": [
+        "av",
+        "packaging",
+        "pillow",
+        "requests",
+        "decord; extra == 'decord'"
+      ],
+      "installer": "uv"
+    },
+    "tilelang": {
+      "version": "0.1.9",
+      "requires_dist": [
+        "apache-tvm-ffi>=0.1.2,~=0.1.0",
+        "torch-c-dlpack-ext; python_version < \"3.14\"",
+        "cloudpickle",
+        "ml-dtypes",
+        "numpy>=1.23.5",
+        "psutil",
+        "torch",
+        "torch>=2.4; platform_system == \"Darwin\"",
+        "setuptools; platform_system == \"Darwin\"",
+        "tqdm>=4.62.3",
+        "typing-extensions>=4.10.0",
+        "z3-solver<4.15.5,>=4.13.0",
+        "ml-dtypes>=0.5.1; extra == \"fp4\"",
+        "matplotlib; extra == \"vis\"",
+        "nvidia-cuda-nvcc>=13.0.48; extra == \"nvcc\"",
+        "nvidia-cuda-cccl>=13.0.50; extra == \"nvcc\""
+      ],
+      "installer": "uv"
+    },
+    "vllm": {
+      "version": "0.24.0",
+      "requires_dist": [
+        "regex",
+        "cachetools",
+        "psutil",
+        "sentencepiece",
+        "numpy",
+        "requests>=2.26.0",
+        "tqdm",
+        "blake3",
+        "py-cpuinfo",
+        "transformers>=5.5.3",
+        "tokenizers>=0.21.1",
+        "safetensors>=0.6.2",
+        "protobuf!=6.30.*,!=6.31.*,!=6.32.*,!=6.33.0.*,!=6.33.1.*,!=6.33.2.*,!=6.33.3.*,!=6.33.4.*,>=5.29.6",
+        "fastapi[standard]<0.137.0,>=0.133.0",
+        "starlette>=1.0.1",
+        "aiohttp>=3.13.3",
+        "openai>=2.0.0",
+        "pydantic>=2.12.0",
+        "prometheus_client>=0.18.0",
+        "pillow",
+        "prometheus-fastapi-instrumentator>=8.0.0",
+        "tiktoken>=0.6.0",
+        "lm-format-enforcer==0.11.3",
+        "llguidance<1.8.0,>=1.7.0; platform_machine == \"x86_64\" or platform_machine == \"arm64\" or platform_machine == \"aarch64\" or platform_machine == \"ppc64le\"",
+        "outlines_core==0.2.14",
+        "diskcache==5.6.3",
+        "lark==1.2.2",
+        "xgrammar<1.0.0,>=0.2.1; platform_machine == \"x86_64\" or platform_machine == \"aarch64\" or platform_machine == \"arm64\" or platform_machine == \"s390x\" or platform_machine == \"ppc64le\"",
+        "typing_extensions>=4.10",
+        "filelock>=3.16.1",
+        "partial-json-parser",
+        "jsonschema>=4.23.0",
+        "pyzmq>=25.0.0",
+        "msgspec",
+        "mistral_common[image]>=1.11.3",
+        "opencv-python-headless>=4.13.0",
+        "pyyaml",
+        "six>=1.16.0; python_version > \"3.11\"",
+        "setuptools<81.0.0,>=77.0.3; python_version > \"3.11\"",
+        "einops",
+        "compressed-tensors==0.17.0",
+        "depyf==0.20.0",
+        "cloudpickle",
+        "watchfiles",
+        "python-json-logger",
+        "ninja",
+        "pybase64",
+        "cbor2",
+        "ijson",
+        "setproctitle",
+        "openai-harmony>=0.0.3",
+        "anthropic>=0.71.0",
+        "model-hosting-container-standards<1.0.0,>=0.1.14",
+        "mcp",
+        "opentelemetry-sdk>=1.27.0",
+        "opentelemetry-api>=1.27.0",
+        "opentelemetry-exporter-otlp>=1.27.0",
+        "opentelemetry-semantic-conventions-ai>=0.4.1",
+        "numba==0.65.0",
+        "torch==2.11.0",
+        "torchaudio==2.11.0",
+        "torchvision==0.26.0",
+        "flashinfer-python==0.6.12",
+        "flashinfer-cubin==0.6.12",
+        "apache-tvm-ffi==0.1.9",
+        "tilelang==0.1.9",
+        "nvidia-cudnn-frontend>=1.19.1",
+        "fastsafetensors>=0.3.2",
+        "nvidia-cutlass-dsl[cu13]==4.5.2",
+        "quack-kernels>=0.3.3",
+        "tokenspeed-mla==0.1.2",
+        "humming-kernels[cu13]==0.1.6",
+        "zentorch==2.11.0.0; extra == \"zen\"",
+        "pandas; extra == \"bench\"",
+        "matplotlib; extra == \"bench\"",
+        "seaborn; extra == \"bench\"",
+        "datasets; extra == \"bench\"",
+        "scipy; extra == \"bench\"",
+        "plotly; extra == \"bench\"",
+        "tensorizer==2.10.1; extra == \"tensorizer\"",
+        "fastsafetensors>=0.3.2; extra == \"fastsafetensors\"",
+        "instanttensor>=0.1.5; extra == \"instanttensor\"",
+        "runai-model-streamer[azure,gcs,s3]>=0.15.7; extra == \"runai\"",
+        "av; extra == \"audio\"",
+        "scipy; extra == \"audio\"",
+        "soundfile; extra == \"audio\"",
+        "soxr; extra == \"audio\"",
+        "mistral_common[audio]; extra == \"audio\"",
+        "helion==1.1.0; extra == \"helion\"",
+        "smg-grpc-servicer[vllm]>=0.5.2; extra == \"grpc\"",
+        "opentelemetry-sdk>=1.26.0; extra == \"otel\"",
+        "opentelemetry-api>=1.26.0; extra == \"otel\"",
+        "opentelemetry-exporter-otlp>=1.26.0; extra == \"otel\"",
+        "opentelemetry-semantic-conventions-ai>=0.4.1; extra == \"otel\"",
+        "vllm-gguf-plugin>=0.0.2; extra == \"extra-quant\""
+      ],
+      "installer": "uv"
+    },
+    "torch": {
+      "version": "2.11.0",
+      "requires_dist": [
+        "filelock",
+        "typing-extensions>=4.10.0",
+        "setuptools<82",
+        "sympy>=1.13.3",
+        "networkx>=2.5.1",
+        "jinja2",
+        "fsspec>=0.8.5",
+        "cuda-toolkit[cublas,cudart,cufft,cufile,cupti,curand,cusolver,cusparse,nvjitlink,nvrtc,nvtx]==13.0.2; platform_system == \"Linux\"",
+        "cuda-bindings<14,>=13.0.3; platform_system == \"Linux\"",
+        "nvidia-cudnn-cu13==9.19.0.56; platform_system == \"Linux\"",
+        "nvidia-cusparselt-cu13==0.8.0; platform_system == \"Linux\"",
+        "nvidia-nccl-cu13==2.28.9; platform_system == \"Linux\"",
+        "nvidia-nvshmem-cu13==3.4.5; platform_system == \"Linux\"",
+        "triton==3.6.0; platform_system == \"Linux\"",
+        "optree>=0.13.0; extra == \"optree\"",
+        "opt-einsum>=3.3; extra == \"opt-einsum\"",
+        "pyyaml; extra == \"pyyaml\""
+      ],
+      "installer": "uv"
+    },
+    "tokenizers": {
+      "version": "0.22.2",
+      "requires_dist": [
+        "huggingface-hub>=0.16.4,<2.0",
+        "pytest ; extra == 'testing'",
+        "pytest-asyncio ; extra == 'testing'",
+        "requests ; extra == 'testing'",
+        "numpy ; extra == 'testing'",
+        "datasets ; extra == 'testing'",
+        "ruff ; extra == 'testing'",
+        "ty ; extra == 'testing'",
+        "sphinx ; extra == 'docs'",
+        "sphinx-rtd-theme ; extra == 'docs'",
+        "setuptools-rust ; extra == 'docs'",
+        "tokenizers[testing] ; extra == 'dev'"
+      ],
+      "installer": "uv"
+    },
+    "tensordict": {
+      "version": "0.10.0",
+      "requires_dist": [
+        "torch",
+        "numpy",
+        "cloudpickle",
+        "packaging",
+        "importlib_metadata",
+        "orjson; python_version < \"3.13\"",
+        "pyvers<0.2.0,>=0.1.0",
+        "pytest; extra == \"tests\"",
+        "pyyaml; extra == \"tests\"",
+        "pytest-instafail; extra == \"tests\"",
+        "pytest-rerunfailures; extra == \"tests\"",
+        "pytest-benchmark; extra == \"tests\"",
+        "torchsnapshot-nightly; extra == \"checkpointing\"",
+        "h5py>=3.8; extra == \"h5\"",
+        "pybind11; extra == \"dev\"",
+        "ninja; extra == \"dev\"",
+        "mypy>=1.0.0; extra == \"typecheck\""
+      ],
+      "installer": "uv"
+    },
+    "opencv-python-headless": {
+      "version": "5.0.0.93",
+      "requires_dist": [
+        "numpy<2.0; python_version < \"3.9\"",
+        "numpy>=2; python_version >= \"3.9\""
+      ],
+      "installer": "uv"
+    },
+    "flash-attn": {
+      "version": "2.8.3",
+      "requires_dist": [
+        "torch",
+        "einops"
+      ],
+      "installer": "uv"
+    },
+    "flash-linear-attention": {
+      "version": "0.5.2",
+      "requires_dist": [
+        "fla-core==0.5.2",
+        "transformers>=4.45.0",
+        "fla-core[cuda]==0.5.2; extra == \"cuda\"",
+        "fla-core[rocm]==0.5.2; extra == \"rocm\"",
+        "fla-core[xpu]==0.5.2; extra == \"xpu\"",
+        "fla-core[npu]==0.5.2; extra == \"npu\"",
+        "fla-core[cpu]==0.5.2; extra == \"cpu\"",
+        "tilelang>=0.1.9; extra == \"tilelang\"",
+        "causal-conv1d>=1.4.0; extra == \"conv1d\"",
+        "matplotlib; extra == \"benchmark\"",
+        "pandas; extra == \"benchmark\"",
+        "datasets>=3.3.0; extra == \"benchmark\"",
+        "pytest; extra == \"test\"",
+        "pytest-xdist; extra == \"test\""
+      ],
+      "installer": "uv"
+    },
+    "torchcodec": {
+      "version": "0.16.0",
+      "requires_dist": [
+        "numpy; extra == \"dev\"",
+        "pytest; extra == \"dev\"",
+        "pillow; extra == \"dev\""
+      ],
+      "installer": "uv"
+    },
+    "ray": {
+      "version": "2.58.0",
+      "requires_dist": [
+        "click>=7.0",
+        "filelock",
+        "jsonschema",
+        "msgpack<2.0.0,>=1.0.0",
+        "packaging>=24.2",
+        "protobuf>=3.20.3",
+        "pyyaml",
+        "requests",
+        "cupy-cuda12x; sys_platform != \"darwin\" and extra == \"cgraph\"",
+        "grpcio!=1.56.0; sys_platform == \"darwin\" and extra == \"client\"",
+        "grpcio; extra == \"client\"",
+        "numpy>=1.20; extra == \"data\"",
+        "pandas>=2.2.3; extra == \"data\"",
+        "pyarrow>=17.0.0; extra == \"data\"",
+        "fsspec; extra == \"data\"",
+        "aiohttp>=3.14.1; extra == \"default\"",
+        "aiohttp_cors; extra == \"default\"",
+        "colorful; extra == \"default\"",
+        "py-spy>=0.2.0; python_version < \"3.12\" and extra == \"default\"",
+        "py-spy>=0.4.0; python_version >= \"3.12\" and extra == \"default\"",
+        "requests; extra == \"default\"",
+        "grpcio>=1.42.0; extra == \"default\"",
+        "opencensus; extra == \"default\"",
+        "opentelemetry-sdk>=1.30.0; extra == \"default\"",
+        "opentelemetry-exporter-prometheus; extra == \"default\"",
+        "opentelemetry-proto; extra == \"default\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"default\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"default\"",
+        "prometheus_client>=0.7.1; extra == \"default\"",
+        "smart_open; extra == \"default\"",
+        "virtualenv!=20.21.1,>=20.0.24; extra == \"default\"",
+        "memray; sys_platform != \"win32\" and extra == \"observability\"",
+        "watchfiles; extra == \"serve\"",
+        "requests; extra == \"serve\"",
+        "opentelemetry-proto; extra == \"serve\"",
+        "mmh3; extra == \"serve\"",
+        "prometheus_client>=0.7.1; extra == \"serve\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"serve\"",
+        "fastapi>=0.133.0; extra == \"serve\"",
+        "colorful; extra == \"serve\"",
+        "uvicorn[standard]; extra == \"serve\"",
+        "aiohttp>=3.14.1; extra == \"serve\"",
+        "grpcio>=1.42.0; extra == \"serve\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"serve\"",
+        "py-spy>=0.4.0; python_version >= \"3.12\" and extra == \"serve\"",
+        "virtualenv!=20.21.1,>=20.0.24; extra == \"serve\"",
+        "py-spy>=0.2.0; python_version < \"3.12\" and extra == \"serve\"",
+        "opentelemetry-sdk>=1.30.0; extra == \"serve\"",
+        "starlette>=1.0.1; extra == \"serve\"",
+        "aiohttp_cors; extra == \"serve\"",
+        "smart_open; extra == \"serve\"",
+        "opentelemetry-exporter-prometheus; extra == \"serve\"",
+        "ray-haproxy<2.9.0,>=2.8.25; sys_platform == \"linux\" and extra == \"serve\"",
+        "opencensus; extra == \"serve\"",
+        "pandas; extra == \"tune\"",
+        "tensorboardX>=1.9; extra == \"tune\"",
+        "requests; extra == \"tune\"",
+        "pyarrow>=17.0.0; extra == \"tune\"",
+        "fsspec; extra == \"tune\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"tune\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"tune\"",
+        "cupy-cuda12x; sys_platform != \"darwin\" and extra == \"adag\"",
+        "watchfiles; extra == \"serve-grpc\"",
+        "requests; extra == \"serve-grpc\"",
+        "opentelemetry-proto; extra == \"serve-grpc\"",
+        "mmh3; extra == \"serve-grpc\"",
+        "prometheus_client>=0.7.1; extra == \"serve-grpc\"",
+        "fastapi>=0.133.0; extra == \"serve-grpc\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"serve-grpc\"",
+        "colorful; extra == \"serve-grpc\"",
+        "uvicorn[standard]; extra == \"serve-grpc\"",
+        "aiohttp>=3.14.1; extra == \"serve-grpc\"",
+        "grpcio>=1.42.0; extra == \"serve-grpc\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"serve-grpc\"",
+        "py-spy>=0.4.0; python_version >= \"3.12\" and extra == \"serve-grpc\"",
+        "virtualenv!=20.21.1,>=20.0.24; extra == \"serve-grpc\"",
+        "pyOpenSSL; extra == \"serve-grpc\"",
+        "py-spy>=0.2.0; python_version < \"3.12\" and extra == \"serve-grpc\"",
+        "opentelemetry-sdk>=1.30.0; extra == \"serve-grpc\"",
+        "starlette>=1.0.1; extra == \"serve-grpc\"",
+        "aiohttp_cors; extra == \"serve-grpc\"",
+        "smart_open; extra == \"serve-grpc\"",
+        "opentelemetry-exporter-prometheus; extra == \"serve-grpc\"",
+        "ray-haproxy<2.9.0,>=2.8.25; sys_platform == \"linux\" and extra == \"serve-grpc\"",
+        "opencensus; extra == \"serve-grpc\"",
+        "watchfiles; extra == \"serve-async-inference\"",
+        "requests; extra == \"serve-async-inference\"",
+        "opentelemetry-proto; extra == \"serve-async-inference\"",
+        "mmh3; extra == \"serve-async-inference\"",
+        "prometheus_client>=0.7.1; extra == \"serve-async-inference\"",
+        "fastapi>=0.133.0; extra == \"serve-async-inference\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"serve-async-inference\"",
+        "colorful; extra == \"serve-async-inference\"",
+        "uvicorn[standard]; extra == \"serve-async-inference\"",
+        "aiohttp>=3.14.1; extra == \"serve-async-inference\"",
+        "grpcio>=1.42.0; extra == \"serve-async-inference\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"serve-async-inference\"",
+        "py-spy>=0.4.0; python_version >= \"3.12\" and extra == \"serve-async-inference\"",
+        "virtualenv!=20.21.1,>=20.0.24; extra == \"serve-async-inference\"",
+        "py-spy>=0.2.0; python_version < \"3.12\" and extra == \"serve-async-inference\"",
+        "celery; extra == \"serve-async-inference\"",
+        "opentelemetry-sdk>=1.30.0; extra == \"serve-async-inference\"",
+        "starlette>=1.0.1; extra == \"serve-async-inference\"",
+        "aiohttp_cors; extra == \"serve-async-inference\"",
+        "smart_open; extra == \"serve-async-inference\"",
+        "opentelemetry-exporter-prometheus; extra == \"serve-async-inference\"",
+        "taskiq; extra == \"serve-async-inference\"",
+        "ray-haproxy<2.9.0,>=2.8.25; sys_platform == \"linux\" and extra == \"serve-async-inference\"",
+        "opencensus; extra == \"serve-async-inference\"",
+        "ray-cpp==2.58.0; extra == \"cpp\"",
+        "pandas; extra == \"rllib\"",
+        "tensorboardX>=1.9; extra == \"rllib\"",
+        "requests; extra == \"rllib\"",
+        "pyarrow>=17.0.0; extra == \"rllib\"",
+        "fsspec; extra == \"rllib\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"rllib\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"rllib\"",
+        "dm_tree; extra == \"rllib\"",
+        "gymnasium==1.2.2; extra == \"rllib\"",
+        "lz4; extra == \"rllib\"",
+        "ormsgpack>=1.7.0; extra == \"rllib\"",
+        "pyyaml; extra == \"rllib\"",
+        "scipy; extra == \"rllib\"",
+        "pandas; extra == \"train\"",
+        "tensorboardX>=1.9; extra == \"train\"",
+        "requests; extra == \"train\"",
+        "pyarrow>=17.0.0; extra == \"train\"",
+        "fsspec; extra == \"train\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"train\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"train\"",
+        "watchfiles; extra == \"air\"",
+        "requests; extra == \"air\"",
+        "fsspec; extra == \"air\"",
+        "opentelemetry-proto; extra == \"air\"",
+        "mmh3; extra == \"air\"",
+        "prometheus_client>=0.7.1; extra == \"air\"",
+        "fastapi>=0.133.0; extra == \"air\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"air\"",
+        "colorful; extra == \"air\"",
+        "tensorboardX>=1.9; extra == \"air\"",
+        "uvicorn[standard]; extra == \"air\"",
+        "pyarrow>=17.0.0; extra == \"air\"",
+        "aiohttp>=3.14.1; extra == \"air\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"air\"",
+        "grpcio>=1.42.0; extra == \"air\"",
+        "pandas; extra == \"air\"",
+        "py-spy>=0.4.0; python_version >= \"3.12\" and extra == \"air\"",
+        "virtualenv!=20.21.1,>=20.0.24; extra == \"air\"",
+        "py-spy>=0.2.0; python_version < \"3.12\" and extra == \"air\"",
+        "opentelemetry-sdk>=1.30.0; extra == \"air\"",
+        "starlette>=1.0.1; extra == \"air\"",
+        "aiohttp_cors; extra == \"air\"",
+        "smart_open; extra == \"air\"",
+        "opentelemetry-exporter-prometheus; extra == \"air\"",
+        "pandas>=2.2.3; extra == \"air\"",
+        "numpy>=1.20; extra == \"air\"",
+        "ray-haproxy<2.9.0,>=2.8.25; sys_platform == \"linux\" and extra == \"air\"",
+        "opencensus; extra == \"air\"",
+        "scipy; extra == \"all\"",
+        "grpcio; extra == \"all\"",
+        "watchfiles; extra == \"all\"",
+        "requests; extra == \"all\"",
+        "fsspec; extra == \"all\"",
+        "cupy-cuda12x; sys_platform != \"darwin\" and extra == \"all\"",
+        "opentelemetry-proto; extra == \"all\"",
+        "mmh3; extra == \"all\"",
+        "prometheus_client>=0.7.1; extra == \"all\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"all\"",
+        "fastapi>=0.133.0; extra == \"all\"",
+        "grpcio!=1.56.0; sys_platform == \"darwin\" and extra == \"all\"",
+        "colorful; extra == \"all\"",
+        "tensorboardX>=1.9; extra == \"all\"",
+        "pyyaml; extra == \"all\"",
+        "uvicorn[standard]; extra == \"all\"",
+        "pyarrow>=17.0.0; extra == \"all\"",
+        "aiohttp>=3.14.1; extra == \"all\"",
+        "grpcio>=1.42.0; extra == \"all\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"all\"",
+        "py-spy>=0.4.0; python_version >= \"3.12\" and extra == \"all\"",
+        "virtualenv!=20.21.1,>=20.0.24; extra == \"all\"",
+        "pandas; extra == \"all\"",
+        "py-spy>=0.2.0; python_version < \"3.12\" and extra == \"all\"",
+        "pyOpenSSL; extra == \"all\"",
+        "celery; extra == \"all\"",
+        "opentelemetry-sdk>=1.30.0; extra == \"all\"",
+        "starlette>=1.0.1; extra == \"all\"",
+        "aiohttp_cors; extra == \"all\"",
+        "smart_open; extra == \"all\"",
+        "opentelemetry-exporter-prometheus; extra == \"all\"",
+        "gymnasium==1.2.2; extra == \"all\"",
+        "taskiq; extra == \"all\"",
+        "pandas>=2.2.3; extra == \"all\"",
+        "numpy>=1.20; extra == \"all\"",
+        "ray-haproxy<2.9.0,>=2.8.25; sys_platform == \"linux\" and extra == \"all\"",
+        "opencensus; extra == \"all\"",
+        "ormsgpack>=1.7.0; extra == \"all\"",
+        "dm_tree; extra == \"all\"",
+        "lz4; extra == \"all\"",
+        "memray; sys_platform != \"win32\" and extra == \"all\"",
+        "scipy; extra == \"all-cpp\"",
+        "grpcio; extra == \"all-cpp\"",
+        "watchfiles; extra == \"all-cpp\"",
+        "requests; extra == \"all-cpp\"",
+        "fsspec; extra == \"all-cpp\"",
+        "cupy-cuda12x; sys_platform != \"darwin\" and extra == \"all-cpp\"",
+        "opentelemetry-proto; extra == \"all-cpp\"",
+        "mmh3; extra == \"all-cpp\"",
+        "prometheus_client>=0.7.1; extra == \"all-cpp\"",
+        "fastapi>=0.133.0; extra == \"all-cpp\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"all-cpp\"",
+        "grpcio!=1.56.0; sys_platform == \"darwin\" and extra == \"all-cpp\"",
+        "colorful; extra == \"all-cpp\"",
+        "tensorboardX>=1.9; extra == \"all-cpp\"",
+        "pyyaml; extra == \"all-cpp\"",
+        "uvicorn[standard]; extra == \"all-cpp\"",
+        "pyarrow>=17.0.0; extra == \"all-cpp\"",
+        "aiohttp>=3.14.1; extra == \"all-cpp\"",
+        "grpcio>=1.42.0; extra == \"all-cpp\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"all-cpp\"",
+        "py-spy>=0.4.0; python_version >= \"3.12\" and extra == \"all-cpp\"",
+        "virtualenv!=20.21.1,>=20.0.24; extra == \"all-cpp\"",
+        "pandas; extra == \"all-cpp\"",
+        "py-spy>=0.2.0; python_version < \"3.12\" and extra == \"all-cpp\"",
+        "pyOpenSSL; extra == \"all-cpp\"",
+        "celery; extra == \"all-cpp\"",
+        "opentelemetry-sdk>=1.30.0; extra == \"all-cpp\"",
+        "starlette>=1.0.1; extra == \"all-cpp\"",
+        "aiohttp_cors; extra == \"all-cpp\"",
+        "smart_open; extra == \"all-cpp\"",
+        "opentelemetry-exporter-prometheus; extra == \"all-cpp\"",
+        "gymnasium==1.2.2; extra == \"all-cpp\"",
+        "taskiq; extra == \"all-cpp\"",
+        "pandas>=2.2.3; extra == \"all-cpp\"",
+        "numpy>=1.20; extra == \"all-cpp\"",
+        "ray-haproxy<2.9.0,>=2.8.25; sys_platform == \"linux\" and extra == \"all-cpp\"",
+        "opencensus; extra == \"all-cpp\"",
+        "ormsgpack>=1.7.0; extra == \"all-cpp\"",
+        "dm_tree; extra == \"all-cpp\"",
+        "ray-cpp==2.58.0; extra == \"all-cpp\"",
+        "lz4; extra == \"all-cpp\"",
+        "memray; sys_platform != \"win32\" and extra == \"all-cpp\"",
+        "pybind11; extra == \"llm\"",
+        "watchfiles; extra == \"llm\"",
+        "requests; extra == \"llm\"",
+        "typer; extra == \"llm\"",
+        "fsspec; extra == \"llm\"",
+        "opentelemetry-proto; extra == \"llm\"",
+        "mmh3; extra == \"llm\"",
+        "prometheus_client>=0.7.1; extra == \"llm\"",
+        "pydantic<3,>=2.13.0; python_version >= \"3.14\" and extra == \"llm\"",
+        "fastapi>=0.133.0; extra == \"llm\"",
+        "colorful; extra == \"llm\"",
+        "uvicorn[standard]; extra == \"llm\"",
+        "pyarrow>=17.0.0; extra == \"llm\"",
+        "aiohttp>=3.14.1; extra == \"llm\"",
+        "ninja; extra == \"llm\"",
+        "vllm[audio]==0.26.0; extra == \"llm\"",
+        "grpcio>=1.42.0; extra == \"llm\"",
+        "pydantic<3,>=2.5.0; python_version < \"3.14\" and extra == \"llm\"",
+        "py-spy>=0.4.0; python_version >= \"3.12\" and extra == \"llm\"",
+        "virtualenv!=20.21.1,>=20.0.24; extra == \"llm\"",
+        "py-spy>=0.2.0; python_version < \"3.12\" and extra == \"llm\"",
+        "opentelemetry-sdk>=1.30.0; extra == \"llm\"",
+        "meson; extra == \"llm\"",
+        "starlette>=1.0.1; extra == \"llm\"",
+        "aiohttp_cors; extra == \"llm\"",
+        "smart_open; extra == \"llm\"",
+        "opentelemetry-exporter-prometheus; extra == \"llm\"",
+        "jsonschema; extra == \"llm\"",
+        "pandas>=2.2.3; extra == \"llm\"",
+        "async-timeout; python_version < \"3.11\" and extra == \"llm\"",
+        "numpy>=1.20; extra == \"llm\"",
+        "nixl-cu13==1.3.1; extra == \"llm\"",
+        "ray-haproxy<2.9.0,>=2.8.25; sys_platform == \"linux\" and extra == \"llm\"",
+        "opencensus; extra == \"llm\"",
+        "nixl==1.3.1; extra == \"llm\"",
+        "jsonref>=1.1.0; extra == \"llm\"",
+        "hf_transfer; extra == \"llm\""
+      ],
+      "installer": "uv"
+    },
+    "huggingface-hub": {
+      "version": "1.33.0",
+      "requires_dist": [
+        "click<9.0.0,>=8.4.2",
+        "filelock>=3.10.0",
+        "fsspec>=2023.5.0",
+        "hf-xet<2.0.0,>=1.6.0; platform_machine == \"x86_64\" or platform_machine == \"amd64\" or platform_machine == \"AMD64\" or platform_machine == \"arm64\" or platform_machine == \"aarch64\"",
+        "httpx<1,>=0.23.0",
+        "packaging>=20.9",
+        "pyyaml>=5.1",
+        "tomli>=1.1.0; python_version < \"3.11\"",
+        "tqdm>=4.42.1",
+        "typing-extensions>=4.1.0",
+        "authlib>=1.3.2; extra == \"oauth\"",
+        "fastapi; extra == \"oauth\"",
+        "httpx; extra == \"oauth\"",
+        "itsdangerous; extra == \"oauth\"",
+        "torch; extra == \"torch\"",
+        "safetensors[torch]; extra == \"torch\"",
+        "toml; extra == \"fastai\"",
+        "fastai>=2.4; extra == \"fastai\"",
+        "fastcore>=1.3.27; extra == \"fastai\"",
+        "hf-xet<2.0.0,>=1.6.0; extra == \"hf-xet\"",
+        "mcp<2.0.0,>=1.9.4; extra == \"mcp\"",
+        "authlib>=1.3.2; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "httpx; extra == \"testing\"",
+        "itsdangerous; extra == \"testing\"",
+        "jedi; extra == \"testing\"",
+        "Jinja2; extra == \"testing\"",
+        "pytest>=8.4.2; extra == \"testing\"",
+        "pytest-cov; extra == \"testing\"",
+        "pytest-env; extra == \"testing\"",
+        "pytest-xdist; extra == \"testing\"",
+        "pytest-vcr; extra == \"testing\"",
+        "pytest-asyncio; extra == \"testing\"",
+        "pytest-rerunfailures>=16.2; extra == \"testing\"",
+        "pytest-mock; extra == \"testing\"",
+        "urllib3<2.0; extra == \"testing\"",
+        "soundfile; extra == \"testing\"",
+        "Pillow; extra == \"testing\"",
+        "numpy; extra == \"testing\"",
+        "duckdb; extra == \"testing\"",
+        "fastapi; extra == \"testing\"",
+        "gradio>=5.0.0; extra == \"gradio\"",
+        "requests; extra == \"gradio\"",
+        "typing-extensions>=4.8.0; extra == \"typing\"",
+        "types-PyYAML; extra == \"typing\"",
+        "types-simplejson; extra == \"typing\"",
+        "types-toml; extra == \"typing\"",
+        "types-tqdm; extra == \"typing\"",
+        "types-urllib3; extra == \"typing\"",
+        "ruff>=0.9.0; extra == \"quality\"",
+        "mypy==1.15.0; extra == \"quality\"",
+        "libcst>=1.4.0; extra == \"quality\"",
+        "ty; extra == \"quality\"",
+        "authlib>=1.3.2; extra == \"all\"",
+        "fastapi; extra == \"all\"",
+        "httpx; extra == \"all\"",
+        "itsdangerous; extra == \"all\"",
+        "jedi; extra == \"all\"",
+        "Jinja2; extra == \"all\"",
+        "pytest>=8.4.2; extra == \"all\"",
+        "pytest-cov; extra == \"all\"",
+        "pytest-env; extra == \"all\"",
+        "pytest-xdist; extra == \"all\"",
+        "pytest-vcr; extra == \"all\"",
+        "pytest-asyncio; extra == \"all\"",
+        "pytest-rerunfailures>=16.2; extra == \"all\"",
+        "pytest-mock; extra == \"all\"",
+        "urllib3<2.0; extra == \"all\"",
+        "soundfile; extra == \"all\"",
+        "Pillow; extra == \"all\"",
+        "numpy; extra == \"all\"",
+        "duckdb; extra == \"all\"",
+        "fastapi; extra == \"all\"",
+        "ruff>=0.9.0; extra == \"all\"",
+        "mypy==1.15.0; extra == \"all\"",
+        "libcst>=1.4.0; extra == \"all\"",
+        "ty; extra == \"all\"",
+        "typing-extensions>=4.8.0; extra == \"all\"",
+        "types-PyYAML; extra == \"all\"",
+        "types-simplejson; extra == \"all\"",
+        "types-toml; extra == \"all\"",
+        "types-tqdm; extra == \"all\"",
+        "types-urllib3; extra == \"all\"",
+        "authlib>=1.3.2; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "httpx; extra == \"dev\"",
+        "itsdangerous; extra == \"dev\"",
+        "jedi; extra == \"dev\"",
+        "Jinja2; extra == \"dev\"",
+        "pytest>=8.4.2; extra == \"dev\"",
+        "pytest-cov; extra == \"dev\"",
+        "pytest-env; extra == \"dev\"",
+        "pytest-xdist; extra == \"dev\"",
+        "pytest-vcr; extra == \"dev\"",
+        "pytest-asyncio; extra == \"dev\"",
+        "pytest-rerunfailures>=16.2; extra == \"dev\"",
+        "pytest-mock; extra == \"dev\"",
+        "urllib3<2.0; extra == \"dev\"",
+        "soundfile; extra == \"dev\"",
+        "Pillow; extra == \"dev\"",
+        "numpy; extra == \"dev\"",
+        "duckdb; extra == \"dev\"",
+        "fastapi; extra == \"dev\"",
+        "ruff>=0.9.0; extra == \"dev\"",
+        "mypy==1.15.0; extra == \"dev\"",
+        "libcst>=1.4.0; extra == \"dev\"",
+        "ty; extra == \"dev\"",
+        "typing-extensions>=4.8.0; extra == \"dev\"",
+        "types-PyYAML; extra == \"dev\"",
+        "types-simplejson; extra == \"dev\"",
+        "types-toml; extra == \"dev\"",
+        "types-tqdm; extra == \"dev\"",
+        "types-urllib3; extra == \"dev\""
+      ],
+      "installer": "uv"
+    },
+    "safetensors": {
+      "version": "0.8.0",
+      "requires_dist": [
+        "safetensors[torch] ; extra == 'all'",
+        "safetensors[numpy] ; extra == 'all'",
+        "safetensors[jax] ; extra == 'all'",
+        "safetensors[paddlepaddle] ; extra == 'all'",
+        "safetensors[convert] ; extra == 'all'",
+        "safetensors[quality] ; extra == 'all'",
+        "safetensors[testing] ; extra == 'all'",
+        "safetensors[torch] ; extra == 'convert'",
+        "huggingface-hub>=1.4 ; extra == 'convert'",
+        "safetensors[all] ; extra == 'dev'",
+        "safetensors[pinned-tf] ; extra == 'dev'",
+        "safetensors[numpy] ; extra == 'jax'",
+        "flax>=0.6.3 ; extra == 'jax'",
+        "jax>=0.3.25 ; extra == 'jax'",
+        "jaxlib>=0.3.25 ; extra == 'jax'",
+        "mlx>=0.0.9 ; extra == 'mlx'",
+        "numpy>=1.24.6 ; extra == 'numpy'",
+        "safetensors[numpy] ; extra == 'paddlepaddle'",
+        "paddlepaddle>=2.4.1 ; extra == 'paddlepaddle'",
+        "safetensors[numpy] ; extra == 'pinned-tf'",
+        "tensorflow==2.18.0 ; extra == 'pinned-tf'",
+        "ruff ; extra == 'quality'",
+        "safetensors[numpy] ; extra == 'tensorflow'",
+        "tensorflow>=2.11.0 ; extra == 'tensorflow'",
+        "safetensors[numpy] ; extra == 'testing'",
+        "h5py>=3.7.0 ; extra == 'testing'",
+        "setuptools-rust>=1.12.0 ; extra == 'testing'",
+        "pytest>=9.0 ; extra == 'testing'",
+        "pytest-benchmark>=5.2 ; extra == 'testing'",
+        "hypothesis>=6.70.2 ; extra == 'testing'",
+        "fsspec>=2024.6.0 ; extra == 'testing'",
+        "s3fs>=2024.6.0 ; extra == 'testing'",
+        "safetensors[numpy] ; extra == 'tf-nightly'",
+        "tf-nightly ; extra == 'tf-nightly'",
+        "safetensors[numpy] ; extra == 'torch'",
+        "torch>=2.4 ; extra == 'torch'"
+      ],
+      "installer": "uv"
+    },
+    "numpy": {
+      "version": "2.3.5",
+      "requires_dist": [],
+      "installer": "uv"
+    }
+  },
+  "module_presence": {
+    "torch": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "0387d8b811b289287479c8bfdf4e1dac3a71b246f938d82da1331cf2dc8bf001"
+    },
+    "torchvision": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "600bded9685ed2210e4b6420925a3f9b18973de196aea11da9a2a0619226b3bf"
+    },
+    "torchaudio": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "f0e07710f18256217b2e0056cbf6d4c99505e872542296d323ca2e446c27ea55"
+    },
+    "transformers": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "b5d7bd82ca7d2929e9485c7e2ef0a54e4f051253712899de0f8b1b5e3aafb540"
+    },
+    "tokenizers": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "148ecb122f3fee03be9abb1fe213dd85fbc691bf2ad2b437ff531533781a6a3a"
+    },
+    "huggingface_hub": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "284be6e001a015f9933e756c6f6552b7266d811aff7b5124c960b5f1fc66b27e"
+    },
+    "diffusers": {
+      "found": false
+    },
+    "diffsynth": {
+      "found": false
+    },
+    "accelerate": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "a479388e26b653f219788e744c45e73d0c172f021d441725ff2af89bf6ee7a8f"
+    },
+    "peft": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "a68ffbaa8b3b0777597c7c48077080a890aca8a31acffcfc1c7469f4d08f7a57"
+    },
+    "swift": {
+      "found": false
+    },
+    "verl": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "513270d0f054e18b66dce36220d001df3030c94b6967864ad3824fbd9d742285"
+    },
+    "vllm": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "fd3708e5a13abe98c566afd79a6b1987cef70c0548bb19c1218ff6a7f9d43346"
+    },
+    "fla": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "b7e0d26abd3162884ce94f37d0210a25344025055383d377937f300a3bf5f45d"
+    },
+    "tilelang": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "f382ca163aa6be1852768035e29d3cdc5514a38dfd19e27b42948726226e54b8"
+    },
+    "triton": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "96e68192cab5dd11fa30d5dc95d56c29ceecda53477fc3af8f37bee026b7a021"
+    },
+    "cv2": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "69192904dead3ada37c8b305ee850ac7474924730881a9d5e0b2390944eb5aae"
+    },
+    "numpy": {
+      "found": true,
+      "module_file": "__init__.py",
+      "entry_file_sha256": "93924ac4b793328947dfd9eb9355e54eccdfd26a92c8dd52188aed2e52c7eb38"
+    }
+  },
+  "pip_check": {
+    "exit_code": 1,
+    "output": "/local/workspace/envs/verl38/bin/python: No module named pip\n"
+  },
+  "GPU": {
+    "checked": true,
+    "basic_tensor_check": {
+      "exit_code": 0,
+      "output": "{\"torch\": \"2.11.0+cu130\", \"cuda_runtime\": \"13.0\", \"cuda_available\": true, \"arch_list\": [\"sm_75\", \"sm_80\", \"sm_86\", \"sm_90\", \"sm_100\", \"sm_120\"], \"gpu\": \"NVIDIA RPBZZZ6\", \"capability\": [12, 0], \"tiny_bfloat16_result\": 4.0}\n"
+    },
+    "training_or_model_inference_run": false
+  },
+  "original_environment_modified": false,
+  "credentials_included": false
+}

