Upload folder using huggingface_hub

README.md

@@ -33,13 +33,13 @@ configs:
 
 Which open-weight AI models actually run on which consumer GPU, and what it takes
 to get them running. 2 646 model×GPU verdicts across 98 models
-and 27 cards, plus 954 full setup guides
-(16 MB of markdown) written against specific hardware.
+and 27 cards, plus 973 full setup guides
+(17 MB of markdown) written against specific hardware.
 
 This is the machine-readable form of [smeltcore.com](https://smeltcore.com). Every row carries a
 `url` back to the page it came from.
 
-Generated 2026-08-28T12:13:44+00:00 from the public read API (`https://api.smeltcore.com/api/v1`) — no private data, no
+Generated 2026-08-31T18:39:19+00:00 from the public read API (`https://api.smeltcore.com/api/v1`) — no private data, no
 credentials, reproducible by anyone.
 
 ## Configs
@@ -47,7 +47,7 @@ credentials, reproducible by anyone.
 | config | rows | what it is |
 |---|---:|---|
 | `compatibility` | 2 646 | **the point of the dataset.** One row per model × GPU, with a verdict |
-| `recipes` | 954 | full setup guides, markdown included, tagged by model / GPU / tool |
+| `recipes` | 973 | full setup guides, markdown included, tagged by model / GPU / tool |
 | `models` | 98 | the catalogue: licence, upstream repo, modality |
 | `gpus` | 27 | the cards: VRAM, vendor, series |
 | `benchmark_sources` | 158 | third-party measurements, normalised and cited |
@@ -65,10 +65,10 @@ The whole dataset turns on this column, so it is worth reading before using it.
 
 | verdict | rows | meaning |
 |---|---:|---|
-| `verified` | 953 | somebody ran it on this exact card and wrote down how — there is a recipe behind the row |
-| `fits` | 663 | inferred: the model's memory floor is under the card's VRAM, and the vendor is supported. Not measured |
-| `unknown` | 456 | no floor established for this model, so no honest call can be made |
-| `too_big` | 574 | the memory floor exceeds this card. This is the one verdict asserted from anywhere, not only from same-vendor evidence |
+| `verified` | 972 | somebody ran it on this exact card and wrote down how — there is a recipe behind the row |
+| `fits` | 659 | inferred: the model's memory floor is under the card's VRAM, and the vendor is supported. Not measured |
+| `unknown` | 455 | no floor established for this model, so no honest call can be made |
+| `too_big` | 560 | the memory floor exceeds this card. This is the one verdict asserted from anywhere, not only from same-vendor evidence |
 
 The asymmetry between `fits` and `too_big` is deliberate: a model is only called
 runnable on evidence from the same vendor's hardware, but it is called *not*
@@ -123,7 +123,7 @@ extend to the measurements themselves, which belong to the sites named in
   title  = {Self-Hosted AI — GPU Compatibility, Recipes and Catalogue},
   author = {smeltcore},
   url    = {https://smeltcore.com},
-  note   = {Generated 2026-08-28T12:13:44+00:00},
+  note   = {Generated 2026-08-31T18:39:19+00:00},
   year   = {2026}
 }
 ```

data/compatibility.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:95a6a2b70d2581e79c23f1ebd031c94a9184251665de24b8799a3c803d40a763
-size 42957
+oid sha256:5a8174b2ef81c1fbc986b7779a1117e8648b1437d153adb250d8a1e3d0b3e11b
+size 43961

data/gpus.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:984288d68251af5834d2f6c21f48da5bb8f27c4d7330f8ad7184ad0b2effbacf
+oid sha256:f9ecdb73813201a798f717bd49bd8a08226f971e3779319f62c587baba06bfda
 size 3360

data/recipes.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ced13bdbd3db2921b4445bd2c5d472ba6076873e9808eca282bbdbdfe9277b2a
-size 2370012
+oid sha256:182e22581cf46805e7e34103a041783ff3bbd9a3c8dd7dec29a255807663c7fe
+size 2723098