GR00T Whole-Body VLA: Training SONIC Controller

Disclosure: This article may contain affiliate or referral links. If you buy or sign up through those links, VnRobo may earn a commission or service credit.

The first three parts focused on GR00T-LeRobot data for VLA. The NVlabs/GR00T-WholeBodyControl repo adds another critical layer: the SONIC whole-body controller. In a G1/GEAR-SONIC workflow, the VLA usually does not command every motor directly. It produces a higher-level action or latent representation, while SONIC and the C++ deployment stack turn it into stable whole-body motion.

This article covers the SONIC data and training pipeline:

text

Bones-SEED / robot motion / SMPL / SOMA
  -> convert motion data
  -> filter impossible motions
  -> PPO training in Isaac Lab
  -> evaluate tracking metrics
  -> export ONNX
  -> C++ deploy stack

4.1 Goal

By the end, you should be able to:

Prepare motion-reference data for SONIC.
Run local debug training.
Understand when you need 8 GPUs, 64+ GPUs, or just a released checkpoint.
Evaluate checkpoints with tracking metrics.
Export ONNX for C++ deployment.
Diagnose common failures: Git LFS, Isaac Lab env, body-name mismatch, motion file paths, TensorRT.

Important distinction:

Layer	Data	Goal
GR00T VLA	GR00T-LeRobot: image/state/action/language	Learn a policy from observation and instruction.
SONIC WBC	motion_lib PKL, SMPL, SOMA, G1 trajectories	Learn a whole-body motion/controller foundation.
C++ deploy	ONNX/TensorRT + ZMQ + robot state	Run realtime control in sim or on real hardware.

When Do You Need To Train SONIC?

Not every project needs custom SONIC training. For manipulation tasks on Unitree G1 using the official workflow, the faster path is usually:

text

real/sim/public LeRobot data
  -> fine-tune GR00T VLA with UNITREE_G1_SONIC
  -> use the released SONIC checkpoint/deploy stack
  -> infer with PolicyServer + launch_inference.py

Train or fine-tune SONIC yourself only when:

You need a controller for a non-G1 embodiment or changed joint/body layout.
The released checkpoint does not cover the motion skills you need.
You are researching controller foundations, motion tracking, or SOMA/SMPL encoders.
You need custom ONNX/TensorRT export for an internal C++ deployment stack.

If your main goal is VLA training from sim or real data, finish Parts 1-3 first. Part 4 is a lower-level controller layer with much higher compute cost.

4.2 Environment And Hardware

The GR00T-WholeBodyControl training docs describe SONIC training as Isaac Lab PPO with many parallel environments. They recommend 64+ GPUs for reasonable convergence time; one 8-GPU node works but is much slower. You do not need 64 GPUs to learn the pipeline:

Goal	Practical GPU/VRAM
Load env + replay motion	1 GPU, 16-24 GB
Debug 100 training iterations	1 GPU, 24-48 GB
Single-node training	8 GPUs, more VRAM is better
Serious convergence	64+ GPUs according to docs
ONNX export	1 GPU or CPU depending on env/checkpoint

Setup:

bash

sudo apt update
sudo apt install -y git git-lfs
git lfs install

git clone https://github.com/NVlabs/GR00T-WholeBodyControl.git
cd GR00T-WholeBodyControl
git lfs pull

python check_environment.py

Install training dependencies:

bash

# Isaac Lab must be installed separately.
pip install -e "gear_sonic/[training]"

Download checkpoint/sample data:

bash

pip install huggingface_hub
python download_from_hf.py --training

Expected layout:

text

GR00T-WholeBodyControl/
├── gear_sonic/
├── gear_sonic_deploy/
├── data/
│   ├── motion_lib_bones_seed/
│   │   ├── robot/
│   │   └── robot_filtered/
│   └── smpl_filtered/
├── sample_data/
│   ├── robot_filtered/
│   └── smpl_filtered/
└── logs_rl/

4.3 Prepare Motion Data

Convert Bones-SEED CSV To motion_lib

BONES-SEED is a large motion dataset with G1 MuJoCo trajectories according to the repo/docs. SONIC expects motion data in motion_lib PKL format.