robot/legacy_sft_grpo/import_validation.json

@@ -0,0 +1,51 @@
+{
+  "torch": {
+    "ok": true,
+    "version": "2.11.0+cu130"
+  },
+  "torchvision": {
+    "ok": true,
+    "version": "0.26.0+cu130"
+  },
+  "torchaudio": {
+    "ok": true,
+    "version": "2.11.0+cu130"
+  },
+  "transformers": {
+    "ok": true,
+    "version": "5.9.0"
+  },
+  "peft": {
+    "ok": true,
+    "version": "0.21.0"
+  },
+  "accelerate": {
+    "ok": true,
+    "version": "1.15.0"
+  },
+  "safetensors": {
+    "ok": true,
+    "version": "0.8.0"
+  },
+  "decord": {
+    "ok": false,
+    "error_type": "ModuleNotFoundError",
+    "error": "No module named 'decord'"
+  },
+  "vllm": {
+    "ok": true,
+    "version": "0.24.0"
+  },
+  "verl": {
+    "ok": true,
+    "version": "0.9.1"
+  },
+  "fla": {
+    "ok": true,
+    "version": "0.5.2"
+  },
+  "tilelang": {
+    "ok": true,
+    "version": "0.1.9"
+  }
+}

robot/legacy_sft_grpo/requirements.core.in

@@ -0,0 +1,27 @@
+# Observed role-specific root requirements; transitive dependencies resolved/locked separately.
+torch==2.11.0+cu130
+torchvision==0.26.0+cu130
+torchaudio==2.11.0+cu130
+triton==3.6.0
+transformers==5.9.0
+tokenizers==0.22.2
+huggingface-hub==1.33.0
+accelerate==1.15.0
+verl==0.9.1
+vllm==0.24.0
+peft==0.21.0
+safetensors==0.8.0
+qwen-vl-utils==0.0.14
+flash-linear-attention==0.5.2
+tilelang==0.1.9
+tensordict==0.10.0
+ray==2.58.0
+torchcodec==0.16.0
+numpy==2.3.5
+pillow==12.3.0
+opencv-python-headless==5.0.0.93
+h5py==3.16.0
+pandas==3.0.6
+pyarrow==25.0.1
+datasets==5.0.1
+tqdm==4.70.1

robot/legacy_sft_grpo/requirements.observed-full.lock.txt

@@ -0,0 +1,256 @@
+absl-py==2.5.0
+accelerate==1.15.0
+agent-detector==2.0.0
+aiohappyeyeballs==2.7.1
+aiohttp==3.14.3
+aiohttp-cors==0.8.1
+aiosignal==1.4.0
+annotated-doc==0.0.5
+annotated-types==0.8.0
+anthropic==1.8.0
+antlr4-python3-runtime==4.9.3
+anyio==4.15.1
+apache-tvm-ffi==0.1.9
+astor==0.8.1
+attrs==26.1.0
+av==18.1.0
+blake3==1.0.9
+cachetools==7.2.0
+cbor2==6.1.4
+certifi==2026.7.22
+cffi==2.1.1
+charset-normalizer==3.5.1
+click==8.5.0
+cloudpickle==3.1.2
+codetiming==1.4.0
+colorful==0.5.8
+compressed-tensors==0.17.0
+cryptography==50.0.1
+cuda-bindings==13.4.3
+cuda-core==1.2.1
+cuda-pathfinder==1.8.2
+cuda-python==13.4.1
+cuda-tile==1.3.0
+cuda-toolkit==13.0.2
+datasets==5.0.1
+depyf==0.20.0
+detect-installer==0.2.1
+dill==0.4.1
+diskcache==5.6.3
+distlib==0.4.3
+dnspython==2.8.0
+docstring-parser==0.18.0
+einops==0.8.2
+email-validator==2.3.0
+fastapi==0.136.3
+fastapi-cli==0.0.32
+fastapi-cloud-cli==0.26.0
+fastar==0.12.0
+fastsafetensors==0.4.0
+filelock==4.0.5
+fla-core==0.5.2
+flash-attn==2.8.3
+flash-linear-attention==0.5.2
+flashinfer-cubin==0.6.12
+flashinfer-python==0.6.12
+frozenlist==1.8.0
+fsspec==2026.6.0
+google-api-core==2.39.0
+google-auth==2.58.1
+googleapis-common-protos==1.75.4
+grpcio==1.84.0
+h11==0.16.0
+h5py==3.16.0
+hf-xet==1.6.0
+httpcore==1.0.9
+httpcore2==2.13.1
+httptools==0.8.0
+httpx==0.28.1
+httpx2==2.13.1
+huggingface-hub==1.33.0
+humming-kernels==0.1.6
+hydra-core==1.3.7
+idna==3.20
+ijson==3.5.1
+importlib-metadata==9.0.1
+iniconfig==2.3.0
+interegular==0.3.3
+jinja2==3.1.6
+jiter==0.17.0
+jmespath==1.1.0
+jsonschema==4.26.0
+jsonschema-specifications==2025.9.1
+lark==1.2.2
+llguidance==1.7.6
+llvmlite==0.47.0
+lm-format-enforcer==0.11.3
+loguru==0.7.3
+markdown==3.11
+markdown-it-py==4.2.0
+markupsafe==3.0.3
+mathruler==0.1.0
+mbridge==0.15.1
+mcp==2.2.0
+mcp-types==2.2.0
+mdurl==0.1.2
+mistral-common==1.12.0
+ml-dtypes==0.6.0
+model-hosting-container-standards==0.1.16
+mpmath==1.3.0
+msgpack==1.2.2
+msgspec==0.21.1
+multidict==6.9.1
+multiprocess==0.70.19
+networkx==3.7
+ninja==1.13.2
+numba==0.65.0
+numpy==2.3.5
+nvidia-cublas==13.1.0.3
+nvidia-cuda-cccl==13.3.4.3.1
+nvidia-cuda-crt==13.4.92
+nvidia-cuda-cupti==13.0.85
+nvidia-cuda-nvcc==13.2.86
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-runtime==13.0.96
+nvidia-cuda-tileiras==13.2.86
+nvidia-cudnn-cu13==9.19.0.56
+nvidia-cudnn-frontend==1.28.0
+nvidia-cufft==12.0.0.61
+nvidia-cufile==1.15.1.6
+nvidia-curand==10.4.0.35
+nvidia-cusolver==12.0.4.66
+nvidia-cusparse==12.6.3.3
+nvidia-cusparselt-cu13==0.8.0
+nvidia-cutlass-dsl==4.5.2
+nvidia-cutlass-dsl-libs-base==4.5.2
+nvidia-cutlass-dsl-libs-cu13==4.5.2
+nvidia-ml-py==13.615.71
+nvidia-nccl-cu13==2.28.9
+nvidia-nvjitlink==13.0.88
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvtx==13.0.85
+nvidia-nvvm==13.2.86
+nvtx==0.2.16
+omegaconf==2.3.1
+openai==3.19.2
+openai-harmony==0.0.8
+opencensus==0.11.4
+opencensus-context==0.1.3
+opencv-python-headless==5.0.0.93
+opentelemetry-api==1.45.0
+opentelemetry-exporter-http-transport==0.66b0
+opentelemetry-exporter-otlp==1.45.0
+opentelemetry-exporter-otlp-common==0.66b0
+opentelemetry-exporter-otlp-proto-common==1.45.0
+opentelemetry-exporter-otlp-proto-grpc==1.45.0
+opentelemetry-exporter-otlp-proto-http==1.45.0
+opentelemetry-exporter-prometheus==0.66b0
+opentelemetry-proto==1.45.0
+opentelemetry-sdk==1.45.0
+opentelemetry-semantic-conventions==0.66b0
+opentelemetry-semantic-conventions-ai==0.5.1
+orjson==3.12.0
+outlines-core==0.2.14
+packaging==25.0
+pandas==3.0.6
+partial-json-parser==0.2.1.1.post7
+peft==0.21.0
+pillow==12.3.0
+platformdirs==4.12.1
+pluggy==1.6.0
+prometheus-client==0.26.0
+prometheus-fastapi-instrumentator==8.1.0
+propcache==0.5.4
+proto-plus==1.28.4
+protobuf==7.36.2
+psutil==7.2.2
+py-cpuinfo==9.0.0
+py-spy==0.4.2
+pyarrow==25.0.1
+pyasn1==0.6.4
+pyasn1-modules==0.4.2
+pybase64==1.5.0
+pybind11==3.1.0
+pycountry==26.2.16
+pycparser==3.0
+pydantic==2.13.5
+pydantic-core==2.46.5
+pydantic-extra-types==2.11.1
+pydantic-settings==2.15.0
+pyelftools==0.33
+pygments==2.21.0
+pyjwt==2.15.0
+pylatexenc==2.11
+pytest==9.1.1
+pytest-asyncio==1.4.0
+pytest-rerunfailures==16.7
+python-dateutil==2.9.0.post0
+python-discovery==1.6.1
+python-dotenv==1.2.3
+python-json-logger==4.2.0
+python-multipart==0.0.32
+pyvers==0.1.0
+pyyaml==6.0.3
+pyzmq==27.2.0
+quack-kernels==0.5.0
+qwen-vl-utils==0.0.14
+ray==2.58.0
+referencing==0.37.0
+regex==2026.9.10
+requests==2.34.2
+rich==15.0.0
+rich-toolkit==0.20.5
+rignore==0.8.1
+rpds-py==2026.6.3
+safetensors==0.8.0
+sentencepiece==0.2.2
+sentry-sdk==2.70.0
+setproctitle==1.3.7
+setuptools==80.10.2
+shellingham==1.5.4
+six==1.17.0
+smart-open==8.0.1
+sniffio==1.3.1
+sse-starlette==3.4.11
+starlette==1.7.0
+supervisor==4.3.0
+sympy==1.14.0
+tabulate==0.10.0
+tensorboard==2.21.0
+tensorboard-data-server==0.7.2
+tensordict==0.10.0
+tiktoken==0.14.0
+tilelang==0.1.9
+tokenizers==0.22.2
+tokenspeed-mla==0.1.2
+tokenspeed-triton==3.8.10.post20260920
+torch==2.11.0
+torch-c-dlpack-ext==0.1.5
+torchaudio==2.11.0
+torchcodec==0.16.0
+torchdata==0.11.0
+torchvision==0.26.0
+tqdm==4.70.1
+transferqueue==0.1.11
+transformers==5.9.0
+triton==3.6.0
+truststore==0.10.4
+typer==0.27.2
+typing-extensions==4.16.0
+typing-inspection==0.4.4
+urllib3==2.8.0
+uvicorn==0.54.0
+uvloop==0.22.1
+verl==0.9.1
+virtualenv==21.13.0
+vllm==0.24.0
+wandb==0.30.0
+watchfiles==1.3.0
+websockets==17.1
+werkzeug==3.1.9
+wrapt==2.5.0
+xgrammar==0.2.3
+xxhash==4.0.1
+yarl==1.25.1
+z3-solver==4.15.4.0
+zipp==4.1.0

