Streaming_VGGT is an open-weight model from Anh Dao. Its published files total 62.3 GB.
Completed one joint R2R + RxR epoch, 3,852 optimizer updates, seed429, from Qwen3.5-4B base revision851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a. 30,815 episodes; effective batch8 (4H100 x1episode xGAS2), full episodes/fullBPTT.
Parameters—
Context—
Weights62.2 GB
License—
AccessOpen weights
Monthly Downloads—
Model Card
Completed one joint R2R + RxR epoch, 3,852 optimizer updates, seed429, from Qwen3.5-4B base revision851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a. 30,815 episodes; effective batch8 (4H100 x1episode xGAS2), full episodes/fullBPTT. Actual one-epoch cosine schedule,116LRwarmup steps; auxiliary ramp771updates. G1/epoch-1/ is the inference export: native Window8 text policy with canonical action-text history, without a dual lane or episodic bank. Frozen online FP32 VGGT-Omega supervises sampled visual tokens during training only. The student projector follows SpatialForcing's affineLayerNorm2560 -> Linear4096 -> GELU -> Linear2048 architecture. No feature pre-extraction or PCA is used. All retained…
Excerpt from the card by Anh Dao.
Identity and Version
- Repository
- anhdao69/Streaming_VGGT
- Publisher
- Anh Dao
- Task
- Not stated by the source
- Modality
- Other
- Library
- pytorch
- Parameters
- Not stated by the source
- Languages
- rxr
- Revision
- 249464698837e6d71e275ed769fd40fa90d41d6e
- First published
- 2026-10-13
- Last updated
- 2026-10-17
Files and Weights
121 files, 62.3 GB in total. The weights are 8 files totalling 62.2 GB in bin, pt.
Weights8 files · 62.2 GB
Configuration87 files · 7.0 MB
Tokenizer12 files · 119.9 MB
Documentation3 files · 12.1 KB
Other10 files · 46.7 KB
Repository1 file · 1.9 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| A0/epoch-1/pytorch_model.bin | Weights | 10.4 GB | 5e63563ddb89 |
| A1/epoch-1/pytorch_model.bin | Weights | 10.4 GB | 0d9078ffd00e |
| G1/epoch-1/pytorch_model.bin | Weights | 10.4 GB | a620ab17b1c0 |
| G1/training_auxiliary_projector.pt | Weights | 37.8 MB | 407d6852a335 |
| G3/epoch-1/pytorch_model.bin | Weights | 10.4 GB | a71be6b5291b |
| G3/training_auxiliary_state_head.pt | Weights | 21.1 MB | 66d9c8e18ed3 |
| G3_M5/epoch-1/pytorch_model.bin | Weights | 10.4 GB | 4e2d2095325f |
| R1/epoch-1/pytorch_model.bin | Weights | 10.4 GB | 2b25416a500d |
| A0/epoch-1/config.json | Configuration | 2.9 KB | — |
| A0/epoch-1/export_info.json | Configuration | 55 B | — |
| A0/epoch-1/navigation.json | Configuration | 3.1 KB | — |
| A0/epoch-1/processor_config.json | Configuration | 1.2 KB | — |
| A0/epoch-1/reload_reference.json | Configuration | 122 B | — |
| A0/epoch-1/research_training.json | Configuration | 320 B | — |
| A0/provenance/epoch-1.json | Configuration | 396.4 KB | — |
| A0/provenance/publication_validation.json | Configuration | 579 B | — |
| A0/provenance/resolved_config.json | Configuration | 2.5 KB | — |
| A0/provenance/run_timing.json | Configuration | 353 B | — |
| A0/provenance/source_manifest.json | Configuration | 18.7 KB | — |
| A0/provenance/supervisor_resources.json | Configuration | 1.9 KB | — |
| A0/provenance/train_results.json | Configuration | 191 B | — |
| A0/provenance/trainer_state.json | Configuration | 719.4 KB | — |