bash

python gear_sonic/data_process/convert_soma_csv_to_motion_lib.py \
  --input /path/to/bones_seed/g1/csv/ \
  --output data/motion_lib_bones_seed/robot \
  --fps 30 \
  --fps_source 120 \
  --individual \
  --num_workers 16

Filter Motions G1 Cannot Perform

The docs filter motions involving furniture, vehicles, acrobatics, and elevated surfaces. They note that about 8.7% are removed, leaving roughly 130K out of 142K motions.

bash

python gear_sonic/data_process/filter_and_copy_bones_data.py \
  --source data/motion_lib_bones_seed/robot \
  --dest data/motion_lib_bones_seed/robot_filtered \
  --workers 16

Dry run:

bash

python gear_sonic/data_process/filter_and_copy_bones_data.py \
  --source data/motion_lib_bones_seed/robot \
  --dest /tmp/robot_filtered_preview \
  --workers 16 \
  --dry-run

Custom filters:

bash

python gear_sonic/data_process/filter_and_copy_bones_data.py \
  --source data/motion_lib_bones_seed/robot \
  --dest data/motion_lib_bones_seed/robot_filtered \
  --workers 16 \
  --add-keywords "jump_from_table" "cartwheel"

SOMA Encoder For `sonic_bones_seed`

The default sonic_release config uses G1, teleop, and SMPL. The sonic_bones_seed config adds a SOMA encoder:

bash

python gear_sonic/data_process/extract_soma_joints_from_bvh.py \
  --input /path/to/bones_seed/bvh/ \
  --output data/motion_lib_bones_seed/soma \
  --fps 30 \
  --num_workers 16 \
  --skip_existing

python gear_sonic/data_process/filter_and_copy_bones_data.py \
  --source data/motion_lib_bones_seed/soma \
  --dest data/motion_lib_bones_seed/soma_filtered \
  --workers 16

Four-encoder layout:

text

data/
├── motion_lib_bones_seed/
│   ├── robot_filtered/
│   └── soma_filtered/
└── smpl_filtered/

4.4 Train Or Fine-Tune SONIC

Debug With Sample Data

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=16 \
  headless=True \
  ++manager_env.commands.motion.motion_lib_cfg.motion_file=sample_data/robot_filtered \
  ++manager_env.commands.motion.motion_lib_cfg.smpl_motion_file=sample_data/smpl_filtered \
  ++algo.config.num_learning_iterations=100

With rendering:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=16 \
  headless=False \
  ++algo.config.num_learning_iterations=100

Full Training Or Fine-Tuning

Train from motion library:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=4096 \
  headless=True \
  ++manager_env.commands.motion.motion_lib_cfg.motion_file=data/motion_lib_bones_seed/robot_filtered \
  ++manager_env.commands.motion.motion_lib_cfg.smpl_motion_file=data/smpl_filtered

Fine-tune from released checkpoint:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  +checkpoint=sonic_release/last.pt \
  num_envs=4096 \
  headless=True \
  ++manager_env.commands.motion.motion_lib_cfg.motion_file=data/motion_lib_bones_seed/robot_filtered \
  ++manager_env.commands.motion.motion_lib_cfg.smpl_motion_file=data/smpl_filtered

Single-node 8 GPU:

bash

accelerate launch --num_processes=8 gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=4096 \
  headless=True

Multi-node:

bash

accelerate launch \
  --multi_gpu \
  --num_machines=8 \
  --num_processes=64 \
  --machine_rank=$MACHINE_RANK \
  --main_process_ip=$MASTER_ADDR \
  --main_process_port=$MASTER_PORT \
  gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=4096 \
  headless=True

Offline W&B:

bash

WANDB_MODE=offline python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=16 headless=True

Disable W&B:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  use_wandb=false \
  num_envs=16 headless=True

Output:

text

logs_rl/TRL_G1_Track/<experiment_name>-<timestamp>/
├── model_step_002000.pt
├── model_step_004000.pt
├── config.yaml
└── wandb/

4.5 Monitoring And Evaluation

Track:

Metric	Good range from docs	Meaning
`rewards/total`	3.0+	Total reward.
`rewards/anchor_pos_err`	< 0.15 m	Root position tracking error.
`rewards/body_pos_err`	< 0.10 m	Body position tracking error.
`throughput/fps`	around 4000+	Training throughput.
`success_rate`	> 0.97	Motions tracked without early termination.
`mpjpe_l`	< 30 mm	Local per-joint position error.
`mpjpe_g`	< 200 mm	Global per-joint position error.