source-overlays/egoexo/code/core/data/build_testset_from_masks.py

@@ -0,0 +1,141 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+from project_paths import seg_mask_root, seg_mask_dir
+import os, sys, json, argparse
+import numpy as np
+import cv2
+
+def frame_stats(mask_path):
+    m = cv2.imread(mask_path, cv2.IMREAD_GRAYSCALE)
+    if m is None:
+        return None
+    tot = m.size
+    hand = float((m == 1).sum()) / tot
+    obj = float((m == 2).sum()) / tot
+    fg = float((m > 0).sum()) / tot
+    return (hand, obj, fg)
+
+def longest_run(bools):
+    best = cur = 0
+    for b in bools:
+        cur = cur + 1 if b else 0
+        best = max(best, cur)
+    return best
+
+def score_clip(cdir, args):
+    ego_dir = seg_mask_dir(os.path.basename(cdir))
+    exo_dir = os.path.join(cdir, 'exo_mask')
+    if not os.path.isdir(ego_dir):
+        return None
+    ego_files = sorted((f for f in os.listdir(ego_dir) if f.endswith('.png')))
+    if not ego_files:
+        return None
+    ego = [frame_stats(os.path.join(ego_dir, f)) for f in ego_files]
+    ego = [e for e in ego if e is not None]
+    n = len(ego)
+    if n == 0:
+        return None
+    hand_ok = [args.hand_min <= h <= args.hand_max for h, o, fg in ego]
+    obj_ok = [args.obj_min <= o <= args.obj_max for h, o, fg in ego]
+    both_ok = [ho and oo for ho, oo in zip(hand_ok, obj_ok)]
+    ego_degen = any((fg > args.degen_thr for h, o, fg in ego))
+    ego_hand_cov = sum(hand_ok) / n
+    ego_obj_cov = sum(obj_ok) / n
+    ego_both_cov = sum(both_ok) / n
+    ego_hand_run = longest_run(hand_ok) / n
+    exo_fg_cov = 0.0
+    exo_degen = False
+    if os.path.isdir(exo_dir):
+        exo_files = sorted((f for f in os.listdir(exo_dir) if f.endswith('.png')))
+        exo = [frame_stats(os.path.join(exo_dir, f)) for f in exo_files]
+        exo = [e for e in exo if e is not None]
+        if exo:
+            exo_fg_cov = sum((1 for h, o, fg in exo if fg > 0)) / len(exo)
+            exo_degen = any((fg > args.degen_thr for h, o, fg in exo))
+    passed = ego_both_cov >= args.min_both and ego_hand_run >= args.min_run and (not ego_degen) and (not exo_degen)
+    score = 0.0 if ego_degen or exo_degen else 0.5 * ego_both_cov + 0.3 * ego_hand_cov + 0.2 * ego_hand_run
+    reason = []
+    if ego_both_cov < args.min_both:
+        reason.append(''.join(['ego_both_cov<', format(args.min_both, ''), '(', format(ego_both_cov, ''.join(['.2f'])), ')']))
+    if ego_hand_run < args.min_run:
+        reason.append(''.join(['ego_hand_run<', format(args.min_run, ''), '(', format(ego_hand_run, ''.join(['.2f'])), ')']))
+    if ego_degen:
+        reason.append('ego_degen(铺满)')
+    if exo_degen:
+        reason.append('exo_degen(铺满)')
+    return {'n_frames': n, 'ego_hand_cov': round(ego_hand_cov, 3), 'ego_obj_cov': round(ego_obj_cov, 3), 'ego_both_cov': round(ego_both_cov, 3), 'ego_hand_run': round(ego_hand_run, 3), 'exo_fg_cov': round(exo_fg_cov, 3), 'ego_degen': ego_degen, 'exo_degen': exo_degen, 'score': round(score, 4), 'passed': passed, 'reject_reason': reason}
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--pool', required=True, help='hoi_dataset_testpool 目录(含 clips/ 与 manifest.jsonl)')
+    ap.add_argument('--hand_min', type=float, default=0.001, help='ego 手面积占比下限(过滤噪点)')
+    ap.add_argument('--hand_max', type=float, default=0.4, help='ego 手面积占比上限(过滤误检)')
+    ap.add_argument('--obj_min', type=float, default=0.0005)
+    ap.add_argument('--obj_max', type=float, default=0.5)
+    ap.add_argument('--degen_thr', type=float, default=0.7, help='fg 占比超此判为铺满误检')
+    ap.add_argument('--min_both', type=float, default=0.6, help='ego 同帧手+物覆盖率阈值')
+    ap.add_argument('--min_run', type=float, default=0.4, help='ego 有手帧最长连续占比阈值')
+    args = ap.parse_args()
+    pool = os.path.abspath(args.pool)
+    clips_root = os.path.join(pool, 'clips')
+    manifest = os.path.join(pool, 'manifest.jsonl')
+    if not os.path.isdir(clips_root):
+        print(''.join(['!! 候选池 clips/ 不存在: ', format(clips_root, '')]))
+        sys.exit(1)
+    recs = []
+    if os.path.isfile(manifest):
+        for line in open(manifest):
+            try:
+                recs.append(json.loads(line))
+            except Exception:
+                pass
+    else:
+        recs = [{'clip_id': d, 'dir': os.path.join('clips', d)} for d in sorted(os.listdir(clips_root))]
+    passed, rejected = ([], [])
+    scores_all = []
+    for r in recs:
+        cid = r['clip_id']
+        cdir = os.path.join(pool, r.get('dir', os.path.join('clips', cid)))
+        s = score_clip(cdir, args)
+        if s is None:
+            rejected.append({'clip_id': cid, 'reject_reason': ['无掩码/读取失败']})
+            continue
+        scores_all.append(s['score'])
+        meta = {}
+        mp = os.path.join(cdir, 'meta.json')
+        if os.path.isfile(mp):
+            try:
+                meta = json.load(open(mp))
+            except Exception:
+                pass
+        entry = {'clip_id': cid, 'take_uid': meta.get('take_uid', cid.rsplit('_', 1)[0]), 'take_name': r.get('take_name') or meta.get('take_name', ''), 'prompt': r.get('prompt') or meta.get('prompt', ''), 'start_sec': meta.get('start_sec'), 'num_frames': s['n_frames'], 'exo_frames': os.path.join(cdir, 'exo_frames'), 'exo_mask': os.path.join(cdir, 'exo_mask'), 'ego_frames': os.path.join(cdir, 'ego_frames'), 'ego_mask': seg_mask_dir(os.path.basename(cdir)), **{k: s[k] for k in ('score', 'ego_hand_cov', 'ego_obj_cov', 'ego_both_cov', 'ego_hand_run', 'exo_fg_cov', 'ego_degen', 'exo_degen')}}
+        if s['passed']:
+            passed.append(entry)
+        else:
+            rejected.append({'clip_id': cid, 'score': s['score'], 'reject_reason': s['reject_reason']})
+    passed.sort(key=lambda e: e['score'], reverse=True)
+    json.dump(passed, open(os.path.join(pool, 'testset_manifest.json'), 'w'), ensure_ascii=False, indent=1)
+    json.dump(rejected, open(os.path.join(pool, 'testset_rejected.json'), 'w'), ensure_ascii=False, indent=1)
+    print(''.join(['\n===== 候选池 ', format(len(recs), ''), ' clip: 通过 ', format(len(passed), ''), '  淘汰 ', format(len(rejected), ''), ' =====']))
+    if scores_all:
+        arr = np.array(scores_all)
+        print(''.join(['score 分布: min=', format(arr.min(), ''.join(['.3f'])), ' p25=', format(np.percentile(arr, 25), ''.join(['.3f'])), ' 中位=', format(np.median(arr), ''.join(['.3f'])), ' p75=', format(np.percentile(arr, 75), ''.join(['.3f'])), ' max=', format(arr.max(), ''.join(['.3f']))]))
+        bins = np.linspace(0, 1, 11)
+        hist, _ = np.histogram(arr, bins=bins)
+        for i, h in enumerate(hist):
+            print(''.join(['  [', format(bins[i], ''.join(['.1f'])), ',', format(bins[i + 1], ''.join(['.1f'])), ') ', format('#' * int(h * 40 / max(1, hist.max())), ''), ' ', format(h, '')]))
+    from collections import Counter
+    rc = Counter()
+    for r in rejected:
+        for reason in r.get('reject_reason', []):
+            rc[reason.split('(')[0]] += 1
+    print('淘汰原因 top:', dict(rc.most_common(8)))
+    print(''.join(['\n写出: ', format(os.path.join(pool, 'testset_manifest.json'), ''), ' (', format(len(passed), ''), ' 条, 按 score 降序)']))
+    print(''.join(['      ', format(os.path.join(pool, 'testset_rejected.json'), ''), ' (', format(len(rejected), ''), ' 条)']))
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/caption_clip_probe.py

@@ -0,0 +1,191 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+from project_paths import seg_mask_root, seg_mask_dir
+import argparse
+import json
+import os
+import re
+import sys
+import numpy as np
+import torch
+from PIL import Image
+CLIP_DIR = _project_path('models/CLIP-ViT-H-14-laion2B-s32B-b79K')
+HERE = os.path.dirname(os.path.abspath(__file__))
+_warned_diff_miss = set()
+
+def mask_union_bbox(mask_dir, pad_frac=0.15):
+    """"""
+    if not os.path.isdir(mask_dir):
+        return None
+    fs = sorted((f for f in os.listdir(mask_dir) if f.endswith('.png')))
+    if not fs:
+        return None
+    H = W = None
+    y0 = y1 = x0 = x1 = None
+    for f in fs:
+        m = np.asarray(Image.open(os.path.join(mask_dir, f)))
+        if m.ndim == 3:
+            m = m[..., 0]
+        H, W = m.shape
+        ys, xs = np.where(m > 0)
+        if len(ys) == 0:
+            continue
+        fy0, fy1, fx0, fx1 = (int(ys.min()), int(ys.max()), int(xs.min()), int(xs.max()))
+        y0 = fy0 if y0 is None else min(y0, fy0)
+        y1 = fy1 if y1 is None else max(y1, fy1)
+        x0 = fx0 if x0 is None else min(x0, fx0)
+        x1 = fx1 if x1 is None else max(x1, fx1)
+    if y0 is None:
+        return None
+    h, w = (y1 - y0 + 1, x1 - x0 + 1)
+    py, px = (int(h * pad_frac), int(w * pad_frac))
+    y0, y1 = (max(0, y0 - py), min(H - 1, y1 + py))
+    x0, x1 = (max(0, x0 - px), min(W - 1, x1 + px))
+    return (y0, y1, x0, x1, H, W)
+
+def crop_by_bbox(img, bbox):
+    y0, y1, x0, x1, H, W = bbox
+    Wf, Hf = img.size
+    sy, sx = (Hf / H, Wf / W)
+    yy0, yy1 = (int(y0 * sy), min(Hf, int(y1 * sy) + 1))
+    xx0, xx1 = (int(x0 * sx), min(Wf, int(x1 * sx) + 1))
+    return img.crop((xx0, yy0, xx1, yy1))
+
+def load_frames(d, n_sample=8, bbox=None, last_k=0):
+    """"""
+    if not os.path.isdir(d):
+        return None
+    fs = sorted((f for f in os.listdir(d) if f.lower().endswith(('.png', '.jpg', '.jpeg'))))
+    if not fs:
+        return None
+    if last_k > 0:
+        fs = fs[-last_k:]
+    elif n_sample > 0 and len(fs) > n_sample:
+        idx = [int(round(i * (len(fs) - 1) / (n_sample - 1))) for i in range(n_sample)]
+        fs = [fs[i] for i in idx]
+    imgs = [Image.open(os.path.join(d, f)).convert('RGB') for f in fs]
+    if bbox is not None:
+        imgs = [crop_by_bbox(im, bbox) for im in imgs]
+    return imgs
+
+def split_sents(t):
+    t = re.sub('===EXO===.*?===EGO===', '', t or '', flags=re.S)
+    s = [x.strip() for x in re.split('(?<=[.!?])\\s+|\\n+', t) if x.strip()]
+    return s or ([t] if t else [])
+
+def net_new_text(full_text, blind_text, sim_lo, sim_hi):
+    """"""
+    sys.path.insert(0, HERE)
+    import build_student_sft as B
+    blind_idx = B.build_blind_index(blind_text)
+    out, cur = ([], None)
+    for blk in B.tile_sentences(full_text):
+        lab = B._label_of(blk)
+        if lab:
+            cur = lab
+        if B.classify_weight(blk, cur, blind_idx, sim_lo, sim_hi) == 'netnew':
+            out.append(blk)
+    return ''.join(out).strip()
+
+@torch.no_grad()
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--real', required=True)
+    ap.add_argument('--shuffle', required=True)
+    ap.add_argument('--empty', default=None, help='empty(旁路关)预测目录, 只用于地板 sim(empty); 省略则不算地板, headline Δ=real-shuffle 不受影响(向后兼容: 老调用照常传)')
+    ap.add_argument('--manifest', required=True)
+    ap.add_argument('--caption_json', required=True)
+    ap.add_argument('--mask_root', default=seg_mask_root())
+    ap.add_argument('--n_frames', type=int, default=8)
+    ap.add_argument('--crop_mask', action='store_true', help='裁 ego HOI mask 并集 bbox 再送 CLIP(默认关闭=整张画面, 向后兼容)')
+    ap.add_argument('--pad_frac', type=float, default=0.15)
+    ap.add_argument('--diff_json', default=None, help='{uid: blind_caption} json, 用净新句作 CLIP 文本目标(默认关闭=整段caption)')
+    ap.add_argument('--sim_lo', type=float, default=0.3)
+    ap.add_argument('--sim_hi', type=float, default=0.5)
+    ap.add_argument('--last_k', type=int, default=0, help='只用末尾 last_k 帧(默认0=关闭,整段均匀采样,向后兼容); 首帧是条件图像本身跟 caption 无关,均匀采样会稀释后段 reveal 信号')
+    args = ap.parse_args()
+    os.environ['HF_HUB_OFFLINE'] = '1'
+    os.environ['TRANSFORMERS_OFFLINE'] = '1'
+    from transformers import CLIPModel, CLIPProcessor
+    dev = 'cuda' if torch.cuda.is_available() else 'cpu'
+    model = CLIPModel.from_pretrained(CLIP_DIR).to(dev).eval()
+    proc = CLIPProcessor.from_pretrained(CLIP_DIR)
+
+    def embed_text(text):
+        sents = split_sents(text)[:24]
+        if not sents:
+            return None
+        tok = proc(text=sents, return_tensors='pt', padding=True, truncation=True, max_length=77).to(dev)
+        e = model.get_text_features(**tok)
+        e = e / e.norm(dim=-1, keepdim=True)
+        return e.mean(0)
+
+    def embed_imgs(imgs):
+        px = proc(images=imgs, return_tensors='pt').to(dev)
+        e = model.get_image_features(**px)
+        e = e / e.norm(dim=-1, keepdim=True)
+        return e.mean(0)
+    man = json.load(open(args.manifest))
+    caps = json.load(open(args.caption_json))
+    blind = json.load(open(args.diff_json)) if args.diff_json else None
+    rows = []
+    tag = 'crop_mask=%s diff_json=%s last_k=%d' % (args.crop_mask, bool(args.diff_json), args.last_k)
+    print('>> %s' % tag)
+    print(''.join([format('clip', ''.join(['22s'])), ' ', format('simReal', ''.join(['>8s'])), ' ', format('simShuf', ''.join(['>8s'])), ' ', format('simEmp', ''.join(['>8s'])), ' ', format('Δ=R-S', ''.join(['>8s']))]))
+    for s in man:
+        u = s['take_uid']
+        cap = caps.get(u)
+        if cap is None:
+            continue
+        text = cap
+        if blind is not None:
+            btext = blind.get(u)
+            if btext is None:
+                if u not in _warned_diff_miss:
+                    print(''.join(['  [警告] ', format(u[:12], ''), ' 无 blind 覆盖, 退回整段 caption 作文本目标']))
+                    _warned_diff_miss.add(u)
+            else:
+                nn = net_new_text(cap, btext, args.sim_lo, args.sim_hi)
+                if nn:
+                    text = nn
+        te = embed_text(text)
+        bbox = mask_union_bbox(os.path.join(args.mask_root, u, 'ego'), args.pad_frac) if args.crop_mask else None
+        fr = load_frames(os.path.join(args.real, u), args.n_frames, bbox, args.last_k)
+        fsh = load_frames(os.path.join(args.shuffle, u), args.n_frames, bbox, args.last_k)
+        fe = load_frames(os.path.join(args.empty, u), args.n_frames, bbox, args.last_k) if args.empty else None
+        if te is None or fr is None or fsh is None or (args.empty and fe is None):
+            continue
+        sr = float(embed_imgs(fr) @ te)
+        ss = float(embed_imgs(fsh) @ te)
+        se = float(embed_imgs(fe) @ te) if fe is not None else float('nan')
+        rows.append((sr, ss, se))
+        print(''.join([format(u[:22], ''.join(['22s'])), ' ', format(sr, ''.join(['8.4f'])), ' ', format(ss, ''.join(['8.4f'])), ' ', format(se, ''.join(['8.4f'])), ' ', format(sr - ss, ''.join(['+8.4f']))]))
+    if not rows:
+        print('!! 没有可用样本(caption/mask/帧目录须同时存在)')
+        return
+    a = np.array(rows)
+    mr, ms, me = (a[:, 0].mean(), a[:, 1].mean(), np.nanmean(a[:, 2]))
+    d = a[:, 0] - a[:, 1]
+    win = int((d > 0).sum())
+    tot = len(d)
+    tval = d.mean() / (d.std(ddof=1) / np.sqrt(tot) + 1e-09)
+    print('\n===== CLIP 方向性汇总 [%s] (n=%d) =====' % (tag, tot))
+    print(''.join(['  sim(real,目标)    = ', format(mr, ''.join(['.4f']))]))
+    print(''.join(['  sim(shuffle,目标) = ', format(ms, ''.join(['.4f']))]))
+    if np.isfinite(me):
+        print(''.join(['  sim(empty,目标)   = ', format(me, ''.join(['.4f']))]))
+    print(''.join(['  Δ = sim(real)-sim(shuffle) = ', format(d.mean(), ''.join(['+.4f'])), '  (real 更贴目标的 clip 数: ', format(win, ''), '/', format(tot, ''), ', t≈', format(tval, ''.join(['.2f'])), ')']))
+    print('\n===== 判读 =====')
+    if d.mean() > 0 and (win >= tot * 0.65 or tval > 1.8):
+        print('  => real 显著比 shuffle 更贴目标: 分支【响应正确】, caption 有用, SSIM 冤枉了它。')
+    elif d.mean() > 0:
+        print('  => real 略高于 shuffle 但不显著: 有微弱正确响应, 证据不足, 需扩大样本/更强指标。')
+    else:
+        print('  => real 不比 shuffle 贴目标: 分支响应方向随机 = caption 确实没被有效利用。')
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/eval_masked_grpo3800_3wan.py

