docs: SOURCES.md add ## 10 UniT + ## 11 TacQuad sections, update all stale row counts (fota_unlabeled, RTM, sims, feelanyforce, etc), add Channel-order normalization section; README: update header tagline frame count

README.md

@@ -94,7 +94,7 @@ configs:
 
 ![overview](assets/combined_overview.png)
 
-A unified, parquet-native collection of **~830K [GelSight Mini](https://www.gelsight.com/gelsightmini/) tactile RGB frames** for self-supervised representation learning. **Ten** public datasets are aggregated under one schema, each filtered through a unified area+intensity contact rule with **per-domain thresholds** (I_min = 12 for real subsets, I_min = 10 for sim subsets, 1.5 % background-diversity keep rate) + per-capture phash dedupe:
+A unified, parquet-native collection of **~833K [GelSight Mini](https://www.gelsight.com/gelsightmini/) tactile RGB frames** for self-supervised representation learning. **Ten** public datasets are aggregated under one schema, each filtered through a unified area+intensity contact rule with **per-domain thresholds** (I_min = 12 for real subsets, I_min = 10 for sim subsets, 1.5 % background-diversity keep rate) + per-capture phash dedupe:
 - **~536K real-world frames** from 10 sources (FoTA labeled+unlabeled, 3DCal, FEATS, GelSLAM, TactileTracking, Real Tactile MNIST, FeelAnyForce, UniT, TacQuad)
 - **400 K simulated frames** from 2 Mini-calibrated Taxim renders (Sim Tactile MNIST, Sim Starstruck)
 - Plus a companion **CC-BY-NC extension** ([`yxma/gelsight-mini-pretrain-nc`](https://huggingface.co/datasets/yxma/gelsight-mini-pretrain-nc)) with another **~66 K** real frames from Meta Sparsh

SOURCES.md

@@ -62,8 +62,8 @@ are excluded.
 
 | Subset           | Frames  | Resolution | Markered / Markerless | Splits           |
 |------------------|--------:|------------|----------------------:|------------------|
-| `fota_labeled`   | 29,494  | 640 × 480  | 10,025 / 19,469       | train 23,588 · val 5,906 |
-| `fota_unlabeled` | 516,523 | 640 × 480  | 190,816 / 325,707     | train 413,193 · val 103,330 |
+| `fota_labeled`   | 26,394  | 640 × 480  | 10,025 / 16,369       | train 21,139 · val 5,255    |
+| `fota_unlabeled` | 66,761  | 640 × 480  | 36,592 / 30,169       | train 66,761 (train-only)    |
 
 13 distinct contact objects, 5 initial-pose indices, 15,148 unique
 (object, pose, side) captures. End-effector range
@@ -429,7 +429,7 @@ empty drops.
 
 | Subset         | Frames  | Resolution | Markered | Unique objects |
 |----------------|--------:|------------|---------:|---------------:|
-| `feelanyforce` | 50,997  | 320 × 240  | 0        | 42             |
+| `feelanyforce` | 48,197  | 320 × 240  | 0        | 42             |
 
 **40 random samples:**
 
@@ -470,7 +470,7 @@ per touch, already at peak contact — no video decode needed).
 
 | Subset                | Frames  | Resolution | Domain |
 |-----------------------|--------:|------------|--------|
-| `sim_tactile_mnist`   | 200,000 | 320 × 240  | sim    |
+| `sim_tactile_mnist`   | 150,601 | 320 × 240  | sim    |
 
 **40 random samples:**
 
@@ -499,8 +499,8 @@ filter, no dedupe, capped at 200 K kept frames.
 
 | Subset           | Split | Frames  | Resolution | Domain |
 |------------------|-------|--------:|------------|--------|
-| `sim_starstruck` | train | 46,400  | 320 × 240  | sim    |
-| `sim_starstruck` | test  | 153,600 | 320 × 240  | sim    |
+| `sim_starstruck` | train | 150,000 | 320 × 240  | sim    |
+| `sim_starstruck` | test  |  16,104 | 320 × 240  | sim    |
 | **Total**        |       | **200,000** |        |        |
 