Replay reference motions:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  ++replay=True \
  num_envs=4 \
  headless=False

Evaluate metrics:

bash

python gear_sonic/eval_agent_trl.py \
  +checkpoint=/path/to/model_step_100000.pt \
  +headless=True \
  ++eval_callbacks=im_eval \
  ++run_eval_loop=False \
  ++num_envs=128 \
  "+manager_env/terminations=tracking/eval" \
  "++manager_env.commands.motion.motion_lib_cfg.max_unique_motions=512"

Render videos:

bash

python gear_sonic/eval_agent_trl.py \
  +checkpoint=/path/to/model_step_100000.pt \
  +headless=True \
  ++eval_callbacks=im_eval \
  ++run_eval_loop=False \
  ++num_envs=8 \
  ++manager_env.config.render_results=True \
  "++manager_env.config.save_rendering_dir=/tmp/renders" \
  ++manager_env.config.env_spacing=10.0 \
  "~manager_env/recorders=empty" "+manager_env/recorders=render"

For the released checkpoint, docs say to override embedded training paths:

bash

"++manager_env.commands.motion.motion_lib_cfg.motion_file=data/motion_lib_bones_seed/robot_filtered"

4.6 Export ONNX For C++ Deployment

bash

python gear_sonic/eval_agent_trl.py \
  +checkpoint=/path/to/model_step_100000.pt \
  +headless=True \
  ++num_envs=1 \
  +export_onnx_only=true

Output under exported/:

text

exported/
├── *_smpl.onnx
├── *_g1.onnx
├── *_teleop.onnx
├── *_encoder.onnx
└── *_decoder.onnx

Choose by input modality:

ONNX	Use when
`*_smpl.onnx`	SMPL pose-estimation input.
`*_g1.onnx`	Robot joint/reference input.
`*_teleop.onnx`	VR teleop tracking input.
`_encoder.onnx` + `_decoder.onnx`	Custom integration.

After export, the C++ deployment stack may need TensorRT engine generation. If TensorRT fails, check TensorRT_ROOT, CUDA version, and deployment docs.

Common Errors And Fixes

Error	Cause	Fix
Mesh/checkpoint is a tiny text file	Git LFS missing	`git lfs install && git lfs pull`.
`ModuleNotFoundError: isaaclab`	Not inside Isaac Lab env	Install and activate the correct Isaac Lab environment.
Empty motion library	Wrong `motion_file`/`smpl_motion_file` path	Use absolute paths while debugging.
Body name mismatch	Robot body names differ from config	Fix robot config/body compatibility.
Robot explodes/falls on first frame	Motion scale/order/quaternion issue	Replay motion before training; inspect joint order and frame convention.
Training is too slow	Too few envs/GPUs or rendering enabled	Use headless, increase `num_envs`, use multi-GPU.
Released checkpoint eval path error	Embedded internal paths	Override `motion_file` as documented.
TensorRT build fails	Wrong TensorRT/CUDA setup	Set `TensorRT_ROOT`, use compatible versions.

Done-Correct Criteria

You are done if:

check_environment.py passes.
git lfs pull downloads real assets.
Motion data conversion and filtering finish.
100 debug iterations run without NaN.
Checkpoints appear under logs_rl/TRL_G1_Track/....
Evaluation metrics run.
Rendered video does not show immediate robot collapse.
ONNX export writes files under exported/.

Sources

GR00T Whole-Body VLA: Training SONIC Controller

Disclosure: This article may contain affiliate or referral links. If you buy or sign up through those links, VnRobo may earn a commission or service credit.

This article covers the SONIC data and training pipeline:

text

Bones-SEED / robot motion / SMPL / SOMA
  -> convert motion data
  -> filter impossible motions
  -> PPO training in Isaac Lab
  -> evaluate tracking metrics
  -> export ONNX
  -> C++ deploy stack

4.1 Goal

By the end, you should be able to:

Prepare motion-reference data for SONIC.
Run local debug training.
Understand when you need 8 GPUs, 64+ GPUs, or just a released checkpoint.
Evaluate checkpoints with tracking metrics.
Export ONNX for C++ deployment.
Diagnose common failures: Git LFS, Isaac Lab env, body-name mismatch, motion file paths, TensorRT.