@@ -0,0 +1,90 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+from project_paths import seg_mask_root, seg_mask_dir
+import argparse
+import json
+import os
+import sys
+import numpy as np
+HERE = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, HERE)
+BASELINE_EVAL = _project_path('eval')
+sys.path.insert(0, BASELINE_EVAL)
+from eval_masked_region_metrics import load_seq, mask_union_bbox, crop_seq, resize_seq
+from metrics import psnr_ssim, lpips_score
+ROOT = os.path.join(HERE, '..')
+VARIANTS = ['base', 'global_raw', 'global_clean', 'sroute_raw', 'sroute_clean']
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--manifest', default=os.path.join(ROOT, 'manifests/eval.json'))
+    ap.add_argument('--mask_root', default=seg_mask_root())
+    ap.add_argument('--pred_prefix', default=os.path.join(ROOT, 'outputs/pred_grpo3800_'))
+    ap.add_argument('--variants', nargs='+', default=VARIANTS)
+    ap.add_argument('--pad_frac', type=float, default=0.15)
+    ap.add_argument('--crop_size', type=int, default=256)
+    ap.add_argument('--device', default='cuda')
+    ap.add_argument('--out_json', default=os.path.join(ROOT, 'outputs/gen_grpo3800_3wan/masked_metrics_5variants.json'))
+    ap.add_argument('--limit', type=int, default=-1)
+    args = ap.parse_args()
+    manifest = json.load(open(args.manifest))
+    if args.limit > 0:
+        manifest = manifest[:args.limit]
+    dirs = {v: args.pred_prefix + v for v in args.variants}
+    agg = {v: {'global': {'psnr': [], 'ssim': [], 'lpips': []}, 'local': {'psnr': [], 'ssim': [], 'lpips': []}} for v in args.variants}
+    n_used = 0
+    for s in manifest:
+        uid = s['take_uid']
+        gt_dir = s.get('gt_dir')
+        bbox = mask_union_bbox(os.path.join(args.mask_root, uid, 'ego'), args.pad_frac)
+        gt = load_seq(gt_dir) if gt_dir else None
+        if bbox is None or gt is None:
+            continue
+        preds = {v: load_seq(os.path.join(dirs[v], uid)) for v in args.variants}
+        if any((p is None for p in preds.values())):
+            continue
+        nf = min([len(gt)] + [len(p) for p in preds.values()])
+        if nf < 1:
+            continue
+        gt_seq = gt[:nf]
+        g_gt = resize_seq(gt_seq, args.crop_size)
+        l_gt = crop_seq(gt_seq, bbox, args.crop_size)
+        for v, frames in preds.items():
+            pf = frames[:nf]
+            g_pred = resize_seq(pf, args.crop_size)
+            l_pred = crop_seq(pf, bbox, args.crop_size)
+            gp, gs = psnr_ssim(g_pred, g_gt)
+            lp, ls = psnr_ssim(l_pred, l_gt)
+            gl = lpips_score(g_pred, g_gt, device=args.device)
+            ll = lpips_score(l_pred, l_gt, device=args.device)
+            agg[v]['global']['psnr'].append(gp)
+            agg[v]['global']['ssim'].append(gs)
+            agg[v]['global']['lpips'].append(gl)
+            agg[v]['local']['psnr'].append(lp)
+            agg[v]['local']['ssim'].append(ls)
+            agg[v]['local']['lpips'].append(ll)
+        n_used += 1
+        if n_used % 200 == 0:
+            print(''.join(['  ... ', format(n_used, ''), ' clip']), flush=True)
+    if n_used == 0:
+        print('!! 无可用样本(须 GT/mask/5变体 pred 同时存在)')
+        return
+
+    def mean(xs):
+        return float(np.mean(xs)) if xs else float('nan')
+    summary = {v: {sc: {k: mean(x) for k, x in d.items()} for sc, d in agg[v].items()} for v in args.variants}
+    print(''.join(['\n===== masked 指标汇总 (n=', format(n_used, ''), ', 5 变体同分母) =====']))
+    print(''.join([format('variant', ''.join(['14s'])), ' ', format('scope', ''.join(['7s'])), ' ', format('SSIM', ''.join(['>8s'])), ' ', format('PSNR', ''.join(['>8s'])), ' ', format('LPIPS', ''.join(['>8s']))]))
+    for v in args.variants:
+        for sc in ('global', 'local'):
+            d = summary[v][sc]
+            print(''.join([format(v, ''.join(['14s'])), ' ', format(sc, ''.join(['7s'])), ' ', format(d['ssim'], ''.join(['8.4f'])), ' ', format(d['psnr'], ''.join(['8.3f'])), ' ', format(d['lpips'], ''.join(['8.4f']))]))
+    json.dump({'summary': summary, 'n': n_used, 'variants': args.variants, 'pad_frac': args.pad_frac, 'crop_size': args.crop_size, 'dirs': dirs}, open(args.out_json, 'w'), ensure_ascii=False, indent=2)
+    print(''.join(['\n已存: ', format(args.out_json, '')]))
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/eval_masked_region_metrics.py

@@ -0,0 +1,153 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+from project_paths import seg_mask_root, seg_mask_dir
+import argparse
+import json
+import os
+import sys
+import numpy as np
+from PIL import Image
+BASELINE_EVAL = _project_path('eval')
+
+def load_seq(d):
+    if not os.path.isdir(d):
+        return None
+    fs = sorted((f for f in os.listdir(d) if f.lower().endswith('.png')))
+    if not fs:
+        return None
+    return [np.asarray(Image.open(os.path.join(d, f)).convert('RGB')) for f in fs]
+
+def mask_union_bbox(mask_dir, pad_frac):
+    """"""
+    if not os.path.isdir(mask_dir):
+        return None
+    fs = sorted((f for f in os.listdir(mask_dir) if f.endswith('.png')))
+    if not fs:
+        return None
+    H = W = None
+    y0 = y1 = x0 = x1 = None
+    for f in fs:
+        m = np.asarray(Image.open(os.path.join(mask_dir, f)))
+        if m.ndim == 3:
+            m = m[..., 0]
+        H, W = m.shape
+        ys, xs = np.where(m > 0)
+        if len(ys) == 0:
+            continue
+        fy0, fy1, fx0, fx1 = (int(ys.min()), int(ys.max()), int(xs.min()), int(xs.max()))
+        y0 = fy0 if y0 is None else min(y0, fy0)
+        y1 = fy1 if y1 is None else max(y1, fy1)
+        x0 = fx0 if x0 is None else min(x0, fx0)
+        x1 = fx1 if x1 is None else max(x1, fx1)
+    if y0 is None:
+        return None
+    h, w = (y1 - y0 + 1, x1 - x0 + 1)
+    py, px = (int(h * pad_frac), int(w * pad_frac))
+    y0, y1 = (max(0, y0 - py), min(H - 1, y1 + py))
+    x0, x1 = (max(0, x0 - px), min(W - 1, x1 + px))
+    return (y0, y1, x0, x1, H, W)
+
+def crop_seq(frames, bbox, size):
+    """"""
+    y0, y1, x0, x1, H, W = bbox
+    out = []
+    for fr in frames:
+        Hf, Wf = fr.shape[:2]
+        sy, sx = (Hf / H, Wf / W)
+        yy0, yy1 = (int(y0 * sy), min(Hf, int(y1 * sy) + 1))
+        xx0, xx1 = (int(x0 * sx), min(Wf, int(x1 * sx) + 1))
+        crop = fr[yy0:yy1, xx0:xx1]
+        out.append(np.asarray(Image.fromarray(crop).resize((size, size))))
+    return np.stack(out)
+
+def resize_seq(frames, size):
+    return np.stack([np.asarray(Image.fromarray(fr).resize((size, size))) for fr in frames])
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--real', required=True)
+    ap.add_argument('--shuffle', required=True)
+    ap.add_argument('--empty', required=True)
+    ap.add_argument('--manifest', default=_project_path('manifests/eval.json'))
+    ap.add_argument('--mask_root', default=seg_mask_root())
+    ap.add_argument('--pad_frac', type=float, default=0.15)
+    ap.add_argument('--crop_size', type=int, default=256)
+    ap.add_argument('--device', default='cuda')
+    ap.add_argument('--out_json', default=None)
+    ap.add_argument('--limit', type=int, default=-1)
+    args = ap.parse_args()
+    sys.path.insert(0, BASELINE_EVAL)
+    from metrics import psnr_ssim, lpips_score
+    manifest = json.load(open(args.manifest))
+    if args.limit > 0:
+        manifest = manifest[:args.limit]
+    modes = {'real': args.real, 'shuffle': args.shuffle, 'empty': args.empty}
+    agg = {m: {'global': {'psnr': [], 'ssim': [], 'lpips': []}, 'local': {'psnr': [], 'ssim': [], 'lpips': []}} for m in modes}
+    per_clip = []
+    n_used = 0
+    for s in manifest:
+        uid = s['take_uid']
+        gt_dir = s.get('gt_dir')
+        mask_dir = os.path.join(args.mask_root, uid, 'ego')
+        bbox = mask_union_bbox(mask_dir, args.pad_frac)
+        gt = load_seq(gt_dir) if gt_dir else None
+        if bbox is None or gt is None:
+            continue
+        preds = {m: load_seq(os.path.join(d, uid)) for m, d in modes.items()}
+        if any((v is None for v in preds.values())):
+            continue
+        nf = min([len(gt)] + [len(v) for v in preds.values()])
+        if nf < 1:
+            continue
+        gt_seq = gt[:nf]
+        clip_row = {'take_uid': uid, 'take_name': s.get('take_name')}
+        for m, frames in preds.items():
+            pf = frames[:nf]
+            g_pred = resize_seq(pf, args.crop_size)
+            g_gt = resize_seq(gt_seq, args.crop_size)
+            l_pred = crop_seq(pf, bbox, args.crop_size)
+            l_gt = crop_seq(gt_seq, bbox, args.crop_size)
+            gp, gs = psnr_ssim(g_pred, g_gt)
+            lp, ls = psnr_ssim(l_pred, l_gt)
+            gl = lpips_score(g_pred, g_gt, device=args.device)
+            ll = lpips_score(l_pred, l_gt, device=args.device)
+            agg[m]['global']['psnr'].append(gp)
+            agg[m]['global']['ssim'].append(gs)
+            agg[m]['global']['lpips'].append(gl)
+            agg[m]['local']['psnr'].append(lp)
+            agg[m]['local']['ssim'].append(ls)
+            agg[m]['local']['lpips'].append(ll)
+            clip_row[m] = {'global': {'psnr': gp, 'ssim': gs, 'lpips': gl}, 'local': {'psnr': lp, 'ssim': ls, 'lpips': ll}}
+        per_clip.append(clip_row)
+        n_used += 1
+        print(''.join(['[', format(n_used, ''), '] ', format(s.get('take_name'), ''), ': ']) + ' | '.join((''.join([format(m, ''), ' local_ssim=', format(clip_row[m]['local']['ssim'], ''.join(['.4f']))]) for m in modes)))
+    if n_used == 0:
+        print('!! 没有可用样本(GT/mask/real/shuffle/empty 目录须同时存在)')
+        return
+
+    def mean(xs):
+        return float(np.mean(xs)) if xs else float('nan')
+    summary = {m: {scope: {k: mean(v) for k, v in d.items()} for scope, d in agg[m].items()} for m in modes}
+    print(''.join(['\n===== 全局 vs 局部(HOI mask区域) SSIM/PSNR/LPIPS 汇总 (n=', format(n_used, ''), ') =====']))
+    print(''.join([format('mode', ''.join(['8s'])), ' ', format('scope', ''.join(['7s'])), ' ', format('SSIM', ''.join(['>8s'])), ' ', format('PSNR', ''.join(['>8s'])), ' ', format('LPIPS', ''.join(['>8s']))]))
+    for m in modes:
+        for scope in ('global', 'local'):
+            d2 = summary[m][scope]
+            print(''.join([format(m, ''.join(['8s'])), ' ', format(scope, ''.join(['7s'])), ' ', format(d2['ssim'], ''.join(['8.4f'])), ' ', format(d2['psnr'], ''.join(['8.3f'])), ' ', format(d2['lpips'], ''.join(['8.4f']))]))
+    d_local = summary['real']['local']['ssim'] - summary['shuffle']['local']['ssim']
+    d_global = summary['real']['global']['ssim'] - summary['shuffle']['global']['ssim']
+    print(''.join(['\nΔSSIM(real-shuffle): 全局=', format(d_global, ''.join(['+.4f'])), '  局部(mask区域)=', format(d_local, ''.join(['+.4f']))]))
+    if d_local > d_global + 0.0001:
+        print("=> 局部区域的 real-shuffle 差距比全局更大: 支持'语义增益被全局指标稀释'的假设。")
+    else:
+        print('=> 局部区域差距不比全局大: mask 区域内也看不出 real 优于 shuffle, 语义增益证据仍不足。')
+    if args.out_json:
+        json.dump({'summary': summary, 'per_clip': per_clip, 'n': n_used, 'real': args.real, 'shuffle': args.shuffle, 'empty': args.empty, 'pad_frac': args.pad_frac, 'crop_size': args.crop_size}, open(args.out_json, 'w'), ensure_ascii=False, indent=2)
+        print(''.join(['\n已存: ', format(args.out_json, '')]))
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/eval_opensource_baselines.py

@@ -0,0 +1,90 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+import os, sys, json, argparse
+import numpy as np
+HERE = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, _project_path('eval'))
+sys.path.insert(0, os.path.dirname(HERE))
+sys.path.insert(0, HERE)
+from eval_masked_region_metrics import load_seq, resize_seq
+from metrics import psnr_ssim, lpips_score, fvd_score
+from paper_protocol import PAPER_BASELINES, FRAMES, METRIC_SIZE, validate_manifest, first_frames
+ROOT = os.path.join(HERE, '..')
+OSB = os.path.join(ROOT, 'outputs/opensource_baselines/preds')
+ARMS = {name: os.path.join(OSB, name) for name in PAPER_BASELINES}
+ARMS.update({'ours_base': os.path.join(ROOT, 'outputs/pred_hoi'), 'ours_global': os.path.join(ROOT, 'outputs/pred_global'), 'ours_sroute': os.path.join(ROOT, 'outputs/pred_spatial')})
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--manifest', default=os.path.join(ROOT, 'manifests/eval.json'))
+    ap.add_argument('--crop', type=int, default=256)
+    ap.add_argument('--device', default='cuda')
+    ap.add_argument('--fvd', action='store_true', help='额外算整批 FVD(较慢)')
+    ap.add_argument('--arms', nargs='+', choices=list(ARMS), default=list(PAPER_BASELINES), help='默认只评论文六种 baseline；其它臂须显式指定')
+    ap.add_argument('--out', default=os.path.join(ROOT, 'outputs/opensource_baselines/metrics.json'))
+    args = ap.parse_args()
+    man = validate_manifest(json.load(open(args.manifest)))
+    if args.crop != METRIC_SIZE:
+        ap.error('Paper metrics require 256 x 256 spatial resize.')
+    arms = {k: ARMS[k] for k in args.arms}
+    missing = [k for k, path in arms.items() if not os.path.isdir(path)]
+    if missing:
+        ap.error(''.join(['Missing baseline prediction directories: ', format(missing, '')]))
+    os.makedirs(os.path.dirname(os.path.abspath(args.out)), exist_ok=True)
+    print(''.join(['>> 评测 ', format(len(man), ''), ' clips × ', format(len(arms), ''), ' arms: ', format(list(arms), '')]), flush=True)
+    cache_dir = os.path.join(ROOT, 'outputs/opensource_baselines/_metric_cache')
+    os.makedirs(cache_dir, exist_ok=True)
+    results = {}
+    for name, pdir in arms.items():
+        cache = os.path.join(cache_dir, ''.join([format(name, ''), '.json']))
+        rows = {}
+        gtser, prser = ([], [])
+        n_new = 0
+        for s in man:
+            uid, gt_dir = (s['take_uid'], s.get('gt_dir'))
+            need_fvd = args.fvd
+            if uid in rows and (not need_fvd):
+                continue
+            pred = load_seq(os.path.join(pdir, uid))
+            gt = load_seq(gt_dir) if gt_dir else None
+            pred, gt = first_frames(pred, gt, uid)
+            nf = FRAMES
+            gp = resize_seq(pred[:nf], args.crop)
+            gg = resize_seq(gt[:nf], args.crop)
+            if uid not in rows:
+                ps, ss = psnr_ssim(gp, gg)
+                lp = lpips_score(gp, gg, device=args.device)
+                rows[uid] = {'psnr': ps, 'ssim': ss, 'lpips': lp, 'nf': nf}
+                n_new += 1
+                if n_new % 100 == 0:
+                    print(''.join(['  [', format(name, ''), '] +', format(n_new, ''), ' (累计 ', format(len(rows), ''), ')']), flush=True)
+            if need_fvd:
+                gtser.append(gg)
+                prser.append(gp)
+        json.dump(rows, open(cache, 'w'))
+        agg = {k: float(np.mean([r[k] for r in rows.values()])) for k in ('psnr', 'ssim', 'lpips')}
+        agg['n'] = len(rows)
+        if args.fvd and len(prser) >= 16:
+            T0 = min(min((x.shape[0] for x in prser)), min((x.shape[0] for x in gtser)))
+            pr = np.stack([x[:T0] for x in prser])
+            gt = np.stack([x[:T0] for x in gtser])
+            agg['fvd'] = fvd_score(pr, gt, device=args.device)
+        results[name] = agg
+        print(''.join(['>> ', format(name, ''), ': ', format(agg, '')]), flush=True)
+    json.dump(results, open(args.out, 'w'), ensure_ascii=False, indent=2)
+    print('\n' + '=' * 78)
+    hdr = ''.join([format('arm', ''.join(['16s'])), format('n', ''.join(['>6s'])), format('PSNR', ''.join(['>9s'])), format('SSIM', ''.join(['>9s'])), format('LPIPS', ''.join(['>9s'])), format('FVD', ''.join(['>10s']))])
+    print(hdr)
+    print('-' * 78)
+    for name, a in results.items():
+        fvd = ''.join([format(a['fvd'], ''.join(['10.2f']))]) if 'fvd' in a else ''.join([format('-', ''.join(['>10s']))])
+        print(''.join([format(name, ''.join(['16s'])), format(a['n'], ''.join(['6d'])), format(a['psnr'], ''.join(['9.3f'])), format(a['ssim'], ''.join(['9.4f'])), format(a['lpips'], ''.join(['9.4f'])), format(fvd, '')]))
+    print('=' * 78)
+    print('>> 存', args.out)
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/gen_ego_caption.py

@@ -0,0 +1,222 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+import os
+import sys
+import json
+import glob
+import argparse
+os.environ.setdefault('HF_HUB_OFFLINE', '1')
+os.environ.setdefault('TRANSFORMERS_OFFLINE', '1')
+os.environ.setdefault('HF_HUB_DISABLE_TELEMETRY', '1')
+import torch
+from PIL import Image
+VLM_DIR = _project_path('models/Qwen3.5-27B')
+DATASET = _project_path('./data/hoi_dataset')
+EVAL_MANIFEST = _project_path('manifests/eval.json')
+EGO_INSTRUCTION = "The video shows a person doing a kitchen task, filmed by a fixed third-person camera. IGNORE where the camera is. Put yourself INSIDE that person, looking out through their own eyes (first-person egocentric view: you would mainly see your own hands and whatever is right in front of you on the counter). Think briefly and ONCE about the person's viewpoint — do NOT draft, critique, or rewrite multiple times. Then output your FINAL answer: 2 to 4 concise sentences describing what YOU would see from your own eyes — your hand(s), the object(s) you hold or touch, where they are in your field of view, and the action you are doing. Only describe what is actually visible — do NOT invent objects that are not there. Write plain visual description, no camera talk. Keep your reasoning short so you finish."
+
+def build_model(device):
+    from transformers import AutoProcessor
+    try:
+        from transformers import Qwen3_5ForConditionalGeneration as VLMClass
+        print('加载 Qwen3.5-27B ...', flush=True)
+    except ImportError:
+        from transformers import AutoModelForImageTextToText as VLMClass
+        print('[warn] 无 Qwen3_5 专用类, 回退 AutoModelForImageTextToText', flush=True)
+    model = VLMClass.from_pretrained(VLM_DIR, dtype=torch.bfloat16, device_map=device, attn_implementation='sdpa').eval()
+    processor = AutoProcessor.from_pretrained(VLM_DIR)
+    return (model, processor)
+
+def load_exo_frames_from_dir(fdir, max_frames=8, max_pixels=200704):
+    """"""
+    if not os.path.isdir(fdir):
+        return None
+    files = sorted((f for f in os.listdir(fdir) if f.endswith('.png')))
+    if not files:
+        return None
+    if len(files) > max_frames:
+        idx = [int(i * (len(files) - 1) / (max_frames - 1)) for i in range(max_frames)]
+        files = [files[i] for i in idx]
+    frames = []
+    for f in files:
+        im = Image.open(os.path.join(fdir, f)).convert('RGB')
+        if max_pixels > 0 and im.width * im.height > max_pixels:
+            r = (max_pixels / (im.width * im.height)) ** 0.5
+            im = im.resize((max(1, int(im.width * r)), max(1, int(im.height * r))))
+        frames.append(im)
+    return frames
+
+def load_exo_frames_from_video(video_path, start_sec, num_frames=8, target_fps=8, max_pixels=200704):
+    """"""
+    import decord
+    try:
+        vr = decord.VideoReader(video_path)
+    except Exception:
+        return None
+    fps = vr.get_avg_fps() or 30.0
+    start = int(round(start_sec * fps))
+    step = max(1, int(round(fps / target_fps)))
+    idxs = [min(start + i * step, len(vr) - 1) for i in range(num_frames)]
+    fr = vr.get_batch(idxs).asnumpy()
+    frames = []
+    for f in fr:
+        im = Image.fromarray(f).convert('RGB')
+        if max_pixels > 0 and im.width * im.height > max_pixels:
+            r = (max_pixels / (im.width * im.height)) ** 0.5
+            im = im.resize((max(1, int(im.width * r)), max(1, int(im.height * r))))
+        frames.append(im)
+    return frames
+
+def extract_final_caption(raw):
+    """"""
+    import re
+    if '</think>' in raw:
+        tail = raw.split('</think>')[-1].strip()
+        if tail:
+            return tail
+    markers = list(re.finditer('(?:Final Polish|Revised Draft|Final Draft|Final)\\s*[:\\-]?\\s*', raw, re.I))
+    if markers:
+        seg = raw[markers[-1].end():].strip()
+        seg = seg.strip('*').strip()
+        m = re.search('"([^"]{20,})"', seg)
+        if m:
+            return m.group(1).strip()
+        seg = re.split("\\n\\s*(?:Wait|Let'?s|Critique|Draft)\\b", seg)[0].strip()
+        if len(seg) > 20:
+            return seg
+    quotes = re.findall('"((?:I |From my |My )[^"]{20,})"', raw)
+    if quotes:
+        return quotes[-1].strip()
+    lines = [l.strip(' *"') for l in raw.splitlines() if re.match('\\s*\\*?\\s*\\"?(I |From my |My )', l)]
+    if lines:
+        return lines[-1].strip()
+    return raw.strip()
+
+@torch.no_grad()
+def run_caption(model, processor, frames, device, max_new_tokens=1536):
+    content = [{'type': 'video', 'video': frames}, {'type': 'text', 'text': EGO_INSTRUCTION}]
+    msgs = [{'role': 'user', 'content': content}]
+    text = processor.apply_chat_template(msgs, tokenize=False, add_generation_prompt=True)
+    inputs = processor(text=[text], videos=[frames], return_tensors='pt', padding=True, do_sample_frames=False)
+    inputs = {k: v.to(device) if hasattr(v, 'to') else v for k, v in inputs.items()}
+    out = model.generate(**inputs, max_new_tokens=max_new_tokens, do_sample=False)
+    gen = out[:, inputs['input_ids'].shape[1]:]
+    raw = processor.batch_decode(gen, skip_special_tokens=True)[0].strip()
+    cap = extract_final_caption(raw)
+    return (cap, raw)
+
+def collect_tasks(src, limit):
+    """"""
+    tasks = []
+    if src == 'eval':
+        manifest = json.load(open(EVAL_MANIFEST))
+        if limit > 0:
+            manifest = manifest[:limit]
+        for s in manifest:
+            tasks.append({'key': s['take_uid'], 'kind': 'video', 'video': s['exo_video'], 'start_sec': s['start_sec'], 'narration': s.get('prompt', '')})
+    else:
+        clips_root = os.path.join(DATASET, 'clips')
+        cids = sorted(os.listdir(clips_root))
+        if limit > 0:
+            cids = cids[:limit]
+        for c in cids:
+            fdir = os.path.join(clips_root, c, 'exo_frames')
+            narration = ''
+            mp = os.path.join(clips_root, c, 'meta.json')
+            if os.path.isfile(mp):
+                try:
+                    narration = json.load(open(mp)).get('prompt', '')
+                except Exception:
+                    pass
+            tasks.append({'key': c, 'kind': 'dir', 'dir': fdir, 'narration': narration})
+    return tasks
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--src', choices=['train', 'eval'], default='eval')
+    ap.add_argument('--limit', type=int, default=5, help='取前N条; -1=全量')
+    ap.add_argument('--num_frames', type=int, default=8, help='喂给VLM的exo帧数')
+    ap.add_argument('--num_shards', type=int, default=1)
+    ap.add_argument('--shard', type=int, default=0)
+    ap.add_argument('--shard_tag', default='', help='给输出文件名加标签(多机避免撞名), 如 nodeB -> ego_caption_train_shardnodeB0.json')
+    ap.add_argument('--reverse', action='store_true', help='从任务列表尾部往前提(第二台机用, 与第一台头部相向而行, 靠全局跳过集在中间去重)')
+    ap.add_argument('--out_json', default=None, help='默认 outputs/ego_caption_<src>[_shard].json')
+    ap.add_argument('--show_think', action='store_true', help='验证时把 thinking 推理过程也打印出来(判断视角推理对不对)')
+    args = ap.parse_args()
+    here = os.path.dirname(os.path.abspath(__file__))
+    out_json = args.out_json or os.path.join(here, '..', 'outputs', ''.join(['ego_caption_', format(args.src, '')]) + (''.join(['_shard', format(args.shard_tag, ''), format(args.shard, '')]) if args.num_shards > 1 else '') + '.json')
+    out_json = os.path.abspath(out_json)
+    os.makedirs(os.path.dirname(out_json), exist_ok=True)
+    tasks = collect_tasks(args.src, args.limit)
+    tasks = [t for i, t in enumerate(tasks) if i % args.num_shards == args.shard]
+    if args.reverse:
+        tasks = tasks[::-1]
+    print(''.join(['[', format(args.src, ''), '] shard ', format(args.shard_tag, ''), format(args.shard, ''), '/', format(args.num_shards, ''), ': ', format(len(tasks), ''), ' 条', format(' (反向)' if args.reverse else '', '')]), flush=True)
+    done = {}
+    if os.path.isfile(out_json):
+        try:
+            done = json.load(open(out_json))
+        except Exception:
+            done = {}
+    global_done = set(done.keys())
+    shard_glob = os.path.join(os.path.dirname(out_json), ''.join(['ego_caption_', format(args.src, ''), '_shard*.json']))
+    for sf in glob.glob(shard_glob):
+        try:
+            d = json.load(open(sf))
+            global_done.update((k for k, v in d.items() if v and str(v).strip()))
+        except Exception:
+            pass
+    print(''.join(['[skip] 全局已提 ', format(len(global_done), ''), ' 条(跨所有分片), 本分片将跳过其中命中的']), flush=True)
+    device = 'cuda'
+    model, processor = build_model(device)
+
+    def refresh_global_done():
+        """"""
+        for sf in glob.glob(shard_glob):
+            try:
+                d = json.load(open(sf))
+                global_done.update((k for k, v in d.items() if v and str(v).strip()))
+            except Exception:
+                pass
+
+    def atomic_dump():
+        """"""
+        tmp = out_json + '.tmp'
+        json.dump(done, open(tmp, 'w'), ensure_ascii=False, indent=2)
+        os.replace(tmp, out_json)
+    n_ok = 0
+    for i, t in enumerate(tasks):
+        if t['key'] in done and done[t['key']]:
+            continue
+        if t['key'] in global_done:
+            continue
+        if t['kind'] == 'video':
+            frames = load_exo_frames_from_video(t['video'], t['start_sec'], args.num_frames)
+        else:
+            frames = load_exo_frames_from_dir(t['dir'], args.num_frames)
+        if not frames:
+            print(''.join(['  [skip] 无帧: ', format(t['key'], '')]), flush=True)
+            continue
+        cap, raw = run_caption(model, processor, frames, device)
+        done[t['key']] = cap
+        n_ok += 1
+        if args.src == 'eval' or n_ok <= 5:
+            print(''.join(['\n[', format(i + 1, ''), '/', format(len(tasks), ''), '] ', format(t['key'], '')]))
+            print(''.join(['  narration : ', format(t['narration'], '')]))
+            if args.show_think and '</think>' in raw:
+                think = raw.split('</think>')[0].replace('<think>', '').strip()
+                print(''.join(['  [think]   : ', format(think, '')]))
+            print(''.join(['  ego-cap   : ', format(cap, '')]), flush=True)
+        if n_ok % 20 == 0:
+            atomic_dump()
+            refresh_global_done()
+            print(''.join(['  ... 已存 ', format(n_ok, ''), ' 条 -> ', format(out_json, ''), ' (全局已提 ', format(len(global_done), ''), ')']), flush=True)
+    atomic_dump()
+    print(''.join(['\n完成: ', format(n_ok, ''), ' 条新增, 共 ', format(len(done), ''), ' 条 -> ', format(out_json, '')]))
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/infer_caption.py