| A0/provenance/training_navigation.json | Configuration | 3.4 KB | — |
| A1/epoch-1/config.json | Configuration | 2.9 KB | — |
| A1/epoch-1/export_info.json | Configuration | 55 B | — |
| A1/epoch-1/navigation.json | Configuration | 3.1 KB | — |
| A1/epoch-1/processor_config.json | Configuration | 1.2 KB | — |
| A1/epoch-1/reload_reference.json | Configuration | 125 B | — |
| A1/epoch-1/research_training.json | Configuration | 1.5 KB | — |
| A1/provenance/epoch-1.json | Configuration | 396.6 KB | — |
| A1/provenance/publication_validation.json | Configuration | 579 B | — |
| A1/provenance/qualification_pass.json | Configuration | 28.5 KB | — |
| A1/provenance/resolved_config.json | Configuration | 3.6 KB | — |
| A1/provenance/run_timing.json | Configuration | 353 B | — |
| A1/provenance/source_manifest.json | Configuration | 18.7 KB | — |
| A1/provenance/supervisor_resources.json | Configuration | 1.9 KB | — |
| A1/provenance/train_results.json | Configuration | 191 B | — |
| A1/provenance/trainer_state.json | Configuration | 719.1 KB | — |
| A1/provenance/training_navigation.json | Configuration | 4.6 KB | — |
| G1/epoch-1/config.json | Configuration | 2.9 KB | — |
| G1/epoch-1/export_info.json | Configuration | 55 B | — |
| G1/epoch-1/navigation.json | Configuration | 2.9 KB | — |
| G1/epoch-1/processor_config.json | Configuration | 1.2 KB | — |
| G1/epoch-1/reload_reference.json | Configuration | 126 B | — |
| G1/epoch-1/research_training.json | Configuration | 1.4 KB | — |
| G1/provenance/epoch-1.json | Configuration | 396.6 KB | — |
| G1/provenance/production_launch.json | Configuration | 647 B | — |
| G1/provenance/publication_validation.json | Configuration | 300 B | — |
| G1/provenance/resolved_config.json | Configuration | 3.3 KB | — |
| G1/provenance/run_timing.json | Configuration | 353 B | — |
| G1/provenance/source_manifest.json | Configuration | 35.8 KB | — |
| G1/provenance/train_results.json | Configuration | 189 B | — |
| G1/provenance/trainer_state.json | Configuration | 719.3 KB | — |
| G1/provenance/training_navigation.json | Configuration | 4.3 KB | — |
| G3/epoch-1/config.json | Configuration | 2.9 KB | — |
| G3/epoch-1/export_info.json | Configuration | 55 B | — |
| G3/epoch-1/navigation.json | Configuration | 2.9 KB | — |
| G3/epoch-1/processor_config.json | Configuration | 1.2 KB | — |
| G3/epoch-1/reload_reference.json | Configuration | 124 B | — |
| G3/epoch-1/research_training.json | Configuration | 1.7 KB | — |
| G3/provenance/epoch-1.json | Configuration | 396.5 KB | — |
| G3/provenance/production_launch.json | Configuration | 1.0 KB | — |
| G3/provenance/publication_validation.json | Configuration | 518 B | — |
| G3/provenance/resolved_config.json | Configuration | 3.6 KB | — |
| G3/provenance/run_timing.json | Configuration | 354 B | — |
| G3/provenance/source_manifest.json | Configuration | 38.3 KB | — |
| G3/provenance/train_results.json | Configuration | 192 B | — |
| G3/provenance/trainer_state.json | Configuration | 719.1 KB | — |
| G3/provenance/training_navigation.json | Configuration | 4.6 KB | — |
| G3_M5/SHA256SUMS.json | Configuration | 2.5 KB | — |
| G3_M5/epoch-1/config.json | Configuration | 2.9 KB | — |
| G3_M5/epoch-1/navigation.json | Configuration | 3.9 KB | — |
| G3_M5/epoch-1/processor_config.json | Configuration | 1.2 KB | — |