Important distinction:

Layer	Data	Goal
GR00T VLA	GR00T-LeRobot: image/state/action/language	Learn a policy from observation and instruction.
SONIC WBC	motion_lib PKL, SMPL, SOMA, G1 trajectories	Learn a whole-body motion/controller foundation.
C++ deploy	ONNX/TensorRT + ZMQ + robot state	Run realtime control in sim or on real hardware.

When Do You Need To Train SONIC?

Not every project needs custom SONIC training. For manipulation tasks on Unitree G1 using the official workflow, the faster path is usually:

text

real/sim/public LeRobot data
  -> fine-tune GR00T VLA with UNITREE_G1_SONIC
  -> use the released SONIC checkpoint/deploy stack
  -> infer with PolicyServer + launch_inference.py

Train or fine-tune SONIC yourself only when:

You need a controller for a non-G1 embodiment or changed joint/body layout.
The released checkpoint does not cover the motion skills you need.
You are researching controller foundations, motion tracking, or SOMA/SMPL encoders.
You need custom ONNX/TensorRT export for an internal C++ deployment stack.

If your main goal is VLA training from sim or real data, finish Parts 1-3 first. Part 4 is a lower-level controller layer with much higher compute cost.

4.2 Environment And Hardware

Goal	Practical GPU/VRAM
Load env + replay motion	1 GPU, 16-24 GB
Debug 100 training iterations	1 GPU, 24-48 GB
Single-node training	8 GPUs, more VRAM is better
Serious convergence	64+ GPUs according to docs
ONNX export	1 GPU or CPU depending on env/checkpoint

Setup:

bash

sudo apt update
sudo apt install -y git git-lfs
git lfs install

git clone https://github.com/NVlabs/GR00T-WholeBodyControl.git
cd GR00T-WholeBodyControl
git lfs pull

python check_environment.py

Install training dependencies:

bash

# Isaac Lab must be installed separately.
pip install -e "gear_sonic/[training]"

Download checkpoint/sample data:

bash

pip install huggingface_hub
python download_from_hf.py --training

Expected layout:

text

GR00T-WholeBodyControl/
├── gear_sonic/
├── gear_sonic_deploy/
├── data/
│   ├── motion_lib_bones_seed/
│   │   ├── robot/
│   │   └── robot_filtered/
│   └── smpl_filtered/
├── sample_data/
│   ├── robot_filtered/
│   └── smpl_filtered/
└── logs_rl/

4.3 Prepare Motion Data

Convert Bones-SEED CSV To motion_lib

BONES-SEED is a large motion dataset with G1 MuJoCo trajectories according to the repo/docs. SONIC expects motion data in motion_lib PKL format.

bash

python gear_sonic/data_process/convert_soma_csv_to_motion_lib.py \
  --input /path/to/bones_seed/g1/csv/ \
  --output data/motion_lib_bones_seed/robot \
  --fps 30 \
  --fps_source 120 \
  --individual \
  --num_workers 16

Filter Motions G1 Cannot Perform

The docs filter motions involving furniture, vehicles, acrobatics, and elevated surfaces. They note that about 8.7% are removed, leaving roughly 130K out of 142K motions.

bash

python gear_sonic/data_process/filter_and_copy_bones_data.py \
  --source data/motion_lib_bones_seed/robot \
  --dest data/motion_lib_bones_seed/robot_filtered \
  --workers 16

Dry run:

bash

python gear_sonic/data_process/filter_and_copy_bones_data.py \
  --source data/motion_lib_bones_seed/robot \
  --dest /tmp/robot_filtered_preview \
  --workers 16 \
  --dry-run

Custom filters:

bash

python gear_sonic/data_process/filter_and_copy_bones_data.py \
  --source data/motion_lib_bones_seed/robot \
  --dest data/motion_lib_bones_seed/robot_filtered \
  --workers 16 \
  --add-keywords "jump_from_table" "cartwheel"