@@ -0,0 +1,173 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+from project_paths import seg_mask_root, seg_mask_dir
+import os
+import sys
+import json
+import argparse
+import numpy as np
+import torch
+from PIL import Image
+from model_resources import wan_base_root, wan_model_paths, wan_tokenizer_path
+VACE = wan_base_root()
+DS = _project_path('third_party/DiffSynth-Studio')
+HERE = os.path.dirname(os.path.abspath(__file__))
+EVAL_MANIFEST = _project_path('manifests/eval.json')
+MASK_ROOT = seg_mask_root()
+CAPTION_JSON = _project_path('data/captions/grpo.json')
+PALETTE = np.array([[0, 0, 0], [220, 40, 40], [40, 200, 80]], dtype=np.uint8)
+NEG = '色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，鱼眼畸变，鱼眼镜头，广角畸变，桶形畸变，镜头畸变，边缘拉伸变形，画面弯曲，fisheye distortion，wide-angle distortion，lens distortion，barrel distortion，可见的相机，三脚架，GoPro，运动相机'
+
+def build_pipe(device, lora_path, branch_ckpt, caption_dim=4096):
+    sys.path.insert(0, DS)
+    sys.path.insert(0, HERE)
+    from diffsynth.pipelines.wan_video import WanVideoPipeline, ModelConfig
+    from caption_cross_attn import inject_caption_cross_attn
+    pipe = WanVideoPipeline.from_pretrained(torch_dtype=torch.bfloat16, device=device, model_configs=[ModelConfig(path=path) for path in wan_model_paths(VACE, strict=True)], tokenizer_config=ModelConfig(path=wan_tokenizer_path(VACE, strict=True)))
+    if lora_path:
+        pipe.load_lora(pipe.vace, lora_path, alpha=1.0)
+        print(''.join(['已加载 Stage1 底座 LoRA: ', format(lora_path, '')]))
+    dit = pipe.dit
+    dev = next(dit.parameters()).device
+    dtype = next(dit.parameters()).dtype
+    from safetensors.torch import load_file
+    sd = load_file(branch_ckpt)
+    ckpt_has_route = any(('spatial_route' in k for k in sd.keys()))
+    env_route = os.environ.get('USE_SPATIAL_ROUTE')
+    use_spatial_route = env_route != '0' if env_route is not None else ckpt_has_route
+    inject_caption_cross_attn(dit, caption_dim=caption_dim, vace_module=pipe.vace, use_spatial_route=use_spatial_route)
+    for block in dit.blocks:
+        block.cross_attn._cap_branch.to(device=dev, dtype=dtype)
+    dit_sd = dit.state_dict()
+    n_cap = 0
+    miss = []
+    for k, v in sd.items():
+        kk = k
+        for pre in ('pipe.dit.', 'dit.', 'model.'):
+            if kk.startswith(pre):
+                kk = kk[len(pre):]
+        if kk in dit_sd:
+            dit_sd[kk].copy_(v.to(dit_sd[kk].dtype).to(dit_sd[kk].device))
+            if '_cap_branch' in kk:
+                n_cap += 1
+        else:
+            miss.append(k)
+    print(''.join(['[caption ckpt] 加载 caption ', format(n_cap, ''), '/', format(len(sd), ''), ' 个参数 <- ', format(branch_ckpt, '')]))
+    if miss:
+        print(''.join(['[caption ckpt] 未匹配 ', format(len(miss), ''), ' 个 key, 样本: ', format(miss[:3], '')]))
+    for block in dit.blocks:
+        cb = block.cross_attn._cap_branch
+        cb.gate_cap.data = cb.gate_cap.data.float()
+    print(''.join(['[caption ckpt] block0 gate_cap=', format(dit.blocks[0].cross_attn._cap_branch.gate_cap.item(), ''.join(['+.4e']))]))
+    return pipe
+
+@torch.no_grad()
+def encode_caption(pipe, cap):
+    """"""
+    if cap is None:
+        return None
+    cap = str(cap).strip()
+    if len(cap) == 0:
+        return None
+    pipe.load_models_to_device(('text_encoder',))
+    ids, mask = pipe.tokenizer(cap, return_mask=True, add_special_tokens=True)
+    ids = ids.to(pipe.device)
+    mask = mask.to(pipe.device)
+    seq_lens = mask.gt(0).sum(dim=1).long()
+    feat = pipe.text_encoder(ids, mask)
+    for i, v in enumerate(seq_lens):
+        feat[:, v:] = 0
+    return feat.to(device=pipe.device, dtype=pipe.torch_dtype)
+
+def load_mask_video(mask_dir, size):
+    """"""
+    fs = sorted((f for f in os.listdir(mask_dir) if f.endswith('.png')))
+    out = []
+    for f in fs:
+        m = np.asarray(Image.open(os.path.join(mask_dir, f)))
+        if m.ndim == 3:
+            m = m[..., 0]
+        col = PALETTE[np.clip(m, 0, 2)]
+        out.append(Image.fromarray(col).resize(size))
+    return out
+
+def main():
+    mode_parser = argparse.ArgumentParser(add_help=False)
+    modes = ['global', 'spatial']
+    mode_parser.add_argument('--variant', choices=modes, default='spatial')
+    mode, _ = mode_parser.parse_known_args()
+    branch = 'global' if mode.variant == 'global' else 'spatial'
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--variant', choices=modes, default=mode.variant)
+    ap.add_argument('--lora', default=_project_path('checkpoints/generator_hoi/ckpt.safetensors'), help='Stage 1 HOI 底座，与分支训练的 preset_lora_path 一致')
+    ap.add_argument('--branch_ckpt', default=_project_path(''.join(['checkpoints/generator_', format(branch, ''), '/ckpt.safetensors'])))
+    ap.add_argument('--out', default=_project_path(''.join(['outputs/pred_', format(mode.variant, '')])))
+    ap.add_argument('--manifest', default=EVAL_MANIFEST)
+    ap.add_argument('--caption_json', default=CAPTION_JSON)
+    ap.add_argument('--caption_dim', type=int, default=4096)
+    ap.add_argument('--height', type=int, default=480)
+    ap.add_argument('--width', type=int, default=480)
+    ap.add_argument('--steps', type=int, default=50)
+    ap.add_argument('--cfg_scale', type=float, default=5.0)
+    ap.add_argument('--vace_scale', type=float, default=1.0)
+    ap.add_argument('--fps', type=float, default=8.0)
+    ap.add_argument('--seed', type=int, default=0)
+    ap.add_argument('--limit', type=int, default=-1)
+    ap.add_argument('--mask_root', default=MASK_ROOT, help='外部 seg 预测 mask 根目录：<root>/<clip_id>/ego/*.png')
+    ap.add_argument('--device', default='cuda')
+    args = ap.parse_args()
+    os.environ['USE_SPATIAL_ROUTE'] = '0' if args.variant == 'global' else '1'
+    sys.path.insert(0, HERE)
+    from caption_cross_attn import set_caption_context
+    manifest = json.load(open(args.manifest))
+    if args.limit > 0:
+        manifest = manifest[:args.limit]
+    captions = json.load(open(args.caption_json))
+    out = os.path.abspath(args.out)
+    os.makedirs(out, exist_ok=True)
+    pipe = build_pipe(args.device, args.lora, args.branch_ckpt, args.caption_dim)
+    size = (args.width, args.height)
+    print(''.join(['开始 caption 旁路推理, ', format(len(manifest), ''), ' 条 | cfg=', format(args.cfg_scale, '')]))
+    n_done = 0
+    for k, s in enumerate(manifest):
+        uid = s['take_uid']
+        if os.path.isfile(os.path.join(out, uid, 'pred.mp4')):
+            print(''.join(['  [skip] 已存在: ', format(uid, '')]))
+            n_done += 1
+            continue
+        mask_dir = os.path.join(args.mask_root, uid, 'ego')
+        if not os.path.isdir(mask_dir):
+            print(''.join(['  [skip] 无mask: ', format(s['take_name'], '')]))
+            continue
+        cap = captions.get(uid)
+        if cap is None:
+            print(''.join(['  [skip] 无caption: ', format(s['take_name'], '')]))
+            continue
+        mask_vid = load_mask_video(mask_dir, size)
+        nf = len(mask_vid)
+        ref = Image.open(s['first_frame']).convert('RGB').resize(size)
+        cap_feat = encode_caption(pipe, cap)
+        set_caption_context(pipe.dit, cap_feat)
+        cap_shape = tuple(cap_feat.shape) if cap_feat is not None else 'None(旁路关)'
+        print(''.join(['[', format(k + 1, ''), '/', format(len(manifest), ''), '] ', format(s['take_name'], ''), ' | ', format(nf, ''), '帧 | variant=', format(args.variant, ''), ' | cap', format(cap_shape, ''), " | '", format(s['prompt'][:40], ''), "'"]))
+        video = pipe(prompt=s['prompt'], negative_prompt=NEG, vace_video=mask_vid, vace_reference_image=ref, vace_scale=args.vace_scale, height=args.height, width=args.width, num_frames=nf, num_inference_steps=args.steps, cfg_scale=args.cfg_scale, seed=args.seed, tiled=True)
+        sd = os.path.join(out, uid)
+        os.makedirs(sd, exist_ok=True)
+        for i, fr in enumerate(video):
+            Image.fromarray(np.asarray(fr)).save(os.path.join(sd, ''.join([format(i, ''.join(['03d'])), '.png'])))
+        try:
+            from diffsynth.utils.data import save_video
+            save_video(video, os.path.join(sd, 'pred.mp4'), fps=args.fps, quality=5)
+        except Exception:
+            pass
+        n_done += 1
+        print(''.join(['    -> ', format(sd, '')]))
+    json.dump({'model': 'Wan2.1-VACE-1.3B + HOI-LoRA + caption 旁路', 'lora': args.lora, 'branch_ckpt': args.branch_ckpt, 'caption_json': args.caption_json, 'height': args.height, 'width': args.width, 'rotate': False, 'vace_scale': args.vace_scale, 'steps': args.steps, 'cfg_scale': args.cfg_scale, 'variant': args.variant, 'seed': args.seed, 'num_samples': n_done, 'note': 'ego HOI mask + narration(原生) + ego dense caption(旁路) 引导生成'}, open(os.path.join(out, 'infer_config.json'), 'w'), ensure_ascii=False, indent=2)
+    print(''.join(['完成 ', format(n_done, ''), ' 条 ->']), out)
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/infer_vace_hoi.py

@@ -0,0 +1,97 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+from project_paths import seg_mask_root, seg_mask_dir
+import os
+import sys
+import json
+import argparse
+import numpy as np
+import torch
+from PIL import Image
+HERE = os.path.dirname(os.path.abspath(__file__))
+from model_resources import wan_base_root, wan_model_paths, wan_tokenizer_path
+VACE = wan_base_root()
+DS = _project_path('third_party/DiffSynth-Studio')
+EVAL_MANIFEST = _project_path('manifests/eval.json')
+MASK_ROOT = seg_mask_root()
+PALETTE = np.array([[0, 0, 0], [220, 40, 40], [40, 200, 80]], dtype=np.uint8)
+NEG = '色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景'
+
+def build_pipe(device, lora_path):
+    sys.path.insert(0, DS)
+    sys.path.insert(0, HERE)
+    from diffsynth.pipelines.wan_video import WanVideoPipeline, ModelConfig
+    pipe = WanVideoPipeline.from_pretrained(torch_dtype=torch.bfloat16, device=device, model_configs=[ModelConfig(path=path) for path in wan_model_paths(VACE, strict=True)], tokenizer_config=ModelConfig(path=wan_tokenizer_path(VACE, strict=True)))
+    if lora_path:
+        pipe.load_lora(pipe.vace, lora_path, alpha=1.0)
+        print(''.join(['已加载 LoRA: ', format(lora_path, '')]))
+    return pipe
+
+def load_mask_video(mask_dir, size):
+    """"""
+    fs = sorted((f for f in os.listdir(mask_dir) if f.endswith('.png')))
+    out = []
+    for f in fs:
+        m = np.asarray(Image.open(os.path.join(mask_dir, f)))
+        if m.ndim == 3:
+            m = m[..., 0]
+        col = PALETTE[np.clip(m, 0, 2)]
+        out.append(Image.fromarray(col).resize(size))
+    return out
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--lora', default=_project_path('checkpoints/generator_hoi/ckpt.safetensors'))
+    ap.add_argument('--out', default=_project_path('outputs/pred_hoi'))
+    ap.add_argument('--height', type=int, default=480)
+    ap.add_argument('--width', type=int, default=480)
+    ap.add_argument('--steps', type=int, default=50)
+    ap.add_argument('--cfg_scale', type=float, default=5.0)
+    ap.add_argument('--vace_scale', type=float, default=1.0)
+    ap.add_argument('--fps', type=float, default=8.0)
+    ap.add_argument('--seed', type=int, default=0)
+    ap.add_argument('--limit', type=int, default=-1)
+    ap.add_argument('--manifest', default=EVAL_MANIFEST)
+    ap.add_argument('--mask_root', default=MASK_ROOT, help='外部 seg 预测 mask 根目录：<root>/<clip_id>/ego/*.png')
+    ap.add_argument('--device', default='cuda')
+    args = ap.parse_args()
+    manifest = json.load(open(args.manifest))
+    if args.limit > 0:
+        manifest = manifest[:args.limit]
+    out = os.path.abspath(args.out)
+    os.makedirs(out, exist_ok=True)
+    pipe = build_pipe(args.device, args.lora)
+    size = (args.width, args.height)
+    print(''.join(['开始 M2 推理, ', format(len(manifest), ''), ' 条']))
+    for k, s in enumerate(manifest):
+        if os.path.isfile(os.path.join(out, s['take_uid'], 'pred.mp4')):
+            print(''.join(['  [skip] 已存在: ', format(s['take_uid'], '')]))
+            continue
+        mask_dir = os.path.join(args.mask_root, s['take_uid'], 'ego')
+        if not os.path.isdir(mask_dir):
+            print(''.join(['  [skip] 无mask: ', format(s['take_name'], '')]))
+            continue
+        mask_vid = load_mask_video(mask_dir, size)
+        nf = len(mask_vid)
+        ref = Image.open(s['first_frame']).convert('RGB').resize(size)
+        print(''.join(['[', format(k + 1, ''), '/', format(len(manifest), ''), '] ', format(s['take_name'], ''), ' | ', format(nf, ''), "帧 | '", format(s['prompt'][:40], ''), "'"]))
+        video = pipe(prompt=s['prompt'], negative_prompt=NEG, vace_video=mask_vid, vace_reference_image=ref, vace_scale=args.vace_scale, height=args.height, width=args.width, num_frames=nf, num_inference_steps=args.steps, cfg_scale=args.cfg_scale, seed=args.seed, tiled=True)
+        sd = os.path.join(out, s['take_uid'])
+        os.makedirs(sd, exist_ok=True)
+        for i, fr in enumerate(video):
+            Image.fromarray(np.asarray(fr)).save(os.path.join(sd, ''.join([format(i, ''.join(['03d'])), '.png'])))
+        try:
+            from diffsynth.utils.data import save_video
+            save_video(video, os.path.join(sd, 'pred.mp4'), fps=args.fps, quality=5)
+        except Exception:
+            pass
+        print(''.join(['    -> ', format(sd, '')]))
+    json.dump({'model': 'Wan2.1-VACE-1.3B + HOI-LoRA', 'lora': args.lora, 'height': args.height, 'width': args.width, 'rotate': False, 'vace_scale': args.vace_scale, 'note': 'seg 预测的 ego HOI mask 引导生成'}, open(os.path.join(out, 'infer_config.json'), 'w'), indent=2)
+    print('完成 ->', out)
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/judge_media_vllm.py

