Repo hygiene: remove latest.pt duplicate + tokenizer vocab; document runtime requirement
Browse files- README.md +2 -0
- latest.pt +0 -3
- samg_tokenizer.vocab +0 -0
README.md
CHANGED
|
@@ -141,6 +141,8 @@ ids = torch.tensor([sp.EncodeAsIds(prompt)])
|
|
| 141 |
|
| 142 |
## Limitations
|
| 143 |
|
|
|
|
|
|
|
| 144 |
- `build_test_cycle` (17%) and the exact-match of `plan_chain`/`risk_gate`
|
| 145 |
(58%) plateau because long, strictly-ordered plans are hard at 30M; decompose
|
| 146 |
long plans app-side into shorter sub-calls.
|
|
|
|
| 141 |
|
| 142 |
## Limitations
|
| 143 |
|
| 144 |
+
- Inference requires the official AMEFORGE runtime. The published weights are not compatible with from-scratch reimplementations, and third-party ports are unsupported.
|
| 145 |
+
|
| 146 |
- `build_test_cycle` (17%) and the exact-match of `plan_chain`/`risk_gate`
|
| 147 |
(58%) plateau because long, strictly-ordered plans are hard at 30M; decompose
|
| 148 |
long plans app-side into shorter sub-calls.
|
latest.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:16f84dfd996cac9c055f7aecf2f805e425bb65b4d832a145e7bb3e4bfe11e84f
|
| 3 |
-
size 362276829
|
|
|
|
|
|
|
|
|
|
|
|
samg_tokenizer.vocab
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|