SOMA Encoder For `sonic_bones_seed`

The default sonic_release config uses G1, teleop, and SMPL. The sonic_bones_seed config adds a SOMA encoder:

bash

python gear_sonic/data_process/extract_soma_joints_from_bvh.py \
  --input /path/to/bones_seed/bvh/ \
  --output data/motion_lib_bones_seed/soma \
  --fps 30 \
  --num_workers 16 \
  --skip_existing

python gear_sonic/data_process/filter_and_copy_bones_data.py \
  --source data/motion_lib_bones_seed/soma \
  --dest data/motion_lib_bones_seed/soma_filtered \
  --workers 16

Four-encoder layout:

text

data/
├── motion_lib_bones_seed/
│   ├── robot_filtered/
│   └── soma_filtered/
└── smpl_filtered/

4.4 Train Or Fine-Tune SONIC

Debug With Sample Data

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=16 \
  headless=True \
  ++manager_env.commands.motion.motion_lib_cfg.motion_file=sample_data/robot_filtered \
  ++manager_env.commands.motion.motion_lib_cfg.smpl_motion_file=sample_data/smpl_filtered \
  ++algo.config.num_learning_iterations=100

With rendering:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=16 \
  headless=False \
  ++algo.config.num_learning_iterations=100

Full Training Or Fine-Tuning

Train from motion library:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=4096 \
  headless=True \
  ++manager_env.commands.motion.motion_lib_cfg.motion_file=data/motion_lib_bones_seed/robot_filtered \
  ++manager_env.commands.motion.motion_lib_cfg.smpl_motion_file=data/smpl_filtered

Fine-tune from released checkpoint:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  +checkpoint=sonic_release/last.pt \
  num_envs=4096 \
  headless=True \
  ++manager_env.commands.motion.motion_lib_cfg.motion_file=data/motion_lib_bones_seed/robot_filtered \
  ++manager_env.commands.motion.motion_lib_cfg.smpl_motion_file=data/smpl_filtered

Single-node 8 GPU:

bash

accelerate launch --num_processes=8 gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=4096 \
  headless=True

Multi-node:

bash

accelerate launch \
  --multi_gpu \
  --num_machines=8 \
  --num_processes=64 \
  --machine_rank=$MACHINE_RANK \
  --main_process_ip=$MASTER_ADDR \
  --main_process_port=$MASTER_PORT \
  gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=4096 \
  headless=True

Offline W&B:

bash

WANDB_MODE=offline python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  num_envs=16 headless=True

Disable W&B:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  use_wandb=false \
  num_envs=16 headless=True

Output:

text

logs_rl/TRL_G1_Track/<experiment_name>-<timestamp>/
├── model_step_002000.pt
├── model_step_004000.pt
├── config.yaml
└── wandb/

4.5 Monitoring And Evaluation

Track:

Metric	Good range from docs	Meaning
`rewards/total`	3.0+	Total reward.
`rewards/anchor_pos_err`	< 0.15 m	Root position tracking error.
`rewards/body_pos_err`	< 0.10 m	Body position tracking error.
`throughput/fps`	around 4000+	Training throughput.
`success_rate`	> 0.97	Motions tracked without early termination.
`mpjpe_l`	< 30 mm	Local per-joint position error.
`mpjpe_g`	< 200 mm	Global per-joint position error.

Replay reference motions:

bash

python gear_sonic/train_agent_trl.py \
  +exp=manager/universal_token/all_modes/sonic_release \
  ++replay=True \
  num_envs=4 \
  headless=False

Evaluate metrics:

bash

python gear_sonic/eval_agent_trl.py \
  +checkpoint=/path/to/model_step_100000.pt \
  +headless=True \
  ++eval_callbacks=im_eval \
  ++run_eval_loop=False \
  ++num_envs=128 \
  "+manager_env/terminations=tracking/eval" \
  "++manager_env.commands.motion.motion_lib_cfg.max_unique_motions=512"

Render videos:

bash

python gear_sonic/eval_agent_trl.py \
  +checkpoint=/path/to/model_step_100000.pt \
  +headless=True \
  ++eval_callbacks=im_eval \
  ++run_eval_loop=False \
  ++num_envs=8 \
  ++manager_env.config.render_results=True \
  "++manager_env.config.save_rendering_dir=/tmp/renders" \
  ++manager_env.config.env_spacing=10.0 \
  "~manager_env/recorders=empty" "+manager_env/recorders=render"

For the released checkpoint, docs say to override embedded training paths:

bash

"++manager_env.commands.motion.motion_lib_cfg.motion_file=data/motion_lib_bones_seed/robot_filtered"

4.6 Export ONNX For C++ Deployment

bash

python gear_sonic/eval_agent_trl.py \
  +checkpoint=/path/to/model_step_100000.pt \
  +headless=True \
  ++num_envs=1 \
  +export_onnx_only=true

Output under exported/:

text

exported/
├── *_smpl.onnx
├── *_g1.onnx
├── *_teleop.onnx
├── *_encoder.onnx
└── *_decoder.onnx

Choose by input modality:

ONNX	Use when
`*_smpl.onnx`	SMPL pose-estimation input.
`*_g1.onnx`	Robot joint/reference input.
`*_teleop.onnx`	VR teleop tracking input.
`_encoder.onnx` + `_decoder.onnx`	Custom integration.

After export, the C++ deployment stack may need TensorRT engine generation. If TensorRT fails, check TensorRT_ROOT, CUDA version, and deployment docs.

Common Errors And Fixes

Error	Cause	Fix
Mesh/checkpoint is a tiny text file	Git LFS missing	`git lfs install && git lfs pull`.
`ModuleNotFoundError: isaaclab`	Not inside Isaac Lab env	Install and activate the correct Isaac Lab environment.
Empty motion library	Wrong `motion_file`/`smpl_motion_file` path	Use absolute paths while debugging.
Body name mismatch	Robot body names differ from config	Fix robot config/body compatibility.
Robot explodes/falls on first frame	Motion scale/order/quaternion issue	Replay motion before training; inspect joint order and frame convention.
Training is too slow	Too few envs/GPUs or rendering enabled	Use headless, increase `num_envs`, use multi-GPU.
Released checkpoint eval path error	Embedded internal paths	Override `motion_file` as documented.
TensorRT build fails	Wrong TensorRT/CUDA setup	Set `TensorRT_ROOT`, use compatible versions.

Done-Correct Criteria

You are done if:

check_environment.py passes.
git lfs pull downloads real assets.
Motion data conversion and filtering finish.
100 debug iterations run without NaN.
Checkpoints appear under logs_rl/TRL_G1_Track/....
Evaluation metrics run.
Rendered video does not show immediate robot collapse.
ONNX export writes files under exported/.

GR00T Whole-Body VLA: Training SONIC Controller

4.1 Goal

When Do You Need To Train SONIC?

4.2 Environment And Hardware

4.3 Prepare Motion Data

Convert Bones-SEED CSV To motion_lib

Filter Motions G1 Cannot Perform

SOMA Encoder For sonic_bones_seed

4.4 Train Or Fine-Tune SONIC

Debug With Sample Data

Full Training Or Fine-Tuning

4.5 Monitoring And Evaluation

4.6 Export ONNX For C++ Deployment

Common Errors And Fixes

Done-Correct Criteria

Related Posts

Sources

Nguyễn Anh Tuấn

Related Posts

GR00T Whole-Body VLA Data: Open Datasets

GR00T Whole-Body VLA Data: Do You Need Real Data?

GR00T Whole-Body VLA Data: Simulation Data

GR00T Whole-Body VLA: Training SONIC Controller

4.1 Goal

When Do You Need To Train SONIC?

4.2 Environment And Hardware

4.3 Prepare Motion Data

Convert Bones-SEED CSV To motion_lib

Filter Motions G1 Cannot Perform

SOMA Encoder For sonic_bones_seed

4.4 Train Or Fine-Tune SONIC

Debug With Sample Data

Full Training Or Fine-Tuning

4.5 Monitoring And Evaluation

4.6 Export ONNX For C++ Deployment

Common Errors And Fixes

Done-Correct Criteria

Related Posts

Sources

Nguyễn Anh Tuấn

Related Posts

GR00T Whole-Body VLA Data: Open Datasets

GR00T Whole-Body VLA Data: Do You Need Real Data?

GR00T Whole-Body VLA Data: Simulation Data

SOMA Encoder For `sonic_bones_seed`

SOMA Encoder For `sonic_bones_seed`