@@ -0,0 +1,192 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+import argparse
+import base64
+import io
+import json
+import os
+import re
+LABELS = ['HANDS', 'OBJECT', 'CONTACT', 'SCENE', 'MOTION', 'LIGHTING']
+JUDGE_RUBRIC = 'You are a strict visual grounding judge. You are given a short first-person (egocentric, head-mounted) video showing what a person actually did, and a CANDIDATE caption that claims to describe that same first-person clip using six labels: HANDS, OBJECT, CONTACT, SCENE, MOTION, LIGHTING.\n\nYour job: judge how FAITHFUL the candidate caption is to what is ACTUALLY visible in the video. Do NOT reward fluent writing; reward correctness. If the caption states something that contradicts the video, that label scores low. If it states something that cannot be verified from the video, treat unverifiable specifics as mild errors.\n\nScore EACH label from 1 to 5:\n  5 = fully accurate, matches the video\n  4 = mostly accurate, minor imprecision\n  3 = partially correct, some wrong or unverifiable claims\n  2 = mostly wrong / contradicts the video\n  1 = entirely wrong or absent\n\nAlso judge:\n  rule_zero_violation (true/false): true ONLY if the caption explicitly names the physical recording apparatus -- e.g. a camera, tripod, GoPro, gimbal, a lens, "fisheye", or words like recording/filming/footage. A first-person caption must never name the capture device. IMPORTANT: describing the vantage or where the viewpoint is looking is NOT a violation -- do NOT flag phrases such as "looking down", "overhead", "top-down view", "from above", "a downward view", or "first-person perspective". Those are legitimate scene descriptions of the viewpoint, not references to the capture apparatus. Only an explicit mention of the physical device counts.\n  grammar_ok (true/false): false if the caption has broken/truncated sentences (e.g. ends a clause with "of." with the noun dropped, or a sentence with no subject like "tilts upward").\n  overall (1-5): holistic faithfulness of the whole caption to the video.\n\nOutput STRICT JSON on a single line, no prose, no markdown, exactly these keys:\n{"HANDS":int,"OBJECT":int,"CONTACT":int,"SCENE":int,"MOTION":int,"LIGHTING":int,"rule_zero_violation":bool,"grammar_ok":bool,"overall":int}'
+JUDGE_USER_TAIL = '\n\nCANDIDATE first-person caption to judge:\n<<<\n{caption}\n>>>\n\nThe images above are frames of the real first-person clip in time order. Score the candidate caption against what is actually visible. Output the STRICT JSON only.'
+
+def load_frames_b64(clip_id, pool, n_ego, max_side):
+    """"""
+    from PIL import Image
+    d = os.path.join(pool, 'clips', clip_id, 'ego_frames')
+    if not os.path.isdir(d):
+        return None
+    pngs = sorted((f for f in os.listdir(d) if f.endswith('.png')))
+    if not pngs:
+        return None
+    if n_ego and n_ego < len(pngs):
+        idx = [round(i * (len(pngs) - 1) / (n_ego - 1)) for i in range(n_ego)]
+        pngs = [pngs[i] for i in sorted(set(idx))]
+    out = []
+    for p in pngs:
+        im = Image.open(os.path.join(d, p)).convert('RGB')
+        w, h = im.size
+        s = max(w, h)
+        if s > max_side:
+            r = max_side / s
+            im = im.resize((max(1, round(w * r)), max(1, round(h * r))))
+        buf = io.BytesIO()
+        im.save(buf, format='JPEG', quality=88)
+        out.append(base64.b64encode(buf.getvalue()).decode())
+    return out
+
+def build_messages(frames_b64, caption):
+    content = []
+    for b in frames_b64:
+        content.append({'type': 'image_url', 'image_url': {'url': 'data:image/jpeg;base64,' + b}})
+    content.append({'type': 'text', 'text': JUDGE_USER_TAIL.format(caption=caption)})
+    return [{'role': 'system', 'content': JUDGE_RUBRIC}, {'role': 'user', 'content': content}]
+
+def extract_json(text):
+    m = re.search('\\{.*\\}', text or '', re.S)
+    if not m:
+        return None
+    try:
+        return json.loads(m.group(0))
+    except Exception:
+        return None
+
+def parse_scores(d):
+    if not d:
+        return None
+    try:
+        labs = [float(d[k]) for k in LABELS]
+    except Exception:
+        return None
+    overall = float(d.get('overall', sum(labs) / len(labs)))
+    rz = bool(d.get('rule_zero_violation', False))
+    gram = bool(d.get('grammar_ok', True))
+    label_mean = sum(labs) / len(labs)
+    comp = overall - (1.0 if rz else 0.0) - (0.3 if not gram else 0.0)
+    row = {'composite': round(comp, 3), 'overall': overall, 'label_mean': round(label_mean, 3), 'rule_zero_violation': int(rz), 'grammar_ok': int(gram)}
+    for i, k in enumerate(LABELS):
+        row[k.lower()] = int(labs[i])
+    return row
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--eval', required=True, help='含 _clip_id 的 eval jsonl(定 clip 集合)')
+    ap.add_argument('--pool', required=True, help='hoi_dataset_testpool_9k 根(clips/<id>/ego_frames)')
+    ap.add_argument('--cap', action='append', required=True, help='name=path.json, 可多个')
+    ap.add_argument('--model', default=_project_path('models/Qwen3.5-27B'))
+    ap.add_argument('--out_dir', required=True)
+    ap.add_argument('--tp', type=int, default=2)
+    ap.add_argument('--gpu_mem_util', type=float, default=0.5)
+    ap.add_argument('--max_model_len', type=int, default=8192)
+    ap.add_argument('--n_ego', type=int, default=13)
+    ap.add_argument('--max_side', type=int, default=420)
+    ap.add_argument('--max_new_tokens', type=int, default=512)
+    ap.add_argument('--batch', type=int, default=48, help='每次 llm.chat 的对话数')
+    ap.add_argument('--limit', type=int, default=-1)
+    args = ap.parse_args()
+    os.makedirs(args.out_dir, exist_ok=True)
+    clips = [json.loads(l)['_clip_id'] for l in open(args.eval) if l.strip()]
+    if args.limit > 0:
+        clips = clips[:args.limit]
+    caps = {}
+    for spec in args.cap:
+        name, path = spec.split('=', 1)
+        caps[name] = json.load(open(path))
+    print(''.join(['>> clips=', format(len(clips), ''), '  models=', format(list(caps), '')]), flush=True)
+    vpath = os.path.join(args.out_dir, 'verdicts.jsonl')
+    done = set()
+    if os.path.exists(vpath):
+        for l in open(vpath):
+            try:
+                r = json.loads(l)
+                done.add((r['model'], r['clip_id']))
+            except Exception:
+                pass
+    print(''.join(['>> 已完成 ', format(len(done), ''), ' 条, 断点续跑']), flush=True)
+    tasks = [(m, c) for m in caps for c in clips if (m, c) not in done and caps[m].get(c)]
+    print(''.join(['>> 待判 ', format(len(tasks), ''), ' 条']), flush=True)
+    if not tasks:
+        aggregate(vpath, list(caps), args.out_dir)
+        return
+    from vllm import LLM, SamplingParams
+    print('>> 加载 vLLM %s (tp=%d, gpu_mem=%.2f) ...' % (args.model, args.tp, args.gpu_mem_util), flush=True)
+    llm = LLM(model=args.model, trust_remote_code=True, tensor_parallel_size=args.tp, gpu_memory_utilization=args.gpu_mem_util, max_model_len=args.max_model_len, limit_mm_per_prompt={'image': max(args.n_ego, 16)}, dtype='bfloat16', enforce_eager=False)
+    sp = SamplingParams(temperature=0.0, max_tokens=args.max_new_tokens)
+    frame_cache = {}
+
+    def get_frames(cid):
+        if cid not in frame_cache:
+            frame_cache[cid] = load_frames_b64(cid, args.pool, args.n_ego, args.max_side)
+        return frame_cache[cid]
+    fo = open(vpath, 'a')
+    n_ok = n_bad = 0
+    for s in range(0, len(tasks), args.batch):
+        chunk = tasks[s:s + args.batch]
+        convs, meta = ([], [])
+        for m, c in chunk:
+            fr = get_frames(c)
+            if not fr:
+                n_bad += 1
+                continue
+            convs.append(build_messages(fr, caps[m][c]))
+            meta.append((m, c))
+        if not convs:
+            continue
+        outs = llm.chat(convs, sampling_params=sp, chat_template_kwargs={'enable_thinking': False})
+        for (m, c), o in zip(meta, outs):
+            txt = o.outputs[0].text if o.outputs else ''
+            row = parse_scores(extract_json(txt))
+            if row is None:
+                n_bad += 1
+                rec = {'model': m, 'clip_id': c, 'ok': False, 'raw': (txt or '')[:200]}
+            else:
+                n_ok += 1
+                rec = {'model': m, 'clip_id': c, 'ok': True, **row}
+            fo.write(json.dumps(rec, ensure_ascii=False) + '\n')
+        fo.flush()
+        print(''.join(['  [', format(min(s + args.batch, len(tasks)), ''), '/', format(len(tasks), ''), '] ok=', format(n_ok, ''), ' bad=', format(n_bad, '')]), flush=True)
+    fo.close()
+    aggregate(vpath, list(caps), args.out_dir)
+
+def aggregate(vpath, model_names, out_dir):
+    import csv
+    import statistics as st
+    rows_by_model = {m: [] for m in model_names}
+    for l in open(vpath):
+        try:
+            r = json.loads(l)
+        except Exception:
+            continue
+        if r.get('ok') and r.get('model') in rows_by_model:
+            rows_by_model[r['model']].append(r)
+    summary = []
+    for m in model_names:
+        rs = rows_by_model[m]
+        if not rs:
+            continue
+        rs_sorted = sorted(rs, key=lambda x: (-x['composite'], -x['overall']))
+        cols = ['clip_id', 'composite', 'overall', 'label_mean'] + [k.lower() for k in LABELS] + ['rule_zero_violation', 'grammar_ok']
+        with open(os.path.join(out_dir, ''.join(['ranked_', format(m, ''), '.csv'])), 'w', newline='') as f:
+            w = csv.DictWriter(f, fieldnames=cols, extrasaction='ignore')
+            w.writeheader()
+            w.writerows(rs_sorted)
+        summary.append({'model': m, 'n': len(rs), 'composite': round(st.mean((x['composite'] for x in rs)), 3), 'overall': round(st.mean((x['overall'] for x in rs)), 3), 'label_mean': round(st.mean((x['label_mean'] for x in rs)), 3), 'rule0_viol_rate': round(st.mean((x['rule_zero_violation'] for x in rs)), 3), 'grammar_ok_rate': round(st.mean((x['grammar_ok'] for x in rs)), 3), **{k.lower(): round(st.mean((x[k.lower()] for x in rs)), 3) for k in LABELS}})
+    summary.sort(key=lambda d: -d['composite'])
+    json.dump(summary, open(os.path.join(out_dir, 'summary.json'), 'w'), indent=2, ensure_ascii=False)
+    cols = ['model', 'n', 'composite', 'overall', 'label_mean', 'rule0_viol_rate', 'grammar_ok_rate'] + [k.lower() for k in LABELS]
+    with open(os.path.join(out_dir, 'summary.csv'), 'w', newline='') as f:
+        w = csv.DictWriter(f, fieldnames=cols)
+        w.writeheader()
+        w.writerows(summary)
+    print('\n===== 媒体裁判(Qwen3.5-27B 看 ego 视频)汇总 =====', flush=True)
+    print(''.join([format('model', ''.join(['12s'])), ' ', format('n', ''.join(['>4'])), ' ', format('comp', ''.join(['>6'])), ' ', format('overall', ''.join(['>7'])), ' ', format('labels', ''.join(['>6'])), ' ', format('rule0%', ''.join(['>6'])), ' ', format('gram%', ''.join(['>6']))]), flush=True)
+    for d in summary:
+        print(''.join([format(d['model'], ''.join(['12s'])), ' ', format(d['n'], ''.join(['>4'])), ' ', format(d['composite'], ''.join(['>6.3f'])), ' ', format(d['overall'], ''.join(['>7.3f'])), ' ', format(d['label_mean'], ''.join(['>6.3f'])), ' ', format(d['rule0_viol_rate'] * 100, ''.join(['>5.1f'])), '% ', format(d['grammar_ok_rate'] * 100, ''.join(['>5.1f'])), '%']), flush=True)
+    print('>> 写出 %s/summary.{json,csv} + ranked_*.csv' % out_dir, flush=True)
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/per_clip_3metric_3variant.py

@@ -0,0 +1,112 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+import argparse
+import json
+import os
+import sys
+import numpy as np
+HERE = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, HERE)
+sys.path.insert(0, _project_path('eval'))
+from eval_masked_region_metrics import load_seq, resize_seq
+from metrics import psnr_ssim, lpips_score
+ROOT = os.path.join(HERE, '..')
+VARS = ('base', 'global', 'sroute')
+
+def clip_metrics(pred_dir, uid, g_gt, nf_gt, crop, device):
+    d = os.path.join(pred_dir, uid)
+    if not os.path.isfile(os.path.join(d, 'pred.mp4')):
+        return None
+    pred = load_seq(d)
+    if pred is None:
+        return None
+    nf = min(len(pred), nf_gt)
+    if nf < 1:
+        return None
+    g_pred = resize_seq(pred[:nf], crop)
+    gg = g_gt[:nf]
+    psnr, ssim = psnr_ssim(g_pred, gg)
+    lp = lpips_score(g_pred, gg, device=device)
+    return {'psnr': psnr, 'ssim': ssim, 'lpips': lp}
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--manifest', default=os.path.join(ROOT, 'manifests/eval.json'))
+    ap.add_argument('--base', default=os.path.join(ROOT, 'outputs/pred_hoi'))
+    ap.add_argument('--global_dir', default=os.path.join(ROOT, 'outputs/pred_global'))
+    ap.add_argument('--sroute', default=os.path.join(ROOT, 'outputs/pred_spatial'))
+    ap.add_argument('--crop_size', type=int, default=256)
+    ap.add_argument('--device', default='cuda')
+    ap.add_argument('--topn', type=int, default=1000)
+    ap.add_argument('--nshards', type=int, default=1)
+    ap.add_argument('--shard', type=int, default=-1)
+    ap.add_argument('--rows_dir', default=os.path.join(ROOT, 'outputs/gen_grpo3800_3wan/metric3_rows'))
+    ap.add_argument('--out_json', default=os.path.join(ROOT, 'outputs/gen_grpo3800_3wan/rank_3metric_top1000.json'))
+    ap.add_argument('--merge', action='store_true')
+    args = ap.parse_args()
+    os.makedirs(args.rows_dir, exist_ok=True)
+    dirs = {'base': args.base, 'global': args.global_dir, 'sroute': args.sroute}
+    if args.merge:
+        _finalize(args)
+        return
+    manifest = json.load(open(args.manifest))
+    if args.shard >= 0:
+        manifest = [m for i, m in enumerate(manifest) if i % args.nshards == args.shard]
+    rows = []
+    n = 0
+    for s in manifest:
+        uid, gt_dir = (s['take_uid'], s.get('gt_dir'))
+        gt = load_seq(gt_dir) if gt_dir else None
+        if gt is None:
+            continue
+        nf_gt = len(gt)
+        g_gt = resize_seq(gt, args.crop_size)
+        mv = {v: clip_metrics(dirs[v], uid, g_gt, nf_gt, args.crop_size, args.device) for v in VARS}
+        if any((m is None for m in mv.values())):
+            continue
+        rows.append({'take_uid': uid, **{v: mv[v] for v in VARS}})
+        n += 1
+        if n % 100 == 0:
+            print(''.join(['  [shard ', format(args.shard, ''), '] 三方齐全已算 ', format(n, ''), ' 片']), flush=True)
+    rf = os.path.join(args.rows_dir, ''.join(['rows_', format(args.shard, ''), '.json']))
+    json.dump(rows, open(rf, 'w'), ensure_ascii=False)
+    print(''.join(['[shard ', format(args.shard, ''), '] ', format(n, ''), ' 片 -> ', format(rf, '')]))
+
+def _finalize(args):
+    rows = []
+    for fn in sorted(os.listdir(args.rows_dir)):
+        if fn.startswith('rows_') and fn.endswith('.json'):
+            rows.extend(json.load(open(os.path.join(args.rows_dir, fn))))
+    n_common = len(rows)
+    ordered = [r for r in rows if r['sroute']['ssim'] > r['global']['ssim'] > r['base']['ssim']]
+    ordered.sort(key=lambda r: r['sroute']['ssim'] - r['base']['ssim'], reverse=True)
+    top = ordered[:args.topn]
+
+    def means(subset):
+        out = {}
+        for v in VARS:
+            out[v] = {k: float(np.mean([r[v][k] for r in subset])) for k in ('psnr', 'ssim', 'lpips')}
+        return out
+    res = {'n_common_3way': n_common, 'n_order_sroute_gt_global_gt_base(by_ssim)': len(ordered), 'topn_returned': len(top), 'means_over_all_common': means(rows) if rows else {}, 'means_over_top1000': means(top) if top else {}, 'dirs': {'base': args.base, 'global': args.global_dir, 'sroute': args.sroute}, 'top': top}
+    json.dump(res, open(args.out_json, 'w'), ensure_ascii=False, indent=2)
+    with open(args.out_json.replace('.json', '_uids.txt'), 'w') as f:
+        for r in top:
+            f.write(r['take_uid'] + '\n')
+    print(''.join(['\n三方齐全 clip: ', format(n_common, '')]))
+    print(''.join(['满足 sroute>global>base (按全局SSIM): ', format(len(ordered), ''), ' 片, 取前 ', format(len(top), '')]))
+    for label, sub in (('前1000', top), ('全体交集', rows)):
+        if not sub:
+            continue
+        m = means(sub)
+        print(''.join(['\n[', format(label, ''), ' n=', format(len(sub), ''), ']  (全局 scope)']))
+        print(''.join([format('变体', ''.join(['8s'])), ' ', format('PSNR', ''.join(['>8s'])), ' ', format('SSIM', ''.join(['>8s'])), ' ', format('LPIPS', ''.join(['>8s']))]))
+        for v in VARS:
+            print(''.join([format(v, ''.join(['8s'])), ' ', format(m[v]['psnr'], ''.join(['8.3f'])), ' ', format(m[v]['ssim'], ''.join(['8.4f'])), ' ', format(m[v]['lpips'], ''.join(['8.4f']))]))
+    print(''.join(['\n已存: ', format(args.out_json, '')]))
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/train_caption.py

@@ -0,0 +1,223 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+import os
+import sys
+import glob
+import re
+import torch
+from tqdm import tqdm
+DS = _project_path('third_party/DiffSynth-Studio')
+sys.path.insert(0, DS)
+sys.path.insert(0, os.path.join(DS, 'examples/wanvideo/model_training'))
+sys.path.insert(0, os.path.dirname(__file__))
+import accelerate
+from diffsynth.core import UnifiedDataset
+from diffsynth.diffusion import ModelLogger, launch_training_task
+from diffsynth.diffusion.runner import initialize_deepspeed_gradient_checkpointing
+import train as ds_train
+from caption_cross_attn import inject_caption_cross_attn, set_caption_context, freeze_except_caption
+from flow_matching_loss import flow_match_sft_loss_logitnormal
+
+class WanCaptionTrainingModule(ds_train.WanTrainingModule):
+    """"""
+
+    def __init__(self, *args, caption_dim=4096, **kwargs):
+        super().__init__(*args, **kwargs)
+        dit = self.pipe.dit
+        dev = next(dit.parameters()).device
+        dtype = next(dit.parameters()).dtype
+        use_spatial_route = os.environ.get('USE_SPATIAL_ROUTE', '1') != '0'
+        inject_caption_cross_attn(dit, caption_dim=caption_dim, vace_module=self.pipe.vace, use_spatial_route=use_spatial_route)
+        for block in dit.blocks:
+            block.cross_attn._cap_branch.to(device=dev, dtype=dtype)
+        freeze_except_caption(dit)
+        gate_init = float(os.environ.get('GATE_INIT', '0.1'))
+        for block in dit.blocks:
+            g = block.cross_attn._cap_branch.gate_cap
+            g.data = torch.full_like(g.data, gate_init).float()
+        print(''.join(['[caption] gate 初始化 = ', format(gate_init, ''.join(['+.3f'])), ' (GATE_INIT, 0=原版零门控)']), flush=True)
+        self.task_to_loss['sft'] = lambda pipe, s, p, n: flow_match_sft_loss_logitnormal(pipe, **s, **p)
+        self.task_to_loss['sft:train'] = self.task_to_loss['sft']
+        print(''.join(['[caption] timestep 采样 = ', format(os.environ.get('TS_SAMPLING', 'logitnormal'), ''), ' (std=', format(os.environ.get('TS_LOGIT_STD', '1.0'), ''), ')']), flush=True)
+
+    @torch.no_grad()
+    def _encode_caption(self, data):
+        """"""
+        cap = data.get('caption', None) if isinstance(data, dict) else None
+        if cap is None:
+            return None
+        if isinstance(cap, (list, tuple)):
+            cap = cap[0] if cap else ''
+        if not isinstance(cap, str):
+            import math
+            if cap is None or (isinstance(cap, float) and math.isnan(cap)):
+                return None
+            cap = str(cap)
+        cap = cap.strip()
+        if len(cap) == 0:
+            return None
+        pipe = self.pipe
+        pipe.load_models_to_device(('text_encoder',))
+        ids, mask = pipe.tokenizer(cap, return_mask=True, add_special_tokens=True)
+        ids = ids.to(pipe.device)
+        mask = mask.to(pipe.device)
+        seq_lens = mask.gt(0).sum(dim=1).long()
+        feat = pipe.text_encoder(ids, mask)
+        for i, v in enumerate(seq_lens):
+            feat[:, v:] = 0
+        return feat.to(device=pipe.device, dtype=pipe.torch_dtype)
+
+    @torch.no_grad()
+    def _probe_delta(self, data, feat):
+        """"""
+        dit = self.pipe.dit
+        cpu_state = torch.get_rng_state()
+        cuda_state = torch.cuda.get_rng_state_all() if torch.cuda.is_available() else None
+
+        def _one_pass(use_cap):
+            torch.set_rng_state(cpu_state)
+            if cuda_state is not None:
+                torch.cuda.set_rng_state_all(cuda_state)
+            set_caption_context(dit, feat if use_cap else None)
+            return super(WanCaptionTrainingModule, self).forward(data, inputs=None).item()
+        loss_off = _one_pass(False)
+        loss_on = _one_pass(True)
+        torch.set_rng_state(cpu_state)
+        if cuda_state is not None:
+            torch.cuda.set_rng_state_all(cuda_state)
+        set_caption_context(dit, feat)
+        return (loss_off, loss_on)
+
+    def forward(self, data, inputs=None):
+        feat = self._encode_caption(data) if isinstance(data, dict) else None
+        self._step = getattr(self, '_step', 0) + 1
+        cap_dropout = float(os.environ.get('CAP_DROPOUT', '0.15'))
+        if feat is not None and cap_dropout > 0.0:
+            import random
+            if random.random() < cap_dropout:
+                feat = None
+        probe_every = int(os.environ.get('PROBE_EVERY', '50'))
+        if probe_every > 0 and feat is not None and (self._step % probe_every == 1):
+            loss_off, loss_on = self._probe_delta(data, feat)
+            delta = loss_off - loss_on
+            print(''.join(['[probe step ', format(self._step, ''), '] loss_off=', format(loss_off, ''.join(['.4f'])), ' loss_on=', format(loss_on, ''.join(['.4f'])), ' Δ(off-on)=', format(delta, ''.join(['+.4e'])), '  <-- Δ单调变大=caption分支在学']), flush=True)
+        set_caption_context(self.pipe.dit, feat)
+        loss = super().forward(data, inputs=inputs)
+        if not hasattr(self, '_cap_params'):
+            self._cap_params = [p for p in self.pipe.dit.parameters() if p.requires_grad]
+        if self._cap_params:
+            loss = loss + 0.0 * sum((p.sum() for p in self._cap_params))
+        if self._step % 10 == 1:
+            g = self.pipe.dit.blocks[0].cross_attn._cap_branch.gate_cap
+            print(''.join(['[step ', format(self._step, ''), '] loss=', format(loss.item(), ''.join(['.4f'])), '  gate0=', format(g.item(), ''.join(['+.3e']))]), flush=True)
+        return loss
+
+def _find_latest_epoch(output_path):
+    """"""
+    best_n, best_p = (None, None)
+    for p in glob.glob(os.path.join(output_path, 'epoch-*.safetensors')):
+        m = re.search('epoch-(\\d+)\\.safetensors$', os.path.basename(p))
+        if m:
+            n = int(m.group(1))
+            if best_n is None or n > best_n:
+                best_n, best_p = (n, p)
+    return (best_n, best_p)
+
+def _load_caption_branch(dit, ckpt_path):
+    """"""
+    from safetensors.torch import load_file
+    sd = load_file(ckpt_path)
+    tgt = dit.state_dict()
+    matched = {k: v for k, v in sd.items() if k in tgt}
+    bad_shape = [k for k in matched if tuple(tgt[k].shape) != tuple(matched[k].shape)]
+    for k in bad_shape:
+        matched.pop(k)
+    dit.load_state_dict(matched, strict=False)
+    return (len(matched), len(sd), bad_shape)
+
+def launch_training_task_resume(accelerator, dataset, model, model_logger, args, start_epoch=0):
+    """"""
+    optimizer = torch.optim.AdamW(model.trainable_modules(), lr=args.learning_rate, weight_decay=args.weight_decay)
+    scheduler = torch.optim.lr_scheduler.ConstantLR(optimizer)
+    dataloader = torch.utils.data.DataLoader(dataset, shuffle=True, collate_fn=lambda x: x[0], num_workers=args.dataset_num_workers)
+    model.to(device=accelerator.device)
+    model, optimizer, dataloader, scheduler = accelerator.prepare(model, optimizer, dataloader, scheduler)
+    initialize_deepspeed_gradient_checkpointing(accelerator)
+    if start_epoch > 0:
+        opt_path = os.path.join(args.output_path, ''.join(['optim-epoch-', format(start_epoch - 1, ''), '.pt']))
+        if os.path.exists(opt_path):
+            try:
+                ckpt = torch.load(opt_path, map_location='cpu')
+                optimizer.load_state_dict(ckpt['optimizer'])
+                accelerator.print(''.join(['>> [resume] 已恢复 optimizer 状态: ', format(opt_path, '')]))
+            except Exception as e:
+                accelerator.print(''.join(['>> [resume] optimizer 恢复失败(', format(e, ''), '), 仅分支权重热续, 动量从零']))
+        else:
+            accelerator.print(''.join(['>> [resume] 无 ', format(opt_path, ''), '(旧版没存 optim): 分支权重已热续, optimizer 从零(Adam 几十步补齐动量, 影响可忽略)']))
+    for epoch_id in range(start_epoch, args.num_epochs):
+        for data in tqdm(dataloader, disable=not accelerator.is_main_process, desc=''.join(['epoch ', format(epoch_id, '')])):
+            with accelerator.accumulate(model):
+                optimizer.zero_grad()
+                if dataset.load_from_cache:
+                    loss = model({}, inputs=data)
+                else:
+                    loss = model(data)
+                accelerator.backward(loss)
+                optimizer.step()
+                model_logger.on_step_end(accelerator, model, args.save_steps, loss=loss)
+                scheduler.step()
+        if args.save_steps is None:
+            model_logger.on_epoch_end(accelerator, model, epoch_id)
+            accelerator.wait_for_everyone()
+            if accelerator.is_main_process:
+                torch.save({'optimizer': optimizer.state_dict(), 'epoch': epoch_id}, os.path.join(args.output_path, ''.join(['optim-epoch-', format(epoch_id, ''), '.pt'])))
+    model_logger.on_training_end(accelerator, model, args.save_steps)
+
+def main():
+    import argparse
+    from generator_defaults import set_training_defaults
+    mode_parser = argparse.ArgumentParser(add_help=False)
+    mode_parser.add_argument('--variant', choices=['global', 'spatial', 'reward'], default='spatial')
+    mode, _ = mode_parser.parse_known_args()
+    parser = set_training_defaults(ds_train.wan_parser(), mode.variant)
+    parser.add_argument('--variant', choices=['global', 'spatial', 'reward'], default=mode.variant)
+    source = 'teacher' if mode.variant == 'reward' else 'grpo'
+    parser.add_argument('--caption_json', default=''.join(['data/captions/', format(source, ''), '.json']))
+    parser.add_argument('--source_csv', default='data/vace_train_v2/metadata.csv')
+    parser.add_argument('--caption_dim', type=int, default=4096, help='caption 特征维度 (umT5-xxl=4096)')
+    parser.add_argument('--resume', action='store_true', help='从 output_path 里最新 epoch-*.safetensors 热续 caption 分支(权重回读; optim 有则恢复)')
+    args = parser.parse_args()
+    os.environ['USE_SPATIAL_ROUTE'] = '1' if args.variant == 'spatial' else '0'
+    import subprocess
+    subprocess.run([sys.executable, _project_path('data/make_caption_branch_csv.py'), '--src_csv', args.source_csv, '--merged_cap', args.caption_json, '--out_csv', args.dataset_metadata_path, '--require_all'], check=True)
+    from training_data import validate_file
+    validate_file(args.dataset_metadata_path, ['video', 'vace_video', 'vace_reference_image', 'caption'], args.dataset_base_path)
+    accelerator = accelerate.Accelerator(gradient_accumulation_steps=args.gradient_accumulation_steps, kwargs_handlers=[accelerate.DistributedDataParallelKwargs(find_unused_parameters=args.find_unused_parameters)])
+    if os.environ.get('SQUEEZE_RESIZE', '1') == '1':
+        from patch_squeeze_resize import patch_squeeze_resize
+        patch_squeeze_resize()
+    dataset = UnifiedDataset(base_path=args.dataset_base_path, metadata_path=args.dataset_metadata_path, repeat=args.dataset_repeat, data_file_keys=args.data_file_keys.split(','), main_data_operator=UnifiedDataset.default_video_operator(base_path=args.dataset_base_path, max_pixels=args.max_pixels, height=args.height, width=args.width, height_division_factor=16, width_division_factor=16, num_frames=args.num_frames, time_division_factor=4, time_division_remainder=1))
+    model = WanCaptionTrainingModule(model_paths=args.model_paths, model_id_with_origin_paths=args.model_id_with_origin_paths, tokenizer_path=args.tokenizer_path, trainable_models=args.trainable_models, lora_base_model=args.lora_base_model, lora_target_modules=args.lora_target_modules, lora_rank=args.lora_rank, lora_checkpoint=args.lora_checkpoint, preset_lora_path=args.preset_lora_path, preset_lora_model=args.preset_lora_model, use_gradient_checkpointing=args.use_gradient_checkpointing, use_gradient_checkpointing_offload=args.use_gradient_checkpointing_offload, extra_inputs=args.extra_inputs, task=args.task, device='cpu' if args.initialize_model_on_cpu else accelerator.device, max_timestep_boundary=args.max_timestep_boundary, min_timestep_boundary=args.min_timestep_boundary, caption_dim=args.caption_dim)
+    model.dataset_base_path = os.path.abspath(args.dataset_base_path)
+    start_epoch = 0
+    if args.resume:
+        n, ckpt = _find_latest_epoch(args.output_path)
+        if ckpt is not None:
+            got, tot, bad = _load_caption_branch(model.pipe.dit, ckpt)
+            start_epoch = n + 1
+            if accelerator.is_main_process:
+                msg = ''.join(['>> [resume] 载入 caption 分支 epoch-', format(n, ''), ': ', format(got, ''), '/', format(tot, ''), ' key 命中'])
+                if bad:
+                    msg += ''.join([' (shape不符跳过: ', format(bad, ''), ')'])
+                print(msg + ''.join([', 从 epoch ', format(start_epoch, ''), ' 续训']), flush=True)
+        elif accelerator.is_main_process:
+            print('>> [resume] output_path 无 epoch-*.safetensors, 从头训练', flush=True)
+    model_logger = ModelLogger(args.output_path, remove_prefix_in_ckpt=args.remove_prefix_in_ckpt)
+    launch_training_task_resume(accelerator, dataset, model, model_logger, args, start_epoch=start_epoch)
+if __name__ == '__main__':
+    main()