 **40 random samples:**
@@ -509,18 +509,100 @@ filter, no dedupe, capped at 200 K kept frames.
 
 ---
 
+## 10 · UniT — continuous 3D-pose tracking (`unit`)
+
+**Intro.** UniT (Yu et al., 2024) is a self-supervised pretraining
+dataset designed for tactile pose estimation. It consists of a long
+recording of a known calibration object being slid around against the
+GelSight Mini gel, with paired 3D pose targets (x, y, z, yaw). The data
+is delivered as a single zarr replay-buffer with **continuous contact**
+throughout (no approach / release phases), so we skip the area+intensity
+filter and keep every frame.
+
+**Source release.**
+- 🔗 [UniT repo](https://github.com/ZeyuYong/UniT) (zarr replay-buffer
+  format)
+- 📜 License: BSD-3-Clause-style permissive (per upstream repo metadata)
+
+**Original format.** A `replay_buffer.zarr` with two arrays:
+- `data/tactile_image` — `(11340, 240, 320, 3)` uint8, zstd-compressed,
+  9-frame chunks
+- `data/3Dpose` — `(11340, 4)` float32 — (x_mm, y_mm, z_mm, yaw)
+
+**How we processed it.** Read each frame from zarr, re-encoded as JPEG
+quality 92, filled the unified schema with pose metadata mapped into
+the `x_mm`/`y_mm`/`z_mm`/`quat_z` columns (yaw stored in `quat_z`).
+Wrote `unit/train-00000-of-00001.parquet`. **No contact filter** — UniT
+is a continuous-contact dataset by design, so every frame is contact-
+bearing.
+
+**Stats after processing.**
+
+| Subset | Frames | Resolution | Markered / Markerless | Splits |
+|--------|-------:|------------|----------------------:|--------|
+| `unit` | 11,340 | 320 × 240  | 0 / 11,340            | train 11,340 |
+
+**40 random samples:**
+
+![unit](assets/samples_40_unit.png)
+
+---
+
+## 11 · TacQuad — quad-sensor benchmark, Mini stream (`tacquad`)
+
+**Intro.** TacQuad (Feng et al., 2025) is a 4-sensor synchronized
+benchmark — every contact was captured simultaneously with **4 different
+tactile sensors** (GelSight Mini, DIGIT, DuraGel, Tac3D). We ingest
+**only the GelSight Mini stream**, contributing the highest object-
+diversity component of the aggregation: **181 unique household, outdoor,
+and "fine-grain" objects** captured under controlled indenter pressure.
+
+**Source release.**
+- 🔗 [TacQuad / AnyTouch project](https://github.com/anytouch-project)
+- 📜 License: CC-BY-4.0
+
+**Original format.** Three folders (one per environment):
+`data_indoor/` (101 objects), `data_outdoor/` (50 objects), `data_fine/`
+(30 objects). Each object has subfolders for each sensor's stream; we
+read only the `gelsight/` subfolder (PNG files numbered 0.png, 1.png…).
+
+**How we processed it.** Walked each environment, read all `gelsight/*.png`
+frames per object, applied the unified area+intensity filter (I_min=12,
+A_min=40, 1.5 % bg-diversity). Each environment becomes one split. Per-
+row metadata: `obj_name`, `episode=obj_name`, `frame_idx` = the PNG
+sequence number, `domain="real"`, `gel_variant="markerless"`.
+
+**Stats after processing.**
+
+| Subset    | Split        | Frames | Resolution | Domain |
+|-----------|--------------|-------:|------------|--------|
+| `tacquad` | data_indoor  |  5,363 | 320 × 240  | real   |
+| `tacquad` | data_outdoor |  3,934 | 320 × 240  | real   |
+| `tacquad` | data_fine    |  2,898 | 320 × 240  | real   |
+| **Total** |              | **12,195** |        |        |
+
+24,866 raw Mini frames in the upstream zips, retention ~49 % after the
+contact filter. Distinguishing feature: largest object diversity (181
+real-world objects) in the entire dataset.
+
+**40 random samples:**
+
+![tacquad](assets/samples_40_tacquad.png)
+
+---
+
 ## Investigated but not included
 
-- **TacQuad** (AnyTouch, ICLR 2025) — the released zip contains
-  pre-processed 120×160 PNGs in non-Mini aspect ratio. The CSV metadata
-  doesn't match the released folder structure (no per-sensor sub-folders
-  visible). The data is shaped for the AnyTouch paper's own 224×224
-  training pipeline, not raw Mini RGB pretraining. Deferred.
-- **UniT** — undocumented license, undocumented crop dimensions, ~300 GB
-  manipulation data overshadows the ~16 K usable pretraining frames.
-  Deferred.
 - **`facebook/gelsight-force-estimation`** — CC-BY-NC-4.0 license is
-  incompatible with our CC-BY-4.0 aggregation. Not included.
+  incompatible with this CC-BY-4.0 repo. (Has been moved to our
+  companion NC repo
+  [`yxma/gelsight-mini-pretrain-nc`](https://huggingface.co/datasets/yxma/gelsight-mini-pretrain-nc),
+  where it is now superseded by `sparsh` — the same data at a larger
+  upstream snapshot.)
+- **TVL — Touch-Vision-Language** (Yang et al. 2024) — has paired RGB +
+  caption labels. Not yet ingested. CC-BY-4.0; ~44K Mini frames available.
+- **Touch and Go** (Yang et al. 2022) — has paired natural-scene RGB.
+  Not yet ingested. CC-BY-4.0; ~13K Mini frames available.
 - **YCB-Sight** — CC-BY-SA-4.0 (viral copyleft), and the sim is not
   Mini-calibrated. Not included.
 - **TACTO**, **MidasTouch**, **DiffTactile**, generic GelSight sims —
@@ -529,25 +611,48 @@ filter, no dedupe, capped at 200 K kept frames.
 