| G3_M5/epoch-1/research_training.json | Configuration | 1.7 KB | — |
| G3_M5/provenance/completion_resources.json | Configuration | 1.9 KB | — |
| G3_M5/provenance/epoch-1.json | Configuration | 396.6 KB | — |
| G3_M5/provenance/publication_validation.json | Configuration | 1.3 KB | — |
| G3_M5/provenance/qualification_pass.json | Configuration | 35.6 KB | — |
| G3_M5/provenance/resolved_config.json | Configuration | 3.8 KB | — |
| G3_M5/provenance/run_timing.json | Configuration | 353 B | — |
| G3_M5/provenance/source_and_completion.json | Configuration | 1.8 KB | — |
| G3_M5/provenance/trainer_state.json | Configuration | 719.2 KB | — |
| R1/SHA256SUMS.json | Configuration | 2.3 KB | — |
| R1/epoch-1/config.json | Configuration | 2.9 KB | — |
| R1/epoch-1/navigation.json | Configuration | 3.7 KB | — |
| R1/epoch-1/processor_config.json | Configuration | 1.2 KB | — |
| R1/epoch-1/research_training.json | Configuration | 7.7 KB | — |
| R1/provenance/completion_resources.json | Configuration | 1.9 KB | — |
| R1/provenance/epoch-1.json | Configuration | 396.5 KB | — |
| R1/provenance/publication_validation.json | Configuration | 3.1 KB | — |
| R1/provenance/resolved_config.json | Configuration | 9.9 KB | — |
| R1/provenance/run_timing.json | Configuration | 352 B | — |
| R1/provenance/source_and_completion.json | Configuration | 1.8 KB | — |
| R1/provenance/trainer_state.json | Configuration | 719.2 KB | — |
| SHA256SUMS.json | Configuration | 8.8 KB | — |
| G3_M5/README.md | Documentation | 2.3 KB | — |
| R1/README.md | Documentation | 1.9 KB | — |
| README.md | Documentation | 7.9 KB | — |
| A0/epoch-1/chat_template.jinja | Other | 7.8 KB | — |
| A0/provenance/source.commit | Other | 41 B | — |
| A1/epoch-1/chat_template.jinja | Other | 7.8 KB | — |
| A1/provenance/source.commit | Other | 41 B | — |
| G1/epoch-1/chat_template.jinja | Other | 7.8 KB | — |
| G1/provenance/source.commit | Other | 41 B | — |
| G3/epoch-1/chat_template.jinja | Other | 7.8 KB | — |
| G3/provenance/source.commit | Other | 41 B | — |
| G3_M5/epoch-1/chat_template.jinja | Other | 7.8 KB | — |
| R1/epoch-1/chat_template.jinja | Other | 7.8 KB | — |
| .gitattributes | Repository | 1.9 KB | — |
| A0/epoch-1/tokenizer.json | Tokenizer | 20.0 MB | 06b9509352d2 |
| A0/epoch-1/tokenizer_config.json | Tokenizer | 1.2 KB | — |
| A1/epoch-1/tokenizer.json | Tokenizer | 20.0 MB | 06b9509352d2 |
| A1/epoch-1/tokenizer_config.json | Tokenizer | 1.2 KB | — |
| G1/epoch-1/tokenizer.json | Tokenizer | 20.0 MB | 06b9509352d2 |
| G1/epoch-1/tokenizer_config.json | Tokenizer | 1.2 KB | — |
| G3/epoch-1/tokenizer.json | Tokenizer | 20.0 MB | 06b9509352d2 |
| G3/epoch-1/tokenizer_config.json | Tokenizer | 1.2 KB | — |
| G3_M5/epoch-1/tokenizer.json | Tokenizer | 20.0 MB | 06b9509352d2 |
| G3_M5/epoch-1/tokenizer_config.json | Tokenizer | 1.2 KB | — |
| R1/epoch-1/tokenizer.json | Tokenizer | 20.0 MB | 06b9509352d2 |
| R1/epoch-1/tokenizer_config.json | Tokenizer | 1.2 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 62.2 GB
Download from Anh Dao
Released by Anh Dao through its official repository on Hugging Face.
Built From
- Derived from Qwen/Qwen3.5-4B
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 62.2 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.