source-overlays/egoexo/code/core/diffusion/train_hoi.py

@@ -0,0 +1,146 @@
+""""""
+from pathlib import Path as _Path
+import sys as _sys
+_bundle_root = _Path(__file__).resolve().parents[1]
+_sys.path.insert(0, str(_bundle_root))
+from project_paths import project_path as _project_path, env as _env, initialize as _initialize
+_initialize(__name__ == '__main__')
+import os, sys, re, glob, argparse, importlib.util
+import torch, accelerate
+from tqdm import tqdm
+DS = _project_path('third_party/DiffSynth-Studio')
+sys.path.insert(0, DS)
+from diffsynth.core import UnifiedDataset
+from diffsynth.core.data.operators import LoadVideo, LoadAudio, ImageCropAndResize, ToAbsolutePath
+from diffsynth.diffusion import ModelLogger
+from diffsynth.diffusion.runner import initialize_deepspeed_gradient_checkpointing
+_spec = importlib.util.spec_from_file_location('wan_train_mod', os.path.join(DS, 'examples/wanvideo/model_training/train.py'))
+_wan = importlib.util.module_from_spec(_spec)
+_spec.loader.exec_module(_wan)
+WanTrainingModule, wan_parser = (_wan.WanTrainingModule, _wan.wan_parser)
+_LAST_TS = {'id': -1}
+
+def flow_match_sft_loss_logitnormal(pipe, **inputs):
+    """"""
+    T = len(pipe.scheduler.timesteps)
+    max_b = int(inputs.get('max_timestep_boundary', 1) * T)
+    min_b = int(inputs.get('min_timestep_boundary', 0) * T)
+    mode = os.environ.get('TS_SAMPLING', 'logitnormal')
+    std = float(os.environ.get('TS_LOGIT_STD', '1.0'))
+    if mode == 'uniform' or std <= 0:
+        timestep_id = torch.randint(min_b, max_b, (1,))
+    else:
+        u = torch.sigmoid(torch.randn(1) * std).item()
+        timestep_id = torch.tensor([min_b + int(u * (max_b - min_b))]).clamp_(min_b, max_b - 1)
+    _LAST_TS['id'] = int(timestep_id.item())
+    timestep = pipe.scheduler.timesteps[timestep_id].to(dtype=pipe.torch_dtype, device=pipe.device)
+    noise = torch.randn_like(inputs['input_latents'])
+    inputs['latents'] = pipe.scheduler.add_noise(inputs['input_latents'], noise, timestep)
+    training_target = pipe.scheduler.training_target(inputs['input_latents'], noise, timestep)
+    if 'first_frame_latents' in inputs:
+        inputs['latents'][:, :, 0:1] = inputs['first_frame_latents']
+    models = {name: getattr(pipe, name) for name in pipe.in_iteration_models}
+    noise_pred = pipe.model_fn(**models, **inputs, timestep=timestep)
+    if 'first_frame_latents' in inputs:
+        noise_pred = noise_pred[:, :, 1:]
+        training_target = training_target[:, :, 1:]
+    loss = torch.nn.functional.mse_loss(noise_pred.float(), training_target.float())
+    loss = loss * pipe.scheduler.training_weight(timestep)
+    return loss
+
+def _find_latest_epoch(output_path):
+    """"""
+    best_n, best_p = (None, None)
+    for p in glob.glob(os.path.join(output_path, 'epoch-*.safetensors')):
+        m = re.search('epoch-(\\d+)\\.safetensors$', os.path.basename(p))
+        if m:
+            n = int(m.group(1))
+            if best_n is None or n > best_n:
+                best_n, best_p = (n, p)
+    return (best_n, best_p)
+
+def launch_training_task_with_log_resume(accelerator, dataset, model, model_logger, args, start_epoch=0):
+    optimizer = torch.optim.AdamW(model.trainable_modules(), lr=args.learning_rate, weight_decay=args.weight_decay)
+    scheduler = torch.optim.lr_scheduler.ConstantLR(optimizer)
+    dataloader = torch.utils.data.DataLoader(dataset, shuffle=True, collate_fn=lambda x: x[0], num_workers=args.dataset_num_workers)
+    model.to(device=accelerator.device)
+    model, optimizer, dataloader, scheduler = accelerator.prepare(model, optimizer, dataloader, scheduler)
+    initialize_deepspeed_gradient_checkpointing(accelerator)
+    if start_epoch > 0:
+        opt_path = os.path.join(args.output_path, ''.join(['optim-epoch-', format(start_epoch - 1, ''), '.pt']))
+        if os.path.exists(opt_path):
+            ckpt = torch.load(opt_path, map_location='cpu')
+            try:
+                optimizer.load_state_dict(ckpt['optimizer'])
+                accelerator.print(''.join(['>> [resume] 已恢复 optimizer 状态: ', format(opt_path, '')]))
+            except Exception as e:
+                accelerator.print(''.join(['>> [resume] optimizer 状态恢复失败(', format(e, ''), '), 仅从权重续训']))
+        else:
+            accelerator.print(''.join(['>> [resume] 未找到 ', format(opt_path, ''), ', 仅从 LoRA 权重续训(optimizer 从零)']))
+    loss_log = os.path.join(args.output_path, 'loss_log.csv')
+    if accelerator.is_main_process:
+        os.makedirs(args.output_path, exist_ok=True)
+        if start_epoch == 0 or not os.path.exists(loss_log):
+            with open(loss_log, 'w') as f:
+                f.write('epoch,global_step,timestep,loss\n')
+    global_step = 0
+    for epoch_id in range(start_epoch, args.num_epochs):
+        pbar = tqdm(dataloader, disable=not accelerator.is_main_process, desc=''.join(['epoch ', format(epoch_id, '')]))
+        running, count = (0.0, 0)
+        for data in pbar:
+            with accelerator.accumulate(model):
+                optimizer.zero_grad()
+                if dataset.load_from_cache:
+                    loss = model({}, inputs=data)
+                else:
+                    loss = model(data)
+                accelerator.backward(loss)
+                optimizer.step()
+                model_logger.on_step_end(accelerator, model, args.save_steps, loss=loss)
+                scheduler.step()
+            global_step += 1
+            lv = loss.item()
+            running += lv
+            count += 1
+            if accelerator.is_main_process:
+                pbar.set_postfix(loss=''.join([format(lv, ''.join(['.4f']))]), avg=''.join([format(running / count, ''.join(['.4f']))]), t=_LAST_TS['id'])
+                if global_step % args.log_interval == 0:
+                    with open(loss_log, 'a') as f:
+                        f.write(''.join([format(epoch_id, ''), ',', format(global_step, ''), ',', format(_LAST_TS['id'], ''), ',', format(lv, ''.join(['.6f'])), '\n']))
+        if accelerator.is_main_process:
+            accelerator.print(''.join(['>> [epoch ', format(epoch_id, ''), '] mean loss = ', format(running / max(count, 1), ''.join(['.6f']))]))
+        if args.save_steps is None:
+            model_logger.on_epoch_end(accelerator, model, epoch_id)
+            accelerator.wait_for_everyone()
+            if accelerator.is_main_process:
+                torch.save({'optimizer': optimizer.state_dict(), 'epoch': epoch_id}, os.path.join(args.output_path, ''.join(['optim-epoch-', format(epoch_id, ''), '.pt'])))
+    model_logger.on_training_end(accelerator, model, args.save_steps)
+if __name__ == '__main__':
+    from generator_defaults import set_training_defaults
+    parser = set_training_defaults(wan_parser(), 'hoi')
+    parser.add_argument('--resume', action='store_true', help='自动从 output_path 里最新的 epoch-*.safetensors 续训')
+    parser.add_argument('--log_interval', type=int, default=10, help='每多少 step 往 loss_log.csv 写一行')
+    args = parser.parse_args()
+    from training_data import validate_file
+    validate_file(args.dataset_metadata_path, ['video', 'vace_video', 'vace_reference_image'], args.dataset_base_path)
+    start_epoch = 0
+    if args.resume:
+        n, ckpt = _find_latest_epoch(args.output_path)
+        if ckpt is not None:
+            args.lora_checkpoint = ckpt
+            start_epoch = n + 1
+            print(''.join(['>> [resume] 发现 epoch-', format(n, ''), ' → 加载 ', format(ckpt, ''), ', 从 epoch ', format(start_epoch, ''), ' 继续']), flush=True)
+        else:
+            print('>> [resume] output_path 里没有 epoch-*.safetensors, 从头训练', flush=True)
+    accelerator = accelerate.Accelerator(gradient_accumulation_steps=args.gradient_accumulation_steps, kwargs_handlers=[accelerate.DistributedDataParallelKwargs(find_unused_parameters=args.find_unused_parameters)])
+    if os.environ.get('SQUEEZE_RESIZE', '1') == '1':
+        from patch_squeeze_resize import patch_squeeze_resize
+        patch_squeeze_resize()
+    dataset = UnifiedDataset(base_path=args.dataset_base_path, metadata_path=args.dataset_metadata_path, repeat=args.dataset_repeat, data_file_keys=args.data_file_keys.split(','), main_data_operator=UnifiedDataset.default_video_operator(base_path=args.dataset_base_path, max_pixels=args.max_pixels, height=args.height, width=args.width, height_division_factor=16, width_division_factor=16, num_frames=args.num_frames, time_division_factor=4, time_division_remainder=1), special_operator_map={'animate_face_video': ToAbsolutePath(args.dataset_base_path) >> LoadVideo(args.num_frames, 4, 1, frame_processor=ImageCropAndResize(512, 512, None, 16, 16)), 'input_audio': ToAbsolutePath(args.dataset_base_path) >> LoadAudio(sr=16000), 'wantodance_music_path': ToAbsolutePath(args.dataset_base_path)})
+    model = WanTrainingModule(model_paths=args.model_paths, model_id_with_origin_paths=args.model_id_with_origin_paths, tokenizer_path=args.tokenizer_path, audio_processor_path=args.audio_processor_path, trainable_models=args.trainable_models, lora_base_model=args.lora_base_model, lora_target_modules=args.lora_target_modules, lora_rank=args.lora_rank, lora_checkpoint=args.lora_checkpoint, preset_lora_path=args.preset_lora_path, preset_lora_model=args.preset_lora_model, use_gradient_checkpointing=args.use_gradient_checkpointing, use_gradient_checkpointing_offload=args.use_gradient_checkpointing_offload, extra_inputs=args.extra_inputs, fp8_models=args.fp8_models, offload_models=args.offload_models, task=args.task, device='cpu' if args.initialize_model_on_cpu else accelerator.device, max_timestep_boundary=args.max_timestep_boundary, min_timestep_boundary=args.min_timestep_boundary)
+    model.task_to_loss['sft'] = lambda pipe, s, p, n: flow_match_sft_loss_logitnormal(pipe, **s, **p)
+    model.task_to_loss['sft:train'] = model.task_to_loss['sft']
+    if accelerator.is_main_process:
+        print(''.join(['>> [ts] timestep 采样 = ', format(os.environ.get('TS_SAMPLING', 'logitnormal'), ''), ' (TS_LOGIT_STD=', format(os.environ.get('TS_LOGIT_STD', '1.0'), ''), ')']), flush=True)
+    model_logger = ModelLogger(args.output_path, remove_prefix_in_ckpt=args.remove_prefix_in_ckpt)
+    launch_training_task_with_log_resume(accelerator, dataset, model, model_logger, args, start_epoch=start_epoch)

source-overlays/egoexo/code/core/generator_defaults.py

@@ -0,0 +1,31 @@
+"""Native Python defaults shared by generator training entries."""
+import json
+from model_resources import wan_base_root, wan_model_paths, wan_tokenizer_path
+
+
+def set_training_defaults(parser, variant):
+    model = wan_base_root()
+    data = "data/vace_train_v2"
+    caption_source = "teacher" if variant == "reward" else "grpo"
+    defaults = dict(
+        dataset_base_path=data,
+        dataset_metadata_path=f"{data}/metadata.csv" if variant == "hoi" else f"{data}/metadata_caption_branch_{caption_source}.csv",
+        data_file_keys="video,vace_video,vace_reference_image",
+        extra_inputs="vace_video,vace_reference_image",
+        height=480, width=480, num_frames=13, dataset_repeat=1, num_epochs=5,
+        model_paths=json.dumps(wan_model_paths(model)),
+        tokenizer_path=wan_tokenizer_path(model),
+        use_gradient_checkpointing_offload=True,
+        output_path=f"outputs/train_{variant}",
+        learning_rate=1e-4 if variant == "hoi" else 3e-4,
+    )
+    if variant == "hoi":
+        defaults.update(lora_base_model="vace", lora_target_modules="q,k,v,o,ffn.0,ffn.2", lora_rank=128)
+    else:
+        defaults.update(preset_lora_path="checkpoints/generator_hoi/ckpt.safetensors", preset_lora_model="vace", find_unused_parameters=True)
+    parser.set_defaults(**defaults)
+    # Upstream required options now have concrete package defaults.
+    for action in parser._actions:
+        if action.dest in defaults:
+            action.required = False
+    return parser

source-overlays/egoexo/code/core/model_resources.py

@@ -0,0 +1,70 @@
+"""Resolve original or size-bounded Wan components without importing torch.
+
+DiffSynth accepts one string or a list of shard paths per ModelConfig. Do not
+pass an index JSON to its binary loader or treat each shard as a separate model.
+WAN_MODEL_ROOT may point to an independently downloaded model bundle.
+"""
+import json
+import os
+from pathlib import Path
+from project_paths import project_path
+
+
+def wan_base_root(root=None):
+    return str(Path(root or os.environ.get('WAN_MODEL_ROOT') or
+                    project_path('models/Wan2.1-VACE-1.3B')).expanduser().resolve())
+
+
+def indexed_shards(index_path):
+    index_path = Path(index_path)
+    index = json.loads(index_path.read_text())
+    mapping = index.get('weight_map')
+    if not isinstance(mapping, dict) or not mapping:
+        raise ValueError('empty or invalid model weight index: ' + str(index_path))
+    names = set()
+    for name in mapping.values():
+        if not isinstance(name, str) or Path(name).name != name or not name.endswith('.safetensors'):
+            raise ValueError('unsafe shard path in model index')
+        names.add(name)
+    paths = [index_path.parent / name for name in sorted(names)]
+    for path in paths:
+        if not path.is_file():
+            raise FileNotFoundError('missing model shard: ' + str(path))
+    return [str(path) for path in paths]
+
+
+def component_path(root, original, indices, strict=False):
+    original_path = root / original
+    if original_path.is_file():
+        return str(original_path)
+    for name in indices:
+        index_path = root / name
+        if index_path.is_file():
+            return indexed_shards(index_path)
+    if strict:
+        raise FileNotFoundError('missing original or indexed model component: ' + str(original_path))
+    # Preserve historical command defaults when no weights are present (e.g. dry-run).
+    return str(original_path)
+
+
+def wan_model_paths(root=None, strict=False):
+    root = Path(wan_base_root(root))
+    return [
+        component_path(root, 'diffusion_pytorch_model.safetensors', (
+            'transformer/diffusion_pytorch_model.safetensors.index.json',
+            'diffusion_pytorch_model.safetensors.index.json',
+        ), strict),
+        component_path(root, 'models_t5_umt5-xxl-enc-bf16.pth', (
+            'text_encoder/model.safetensors.index.json',
+            'text_encoder/models_t5_umt5-xxl-enc-bf16.safetensors.index.json',
+            'models_t5_umt5-xxl-enc-bf16.safetensors.index.json',
+        ), strict),
+        component_path(root, 'Wan2.1_VAE.pth', (), strict),
+    ]
+
+
+def wan_tokenizer_path(root=None, strict=False):
+    path = Path(wan_base_root(root)) / 'google/umt5-xxl'
+    if strict and not (path / 'tokenizer.json').is_file() and not (path / 'spiece.model').is_file():
+        raise FileNotFoundError('missing Wan umT5 tokenizer: ' + str(path))
+    return str(path)

tools/apply_source_overlay.py

@@ -0,0 +1,18 @@
+#!/usr/bin/env python3
+"""Apply only exact version/hash-matched code overlays; dry-run unless execute."""
+import argparse,hashlib,json,subprocess
+from pathlib import Path
+ROOT=Path(__file__).absolute().parents[1]
+def sha(p):return hashlib.sha256(Path(p).read_bytes()).hexdigest()
+def main():
+ p=argparse.ArgumentParser();p.add_argument('--code-checkout',type=Path,required=True);p.add_argument('--execute',action='store_true');a=p.parse_args();m=json.loads((ROOT/'SOURCE_OVERLAY_MANIFEST.json').read_text())
+ revision=subprocess.check_output(['git','-C',str(a.code_checkout),'rev-parse','HEAD'],text=True).strip()
+ if revision!=m['revision']:raise ValueError('Checkout revision differs from locked source')
+ for r in m['files']:
+  target=a.code_checkout/r['path'];source=ROOT/r['patch_file']
+  if target.exists() and sha(target)==r['patched_sha256']:continue
+  if target.exists() and sha(target)!=r['expected_git_source_sha256']:raise ValueError('Custom source modifications found; refusing to overwrite')
+  if not target.exists() and r['expected_git_source_sha256']:raise ValueError('Expected source file is missing')
+  print('PLAN',r['path'])
+  if a.execute:target.parent.mkdir(parents=True,exist_ok=True);target.write_bytes(source.read_bytes())
+if __name__=='__main__':main()