 ---
 
+## Channel-order normalization
+
+Some upstream sources stored their RGB frames as **BGR**, which would
+have produced color-inverted tactile patterns when loaded as RGB. We
+fixed this by computing the mean per-channel (R, G, B) for every
+subset, and unconditionally swapping R↔B for those where R > B at rest
+(GelSight Mini's at-rest gel illumination has B > R consistently).
+
+The affected subsets (now corrected to RGB) were:
+- `fota_unlabeled` — globally BGR-stored upstream
+- `unit` — globally BGR-stored upstream
+- (NC repo) `faf_force_estimation` — globally BGR-stored
+- (NC repo) `sparsh` — **mixed** (some files RGB, some BGR);
+  per-image conditional swap applied based on per-image R > B
+
+The diagnostic is published as `assets/channel_order_diagnosis.json`
+(both repos). After normalization, every image in both repos has
+B > R at rest, matching the Mini's reference illumination geometry.
+
+---
+
 ## Aggregate statistics
 
 ![composition](assets/composition.png)
 
 | Subset                | Domain | Frames  | Bytes (GB) | Resolution | Markered | Markerless |
 |-----------------------|--------|--------:|-----------:|------------|---------:|-----------:|
-| `fota_labeled`        | real   | 29,494  | 0.50       | 640 × 480  | 10,025   | 19,469     |
-| `fota_unlabeled`      | real   | 516,523 | 21.09      | 640 × 480  | 190,816  | 325,707    |
-| `threedcal`           | real   | 36,270  | 0.29       | 320 × 240  | 0        | 36,270     |
-| `feats`               | real   | 16,711  | 0.27       | 320 × 240  | 16,711   | 0          |
-| `gelslam`             | real   | 89,612  | 0.82       | 320 × 240  | 0        | 89,612     |
-| `tactile_tracking`    | real   | 1,605   | 0.01       | 320 × 240  | 0        | 1,605      |
-| `real_tactile_mnist`  | real   | 16,961  | 0.08       | 320 × 240  | 0        | 16,961     |
-| `feelanyforce`        | real   | 50,997  | 0.47       | 320 × 240  | 0        | 50,997     |
-| `sim_tactile_mnist`   | sim    | 200,000 | 1.36       | 320 × 240  | 0        | 200,000    |
-| `sim_starstruck`      | sim    | 200,000 | 1.14       | 320 × 240  | 0        | 200,000    |
-| **Real total**        |        | **758,173** | **~23.5** |        | **217,552** | **540,621** |
-| **Sim total**         |        | **400,000** | **~2.5**  |        | **0**     | **400,000** |
-| **Grand total**       |        | **1,158,173** | **~26.0** |       | **217,552** | **940,621** |
+| `fota_labeled`        | real   | 26,394  | 0.46       | 640 × 480  | 10,025   | 16,369     |
+| `fota_unlabeled`      | real   | 66,761  | 3.02       | 640 × 480  | 36,592   | 30,169     |
+| `threedcal`           | real   |  6,924  | 0.05       | 320 × 240  | 0        |  6,924     |
+| `feats`               | real   | 16,969  | 0.28       | 320 × 240  | 16,969   | 0          |
+| `gelslam`             | real   | 114,019 | 1.03       | 320 × 240  | 0        | 114,019    |
+| `tactile_tracking`    | real   |  2,408  | 0.02       | 320 × 240  | 0        |  2,408     |
+| `real_tactile_mnist`  | real   | 30,956  | 0.16       | 320 × 240  | 0        | 30,956     |
+| `feelanyforce`        | real   | 48,197  | 0.45       | 320 × 240  | 0        | 48,197     |
+| `unit`                | real   | 11,340  | 0.01       | 320 × 240  | 0        | 11,340     |
+| `tacquad`             | real   | 12,195  | 0.11       | 320 × 240  | 0        | 12,195     |
+| `sim_tactile_mnist`   | sim    | 150,601 | 1.28       | 320 × 240  | 0        | 150,601    |
+| `sim_starstruck`      | sim    | 166,104 | 1.39       | 320 × 240  | 0        | 166,104    |
+| **Real total**        |        | **536,163** | **~5.59** |        | **63,586** | **472,577** |
+| **Sim total**         |        | **316,705** | **~2.67** |        | **0**     | **316,705** |
+| **Grand total**       |        | **852,868** | **~8.26** |        | **63,586** | **789,282** |
 
 ![summary pies](assets/summary_pies.png)
 
@@ -557,11 +662,8 @@ filter, no dedupe, capped at 200 K kept frames.
 
 **Gel-variant pool sizes after aggregation:**
 
-- **Markerless pool** (`fota_labeled[markerless]` + `fota_unlabeled[markerless]` +
-  `threedcal` + `gelslam` + `tactile_tracking` + `real_tactile_mnist` +
-  `feelanyforce`): **648,168 frames**
-- **Markered pool** (`feats` + `fota_labeled[markered]` +
-  `fota_unlabeled[markered]`): **217,552 frames**
+- **Markerless pool** (all `_*[markerless]` and pure-markerless subsets — `fota_*[markerless]` + `threedcal` + `gelslam` + `tactile_tracking` + `real_tactile_mnist` + `feelanyforce` + `unit` + `tacquad`): **472,577 real markerless frames** (789,282 if including sims)
+- **Markered pool** (`feats` + `fota_labeled[markered]` + `fota_unlabeled[markered]`): **63,586 frames**
 
 Filter examples: