Use default Hub download counting query file
Browse files
README.md
CHANGED
|
@@ -1,6 +1,5 @@
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
-
library_name: pytorch
|
| 4 |
pipeline_tag: robotics
|
| 5 |
tags:
|
| 6 |
- robotics
|
|
@@ -41,6 +40,7 @@ The InternW0 architecture uses:
|
|
| 41 |
|
| 42 |
| File | Purpose | Size | SHA-256 |
|
| 43 |
|---|---|---:|---|
|
|
|
|
| 44 |
| `InternW0-Base.pt` | Model-only native checkpoint | 29,410,738,430 bytes | `131d00c241c88fdd1f14cafb3d9fe7c33388f3bff55ff10397105d9fe6e5f3ac` |
|
| 45 |
| `normalization.json` | RoboTwin action/proprio normalization contract | 4,047 bytes | `0ba957c35cd08f05b55f6a4a08b7e19c784e3816dd15fdb73135a4d1b334340f` |
|
| 46 |
| `InternW0-Base.transfer.json` | Full source-to-target transfer/provenance sidecar | 307,131 bytes | `3d911a075ef680cc8edfa009b4e06a050dc000bc04182341772d9ec35207f568` |
|
|
@@ -64,7 +64,7 @@ Checkpoint summary:
|
|
| 64 |
|
| 65 |
The conversion copies shared VideoDiT, ActionDiT, OVCR, and projector weights from pretraining step 30,000. The one-row RoboTwin domain bank, action encoder/head, soft-prompt bank, and default proprio encoder are initialized for the target interface with seed 42. This distinction is recorded tensor-by-tensor in `InternW0-Base.transfer.json`.
|
| 66 |
|
| 67 |
-
Hugging Face
|
| 68 |
|
| 69 |
## Required external assets
|
| 70 |
|
|
@@ -98,7 +98,7 @@ pip install -e .
|
|
| 98 |
|
| 99 |
```bash
|
| 100 |
hf download InternRobotics/InternW0-Base \
|
| 101 |
-
InternW0-Base.pt normalization.json InternW0-Base.transfer.json \
|
| 102 |
--local-dir assets/robotwin
|
| 103 |
```
|
| 104 |
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
|
|
|
| 3 |
pipeline_tag: robotics
|
| 4 |
tags:
|
| 5 |
- robotics
|
|
|
|
| 40 |
|
| 41 |
| File | Purpose | Size | SHA-256 |
|
| 42 |
|---|---|---:|---|
|
| 43 |
+
| `config.json` | Model/download-counting manifest for the Hub | 1,129 bytes | `0cf84ddb331416e13e174a2c165ffa3eb5466ca2787a8b36d9e262bd7a489bb0` |
|
| 44 |
| `InternW0-Base.pt` | Model-only native checkpoint | 29,410,738,430 bytes | `131d00c241c88fdd1f14cafb3d9fe7c33388f3bff55ff10397105d9fe6e5f3ac` |
|
| 45 |
| `normalization.json` | RoboTwin action/proprio normalization contract | 4,047 bytes | `0ba957c35cd08f05b55f6a4a08b7e19c784e3816dd15fdb73135a4d1b334340f` |
|
| 46 |
| `InternW0-Base.transfer.json` | Full source-to-target transfer/provenance sidecar | 307,131 bytes | `3d911a075ef680cc8edfa009b4e06a050dc000bc04182341772d9ec35207f568` |
|
|
|
|
| 64 |
|
| 65 |
The conversion copies shared VideoDiT, ActionDiT, OVCR, and projector weights from pretraining step 30,000. The one-row RoboTwin domain bank, action encoder/head, soft-prompt bank, and default proprio encoder are initialized for the target interface with seed 42. This distinction is recorded tensor-by-tensor in `InternW0-Base.transfer.json`.
|
| 66 |
|
| 67 |
+
Hugging Face uses library-specific query files for model download statistics. This repository uses the Hub's default query file, `config.json`; HTTP `GET` and `HEAD` requests to that manifest are counted server-side. The badge above reads the public `downloads` field from the Hub API, which corresponds to downloads in the last 30 days. Hub analytics may lag briefly while counters refresh.
|
| 68 |
|
| 69 |
## Required external assets
|
| 70 |
|
|
|
|
| 98 |
|
| 99 |
```bash
|
| 100 |
hf download InternRobotics/InternW0-Base \
|
| 101 |
+
config.json InternW0-Base.pt normalization.json InternW0-Base.transfer.json \
|
| 102 |
--local-dir assets/robotwin
|
| 103 |
```
|
| 104 |
|