kissin42 commited on
Commit
65c22ce
·
verified ·
1 Parent(s): c5373dc

Replace Ant-v5 with bundle 21 and refresh context sweep

Browse files

MuJoCo 3.2.3, seeds 0..49, max_steps=1000, BOS discard. Main score uses kv_cache_max_len=32.

Files changed (2) hide show
  1. README.md +2 -2
  2. ant-v5/model.safetensors +1 -1
README.md CHANGED
@@ -79,7 +79,7 @@ Notebook version: [examples/hub_quickstart.ipynb](https://github.com/ccnets-team
79
 
80
  | Env | Bundle | Ctx | Return | Norm. | Simple Ref. | Medium Ref. |
81
  |---|---|---:|---:|---:|---:|---:|
82
- | `Ant-v5` | `ant-v5` | 32 | 5021.14±1500.60 | 75.50±22.25 | 59.99 ✓ | 86.54 ✗ |
83
  | `HalfCheetah-v5` | `halfcheetah-v5` | 32 | 6816.48±3135.53 | 42.87±19.01 | 43.54 ✗ | 74.83 ✗ |
84
  | `Hopper-v5` | `hopper-v5` | 32 | 2713.66±1075.57 | 70.21±28.01 | 42.65 ✓ | 72.91 ✗ |
85
  | `Walker2d-v5` | `walker2d-v5` | 32 | 3899.88±706.57 | 56.93±10.32 | 59.51 ✗ | 83.26 ✗ |
@@ -109,7 +109,7 @@ seeds `0..49`, `max_steps=1000`):
109
 
110
  | Env | `kv=16` (0.5×) | `kv=32` (1×) | `kv=64` (2×) | Trend |
111
  |---|---:|---:|---:|---|
112
- | `Ant-v5` | **5467.20±1153.95** | 5021.14±1500.60 | 4957.91±1686.26 | **0.5× recommended** |
113
  | `HalfCheetah-v5` | 6793.17±2939.17 | 6816.48±3135.53 | 6468.21±3234.51 | ≈ flat |
114
  | `Hopper-v5` | 3361.71±103.69 | 2713.66±1075.57 | 992.92±445.63 | shorter ↑, 2× collapses |
115
  | `Walker2d-v5` | 3950.09±459.01 | 3899.88±706.57 | 3842.19±718.60 | ≈ flat |
 
79
 
80
  | Env | Bundle | Ctx | Return | Norm. | Simple Ref. | Medium Ref. |
81
  |---|---|---:|---:|---:|---:|---:|
82
+ | `Ant-v5` | `ant-v5` | 32 | 5126.03±1568.43 | 77.05±23.26 | 59.99 ✓ | 86.54 ✗ |
83
  | `HalfCheetah-v5` | `halfcheetah-v5` | 32 | 6816.48±3135.53 | 42.87±19.01 | 43.54 ✗ | 74.83 ✗ |
84
  | `Hopper-v5` | `hopper-v5` | 32 | 2713.66±1075.57 | 70.21±28.01 | 42.65 ✓ | 72.91 ✗ |
85
  | `Walker2d-v5` | `walker2d-v5` | 32 | 3899.88±706.57 | 56.93±10.32 | 59.51 ✗ | 83.26 ✗ |
 
109
 
110
  | Env | `kv=16` (0.5×) | `kv=32` (1×) | `kv=64` (2×) | Trend |
111
  |---|---:|---:|---:|---|
112
+ | `Ant-v5` | 5109.97±1711.18 | 5126.03±1568.43 | **5256.33±1429.58** | **2× highest mean (≈)** |
113
  | `HalfCheetah-v5` | 6793.17±2939.17 | 6816.48±3135.53 | 6468.21±3234.51 | ≈ flat |
114
  | `Hopper-v5` | 3361.71±103.69 | 2713.66±1075.57 | 992.92±445.63 | shorter ↑, 2× collapses |
115
  | `Walker2d-v5` | 3950.09±459.01 | 3899.88±706.57 | 3842.19±718.60 | ≈ flat |
ant-v5/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:59229ce58c33fd7c18ed66a634c25b5da7fe5d9da0ee1ff53e101f9c36cce53a
3
  size 16931180
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5815216dbb2dd97e5fc92dea0864cc6a3a1e4425ac4ae5e1a83cdd058c4e5732
3
  size 16931180