docs(card): refresh Ant-v5 to shipped bundle 30

Ant-v5 weights were replaced with bundle 30 in f8e8c77 without updating the card, so the table still advertised bundle 21 numbers.

Main table: 5126.03+-1568.43 -> 5434.12+-1298.24, norm 77.05+-23.26 -> 81.62+-19.25.
KV sweep: 16/32/64 -> 5292.09+-1338.88 / 5434.12+-1298.24 / 5323.72+-1635.79; trend is now 1x highest mean (0.5x -2.6%, 2x -2.0%, within noise).
Measured over 50 episodes, seeds 0..49, max_steps=1000, on the card's declared reference stack (causal-gpt-rl 0.14.0, mujoco 3.2.3, gymnasium 1.2.3, torch 2.8.0+cu129); the evaluated bundle is byte-identical to the published ant-v5/model.safetensors.

Also list hopper-v5 in the root config.json bundles array, which was missing.

README.md

@@ -79,7 +79,7 @@ Notebook version: [examples/hub_quickstart.ipynb](https://github.com/ccnets-team
 
 | Env | Bundle | Ctx | Return | Norm. | Simple Ref. | Medium Ref. |
 |---|---|---:|---:|---:|---:|---:|
-| `Ant-v5` | `ant-v5` | 32 | 5126.03±1568.43 | 77.05±23.26 | 59.99 ✓ | 86.54 ✗ |
+| `Ant-v5` | `ant-v5` | 32 | 5434.12±1298.24 | 81.62±19.25 | 59.99 ✓ | 86.54 ✗ |
 | `HalfCheetah-v5` | `halfcheetah-v5` | 32 | 6816.48±3135.53 | 42.87±19.01 | 43.54 ✗ | 74.83 ✗ |
 | `Hopper-v5` | `hopper-v5` | 32 | 3199.65±21.74 | 82.87±0.57 | 42.65 ✓ | 72.91 ✓ |
 | `Walker2d-v5` | `walker2d-v5` | 32 | 4122.68±299.84 | 60.19±4.38 | 59.51 ✓ | 83.26 ✗ |
@@ -109,7 +109,7 @@ seeds `0..49`, `max_steps=1000`):
 
 | Env | `kv=16` (0.5×) | `kv=32` (1×) | `kv=64` (2×) | Trend |
 |---|---:|---:|---:|---|
-| `Ant-v5` | 5109.97±1711.18 | 5126.03±1568.43 | **5256.33±1429.58** | **2× highest mean (≈)** |
+| `Ant-v5` | 5292.09±1338.88 | **5434.12±1298.24** | 5323.72±1635.79 | **1× highest mean (≈)** |
 | `HalfCheetah-v5` | 6793.17±2939.17 | 6816.48±3135.53 | 6468.21±3234.51 | ≈ flat |
 | `Hopper-v5` | 3189.53±22.58 | **3199.65±21.74** | 3190.09±23.75 | **≈ flat and stable** |
 | `Walker2d-v5` | 4021.00±573.17 | **4122.68±299.84** | 2659.11±1297.61 | **1× best; 2× collapses** |

config.json

@@ -1,4 +1,4 @@
 {
   "model_type": "causal-gpt-rl",
-  "bundles": ["ant-v5", "halfcheetah-v5", "humanoid-v5", "walker2d-v5"]
+  "bundles": ["ant-v5", "halfcheetah-v5", "hopper-v5", "humanoid-v5", "walker2d-v5"]
 }