tools/capture_environment.py

@@ -0,0 +1,73 @@
+#!/usr/bin/env python3
+"""Read-only interpreter/package provenance capture; no secrets/model downloads."""
+import argparse,hashlib,importlib.metadata as md,importlib.util,json,os,platform,re,subprocess,sys,sysconfig
+from pathlib import Path
+from urllib.parse import urlparse,urlunparse
+CORE={'torch','torchvision','torchaudio','transformers','tokenizers','huggingface-hub','diffusers','diffsynth','accelerate','peft','ms-swift','deepspeed','flash-attn','vllm','verl','flash-linear-attention','tilelang','triton','ray','tensordict','torchcodec','numpy','scipy','pandas','pyarrow','pillow','opencv-python','opencv-python-headless','safetensors','qwen-vl-utils'}
+MODULES=['torch','torchvision','torchaudio','transformers','tokenizers','huggingface_hub','diffusers','diffsynth','accelerate','peft','swift','verl','vllm','fla','tilelang','triton','cv2','numpy']
+
+def normalized(x):return re.sub('[-_.]+','-',x).lower()
+def redact(value):
+ text=str(value)
+ text=re.sub(r'\b(?:hf_|ghp_|github_pat_|sk-|pk-)[A-Za-z0-9_-]{20,}','[REDACTED]',text)
+ text=re.sub(r'/pfs/[^/\s\"\']+','/local/workspace',text)
+ text=re.sub(r'/root/','/local/home/',text)
+ return text
+
+def safe_url(value):
+ u=urlparse(value)
+ if u.scheme=='file':return {'type':'local_source','portable':False,'directory_name':Path(u.path).name}
+ clean_netloc=u.hostname or ''
+ if u.port:clean_netloc+=':'+str(u.port)
+ return {'type':'network_source','url':urlunparse((u.scheme,clean_netloc,u.path,'','',''))}
+
+def run(command):
+ p=subprocess.run(command,stdout=subprocess.PIPE,stderr=subprocess.STDOUT,text=True,env={**os.environ,'PYTHONDONTWRITEBYTECODE':'1','HF_HUB_OFFLINE':'1','TRANSFORMERS_OFFLINE':'1','MODELSCOPE_OFFLINE':'1'},timeout=180)
+ return {'exit_code':p.returncode,'output':redact(p.stdout[-20000:])}
+
+def main():
+ p=argparse.ArgumentParser(description=__doc__);p.add_argument('--name',required=True);p.add_argument('--out',type=Path,required=True);p.add_argument('--private-out',type=Path,required=True);p.add_argument('--gpu-basic-check',action='store_true');a=p.parse_args();a.out.mkdir(parents=True,exist_ok=True);a.private_out.mkdir(parents=True,exist_ok=True)
+ packages={};origins={};raw_origins={}
+ for d in md.distributions():
+  name=d.metadata.get('Name')
+  if not name:continue
+  key=normalized(name);packages[key]=d.version
+  if key in CORE:
+   req=d.requires or []
+   origins[key]={'version':d.version,'requires_dist':req,'installer':(d.read_text('INSTALLER') or '').strip()}
+   direct=d.read_text('direct_url.json')
+   if direct:
+    obj=json.loads(direct);raw_origins[key]=obj;origins[key]['origin']=safe_url(obj.get('url',''))
+    for extra in ('vcs_info','dir_info','archive_info'):
+     if extra in obj:origins[key][extra]=obj[extra]
+ modules={};private_modules={}
+ for name in MODULES:
+  try:s=importlib.util.find_spec(name)
+  except Exception as e:modules[name]={'found':False,'error_type':type(e).__name__};continue
+  if not s:modules[name]={'found':False};continue
+  path=s.origin;private_modules[name]=path
+  modules[name]={'found':True,'module_file':Path(path).name if path else None}
+  if path and Path(path).is_file():modules[name]['entry_file_sha256']=hashlib.sha256(Path(path).read_bytes()).hexdigest()
+ check=run([sys.executable,'-m','pip','check'])
+ conda=[];prefix=Path(sys.prefix)
+ if (prefix/'conda-meta').is_dir():
+  for meta in sorted((prefix/'conda-meta').glob('*.json')):
+   d=json.loads(meta.read_text());url=d.get('url')
+   if url and 'file://' not in url:
+    clean=safe_url(url)['url'];conda.append(clean+('#'+d['md5'] if d.get('md5') else ''))
+  (a.out/'conda-explicit-linux-64.txt').write_text('@EXPLICIT\n'+'\n'.join(conda)+'\n')
+ gpu={'checked':False,'training_or_model_inference_run':False}
+ if a.gpu_basic_check:
+  result=run([sys.executable,'-c',"import json,torch; d={'torch':torch.__version__,'cuda_runtime':torch.version.cuda,'cuda_available':torch.cuda.is_available(),'arch_list':torch.cuda.get_arch_list()};\nif torch.cuda.is_available():\n d.update(gpu=torch.cuda.get_device_name(0),capability=list(torch.cuda.get_device_capability(0)));x=torch.ones((2,2),device='cuda',dtype=torch.bfloat16);d['tiny_bfloat16_result']=float(x.sum());torch.cuda.synchronize()\nprint(json.dumps(d))"])
+  gpu={'checked':True,'basic_tensor_check':result,'training_or_model_inference_run':False}
+ snapshot={'profile':a.name,'capture_scope':'observed current retained runtime; not proof of original launch-time environment',
+  'python':platform.python_version(),'implementation':platform.python_implementation(),'SOABI':sysconfig.get_config_var('SOABI'),'machine':platform.machine(),'platform':platform.system(),'libc':list(platform.libc_ver()),
+  'packages':dict(sorted(packages.items())),'core_package_provenance':origins,'module_presence':modules,'pip_check':check,'GPU':gpu,
+  'original_environment_modified':False,'credentials_included':False}
+ text=json.dumps(snapshot,ensure_ascii=False,indent=2)+'\n'
+ if re.search(r'\b(?:hf_|ghp_|github_pat_|sk-|pk-)[A-Za-z0-9_-]{25,}',text):raise ValueError('Unsafe credential-like capture')
+ (a.out/'environment.snapshot.json').write_text(text)
+ (a.out/'requirements.observed-full.lock.txt').write_text('\n'.join(k+'=='+v for k,v in sorted(packages.items()))+'\n')
+ (a.private_out/(a.name+'.private.json')).write_text(json.dumps({'python_executable':sys.executable,'prefix':sys.prefix,'module_paths':private_modules,'raw_direct_origins':raw_origins},indent=2)+'\n')
+ print(json.dumps({'profile':a.name,'python':snapshot['python'],'packages':len(packages),'core_versions':{k:packages.get(k) for k in sorted(CORE) if k in packages},'pip_check_exit':check['exit_code'],'GPU_basic_check':gpu},indent=2),flush=True)
+if __name__=='__main__':main()

tools/fetch_training_data.py

@@ -0,0 +1,53 @@
+#!/usr/bin/env python3
+"""Fetch pinned training files from the JSON index; dry-run unless --execute."""
+import argparse,hashlib,json,shutil
+from pathlib import Path,PurePosixPath
+
+
+def safe(value):
+ p=PurePosixPath(value)
+ if p.is_absolute() or '..' in p.parts or not p.parts:raise ValueError('Unsafe relative data path')
+ return Path(*p.parts)
+
+def sha(path):
+ h=hashlib.sha256()
+ with Path(path).open('rb') as f:
+  for b in iter(lambda:f.read(8*1024*1024),b''):h.update(b)
+ return h.hexdigest()
+
+def main():
+ p=argparse.ArgumentParser(description=__doc__)
+ directory=Path(__file__).absolute().parent
+ default_index=directory/'TRAINING_DATA_INDEX.json'
+ if not default_index.exists():default_index=directory.parent/'TRAINING_DATA_INDEX.json'
+ p.add_argument('--index',type=Path,default=default_index)
+ p.add_argument('--dataset',default='legacy_mix6_sft')
+ p.add_argument('--output',type=Path,default=Path('training_data_download'))
+ p.add_argument('--execute',action='store_true');a=p.parse_args()
+ index=json.loads(a.index.read_text());datasets={d['dataset_id']:d for d in index['datasets']}
+ if a.dataset=='all_private':ids={r['artifact_id'] for r in index['private_artifacts']}
+ elif a.dataset in datasets:ids=set(datasets[a.dataset].get('artifact_ids',[]))
+ else:raise ValueError('Unknown dataset ID; inspect index.datasets')
+ rows=[r for r in index['private_artifacts'] if r['artifact_id'] in ids]
+ if not rows:raise ValueError('No private snapshot artifacts for this dataset; inspect public download location')
+ source=index['private_snapshot']
+ if not source['repo_id'].startswith('diffusion-gif/training-data-vault-') or source['repo_type']!='dataset' or not source['private']:raise ValueError('Unexpected private data provider')
+ print(json.dumps({'dataset':a.dataset,'files':len(rows),'bytes':sum(r['bytes'] for r in rows),'source_repo':source['repo_id'],'revision':source['revision'],'login_required':True,'execute':a.execute},indent=2))
+ if not a.execute:return
+ from huggingface_hub import get_token,hf_hub_download
+ token=get_token()
+ if not token:raise RuntimeError('Run hf auth login locally using an account authorized for the private vault. Never put tokens in this file or chat.')
+ a.output.mkdir(parents=True,exist_ok=True)
+ for r in rows:
+  target=a.output/safe(r['path'])
+  if target.exists():
+   if not target.is_file() or target.stat().st_size!=r['bytes'] or sha(target)!=r['sha256']:raise ValueError('Existing user file differs; will not overwrite')
+   continue
+  cached=Path(hf_hub_download(repo_id=source['repo_id'],repo_type='dataset',revision=source['revision'],filename=r['path'],token=token,endpoint='https://huggingface.co'))
+  if cached.stat().st_size!=r['bytes'] or sha(cached)!=r['sha256']:raise ValueError('Downloaded file SHA256 mismatch')
+  target.parent.mkdir(parents=True,exist_ok=True);temporary=target.with_name(target.name+'.copy-partial')
+  shutil.copyfile(cached,temporary)
+  if sha(temporary)!=r['sha256']:raise ValueError('Copied file SHA256 mismatch')
+  temporary.replace(target);print('VERIFIED',r['path'],flush=True)
+ print('Done: all selected frozen files downloaded and SHA256 verified. See restore_legacy_mix6.py for a relocatable SFT working copy.')
+if __name__=='__main__':main()

tools/install_environment.py

@@ -0,0 +1,39 @@
+#!/usr/bin/env python3
+"""Prepare a NEW environment from observed version constraints; default dry-run.
+
+Never modifies the original training environments. GPU source builds require a
+matching CUDA toolkit, not merely a CUDA-capable driver/runtime wheel.
+"""
+import argparse,json,os,subprocess,sys
+from pathlib import Path
+ROOT=Path(__file__).absolute().parents[1]
+PROFILES={'egoexo-generator':'egoexo/generator','egoexo-captioner-phiseg':'egoexo/captioner_and_phiseg','robot-sft-grpo':'robot/legacy_sft_grpo','robot-data-cpu':'robot/data_cpu'}
+
+def main():
+ p=argparse.ArgumentParser(description=__doc__);p.add_argument('--profile',choices=PROFILES,required=True);p.add_argument('--prefix',type=Path,required=True);p.add_argument('--python',default=sys.executable);p.add_argument('--cuda-home',type=Path);p.add_argument('--execute',action='store_true');p.add_argument('--flash-attn',action='store_true');a=p.parse_args()
+ folder=ROOT/PROFILES[a.profile];snap=json.loads((folder/'environment.snapshot.json').read_text());version=snap['python'];gpu=a.profile!='robot-data-cpu'
+ cuda='cu128' if a.profile=='egoexo-captioner-phiseg' else 'cu130'
+ commands=[[a.python,'-m','venv',str(a.prefix)],
+           [str(a.prefix/'bin/python'),'-m','pip','install','pip','setuptools','wheel','packaging','ninja']]
+ if a.profile=='robot-data-cpu' and (folder/'requirements.install.lock.txt').exists():
+  commands.append([str(a.prefix/'bin/python'),'-m','pip','install','--require-hashes','-r',str(folder/'requirements.install.lock.txt')])
+ else:
+  commands.append([str(a.prefix/'bin/python'),'-m','pip','install','-r',str(folder/'requirements.core.in'),'-c',str(folder/'requirements.observed-full.lock.txt'),'--index-url','https://pypi.org/simple','--extra-index-url','https://download.pytorch.org/whl/'+cuda])
+ if a.profile.startswith('egoexo'):
+  commands.append([str(a.prefix/'bin/python'),'-m','pip','install','decord==0.6.0'])
+ if a.flash_attn:
+  if a.profile!='robot-sft-grpo':raise ValueError('This exact flash-attn pin belongs to the robot torch2.11/cu130 profile')
+  commands.append([str(a.prefix/'bin/python'),'-m','pip','install','--no-build-isolation','flash-attn==2.8.3'])
+ print(json.dumps({'profile':a.profile,'required_python':version,'cuda_wheel_track':cuda if gpu else 'CPU only',
+                  'commands':commands,'execute':a.execute,'original_environment_modified':False,
+                  'fresh_GPU_install_and_8GPU_training_previously_verified':False},indent=2))
+ if not a.execute:return
+ current=subprocess.check_output([a.python,'-c','import platform;print(platform.python_version())'],text=True).strip()
+ if current!=version:raise ValueError('Interpreter must match captured Python '+version+'; create/select it explicitly before running this installer')
+ if a.prefix.exists():raise ValueError('Prefix already exists; refusing to modify an existing environment')
+ env={**os.environ,'PYTHONDONTWRITEBYTECODE':'1','MAX_JOBS':os.environ.get('MAX_JOBS','4'),'DS_BUILD_OPS':'0'}
+ if a.cuda_home:env['CUDA_HOME']=str(a.cuda_home);env['PATH']=str(a.cuda_home/'bin')+':'+env.get('PATH','')
+ if a.flash_attn and not a.cuda_home:raise ValueError('FlashAttention source build needs --cuda-home pointing to a CUDA13 toolkit matching torch cu130')
+ for command in commands:subprocess.run(command,env=env,check=True)
+ print('Installed role-specific pinned dependencies. Run verify_environment.py before training; optional source dependencies are specified in SOURCE_LOCK.json.')
+if __name__=='__main__':main()

tools/restore_legacy_mix6.py

@@ -0,0 +1,40 @@
+#!/usr/bin/env python3
+"""Restore legacy SFT observations/JSON working copies; never alter frozen files."""
+import argparse,hashlib,json,tarfile
+from pathlib import Path
+
+def main():
+ p=argparse.ArgumentParser(description=__doc__);p.add_argument('--snapshot-root',required=True,type=Path);p.add_argument('--output',required=True,type=Path);p.add_argument('--execute',action='store_true');a=p.parse_args()
+ src=a.snapshot_root/'legacy_mix6';locations={}
+ for line in (src/'observations/locations.private.jsonl').open():
+  r=json.loads(line);key=(r['split'],r['sid'])
+  if key in locations:raise ValueError('Duplicate observation ID')
+  if Path(r['archive']).name!=r['archive'] or Path(r['member']).name!=r['member']:raise ValueError('Unsafe archive/member path')
+  locations[key]=r
+ if not a.execute:
+  print(json.dumps({'records':len(locations),'output':str(a.output),'source_snapshot_modified':False,'execute':False},indent=2));return
+ if a.output.exists() and any(a.output.iterdir()):raise ValueError('Output must be a new/empty directory; will not overwrite user files')
+ a.output.mkdir(parents=True,exist_ok=True);by_archive={}
+ for key,r in locations.items():by_archive.setdefault(r['archive'],{})[r['member']]=(key,r)
+ for name,entries in by_archive.items():
+  with tarfile.open(src/'observations'/name,'r:') as archive:
+   for member in archive:
+    entry=entries.pop(member.name,None)
+    if entry is None or not member.isfile():raise ValueError('Unexpected archive entry')
+    (split,sid),r=entry;data=archive.extractfile(member).read()
+    if len(data)!=r['bytes'] or hashlib.sha256(data).hexdigest()!=r['sha256']:raise ValueError('Original NPZ checksum differs')
+    filename=hashlib.sha256(sid.encode()).hexdigest()[:32]+'.npz'
+    target=a.output/'frames'/split/filename;target.parent.mkdir(parents=True,exist_ok=True)
+    if target.exists():raise ValueError('Duplicate target frame file')
+    target.write_bytes(data)
+  if entries:raise ValueError('Missing observation member')
+ for split in ('train','heldout'):
+  with (src/'sft'/(split+'.jsonl')).open() as inp,(a.output/(split+'.jsonl')).open('w') as out:
+   for line in inp:
+    r=json.loads(line)
+    if (split,r['sid']) not in locations:raise ValueError('Input record lacks backed-up observation')
+    r['frames_root']=str((a.output/'frames'/split).resolve())
+    r['npz']=hashlib.sha256(r['sid'].encode()).hexdigest()[:32]+'.npz'
+    out.write(json.dumps(r,ensure_ascii=False)+'\n')
+ print('Restored exact NPZ bytes and re-based SFT train/heldout JSONL. Only npz/frames_root fields in WORKING copies changed; frozen originals are intact.')
+if __name__=='__main__':main()

tools/run_training_recipe.py

@@ -0,0 +1,29 @@
+#!/usr/bin/env python3
+"""Print portable training commands; only --execute starts training."""
+import argparse,json,os,subprocess,sys
+from pathlib import Path
+
+def main():
+ p=argparse.ArgumentParser(description=__doc__);p.add_argument('--recipe',choices=['egoexo-hoi','egoexo-global','egoexo-spatial','egoexo-phiseg','robot-sft','robot-grpo'],required=True);p.add_argument('--code-root',type=Path,required=True);p.add_argument('--python',required=True);p.add_argument('--data-root',type=Path,required=True);p.add_argument('--model-root',type=Path,required=True);p.add_argument('--output',type=Path,required=True);p.add_argument('--diffsynth-root',type=Path);p.add_argument('--gpus',type=int,default=8);p.add_argument('--execute',action='store_true');a=p.parse_args()
+ env={'OMP_NUM_THREADS':'1','TOKENIZERS_PARALLELISM':'false','PYTHONDONTWRITEBYTECODE':'1','HF_HUB_OFFLINE':'1','MODELSCOPE_OFFLINE':'1','TRANSFORMERS_OFFLINE':'1'}
+ if a.recipe.startswith('robot'):
+  env.update(ROBOT_PYTHON=a.python,ROBOT_TORCHRUN=str(Path(a.python).parent/'torchrun'),ROBOT_USER_ROOT=str(a.output),ROBOT_EXPERIMENT_ROOT=str(a.output),ROBOT_DATA_ROOT=str(a.data_root),ROBOT_MODEL_ROOT=str(a.model_root),DATASET='mix6',NPROC=str(a.gpus),TRAJ2D_SFT_DIR=str(a.data_root/'sft'),TRAJ2D_GRPO_DIR=str(a.data_root/'grpo'),GRPO_TRAIN_LIMIT='0',GRPO_HELDOUT_LIMIT='0',MODEL=str(a.model_root),ROBOT_TRAIN_SEED='42')
+  cwd=a.code_root/'scripts'
+  command=['bash','run_sft_traj2d.sh'] if a.recipe=='robot-sft' else ['bash','run_grpo_traj2d.sh','full']
+ else:
+  cwd=a.code_root
+  env.update(MODEL_ROOT=str(a.model_root),WAN_MODEL_ROOT=str(a.model_root/'models/Wan2.1-VACE-1.3B'),SQUEEZE_RESIZE='1')
+  env['PYTHONPATH']=str(cwd)+((':'+str(a.diffsynth_root)) if a.diffsynth_root else '')
+  base=[a.python,'-m','torch.distributed.run','--nproc_per_node',str(a.gpus)]
+  if a.recipe=='egoexo-phiseg':command=base+['diffusion/train_phiseg.py','--roots',str(a.data_root/'hoi_dataset'),str(a.data_root/'hoi_dataset_v2'),str(a.data_root/'hoi_dataset_extra'),'--output_dir',str(a.output),'--exo_full_context']
+  else:
+   script='diffusion/train_hoi.py' if a.recipe=='egoexo-hoi' else 'diffusion/train_caption.py'
+   command=base+[script,'--dataset_base_path',str(a.data_root),'--dataset_metadata_path',str(a.data_root/'vace_train_v2/metadata.csv') if a.recipe=='egoexo-hoi' else str(a.output/'derived_caption_training.csv'),'--output_path',str(a.output),'--height','480','--width','480','--num_frames','13']
+   if a.recipe!='egoexo-hoi':
+    variant='global' if a.recipe=='egoexo-global' else 'spatial';command+=['--variant',variant,'--source_csv',str(a.data_root/'vace_train_v2/metadata.csv'),'--caption_json',str(a.data_root/'captions/teacher.json'),'--preset_lora_path',str(a.model_root/'checkpoints/generator_hoi/ckpt.safetensors')]
+ print(json.dumps({'recipe':a.recipe,'command':command,'cwd':str(cwd),'environment':env,'execute':a.execute,'not_a_claim_of_original_metrics_or_exact_launch_reconstruction':True},indent=2))
+ if not a.execute:return
+ if a.output.exists():raise ValueError('Training output must be new; will not overwrite checkpoints')
+ if a.recipe=='robot-grpo' and a.gpus!=8:raise ValueError('Historical GRPO configuration is TP=4 / 8 GPUs. Do not silently change topology')
+ a.output.mkdir(parents=True);subprocess.run(command,cwd=cwd,env={**os.environ,**env},check=True)
+if __name__=='__main__':main()

tools/verify_environment.py

@@ -0,0 +1,26 @@
+#!/usr/bin/env python3
+"""Preflight versions/imports/driver vs toolkit; no weights loaded or training."""
+import argparse,importlib,json,os,platform,shutil,subprocess,sys
+from pathlib import Path
+ROOT=Path(__file__).absolute().parents[1]
+MAP={'egoexo-generator':'egoexo/generator','egoexo-captioner-phiseg':'egoexo/captioner_and_phiseg','robot-sft-grpo':'robot/legacy_sft_grpo','robot-data-cpu':'robot/data_cpu'}
+def main():
+ p=argparse.ArgumentParser(description=__doc__);p.add_argument('--profile',choices=MAP,required=True);p.add_argument('--gpu-basic-check',action='store_true');a=p.parse_args();folder=ROOT/MAP[a.profile];snapshot=json.loads((folder/'environment.snapshot.json').read_text())
+ result={'profile':a.profile,'python':platform.python_version(),'expected_python':snapshot['python'],'weights_loaded':False,'training_run':False,'modules':{}}
+ modules=['numpy','PIL','huggingface_hub'] if a.profile=='robot-data-cpu' else ['torch','torchvision','torchaudio','transformers','accelerate','safetensors']
+ if a.profile=='robot-sft-grpo':modules+=['verl','vllm','fla','tilelang']
+ if a.profile=='egoexo-captioner-phiseg':modules+=['swift']
+ for name in modules:
+  try:m=importlib.import_module(name);result['modules'][name]={'ok':True,'version':str(getattr(m,'__version__','unknown'))}
+  except Exception as e:result['modules'][name]={'ok':False,'error_type':type(e).__name__}
+ if 'torch' in sys.modules:
+  import torch
+  result.update(torch_cuda_runtime=torch.version.cuda,gpu_available=torch.cuda.is_available(),gpu_count=torch.cuda.device_count())
+  nvcc=shutil.which('nvcc');result['cuda_toolkit_nvcc_found']=bool(nvcc)
+  if nvcc:result['nvcc_version']=subprocess.check_output([nvcc,'--version'],text=True).strip().splitlines()[-2:]
+  if a.gpu_basic_check and torch.cuda.is_available():
+   x=torch.ones((2,2),device='cuda',dtype=torch.bfloat16);result['tiny_bfloat16_sum']=float(x.sum());torch.cuda.synchronize()
+  if a.profile=='robot-sft-grpo' and torch.cuda.device_count()<8:result['8GPU_GRPO_preconditions_met']=False
+ print(json.dumps(result,indent=2))
+ if result['python']!=result['expected_python'] or any(not x['ok'] for x in result['modules'].values()):raise SystemExit(1)
+if __name__=='__main__':main()