Remove validation progress from model card
Browse files
README.md
CHANGED
|
@@ -52,9 +52,7 @@ checkpoint_dir = snapshot_download(
|
|
| 52 |
|
| 53 |
Replace `llama3-1b` with another directory above to download its parameters. See the code repository for evaluation instructions. The current directory-resume workflow additionally expects `router_stats.pth`; that file is not included in this release, so the downloaded pair alone is not a complete directory-resume bundle.
|
| 54 |
|
| 55 |
-
##
|
| 56 |
-
|
| 57 |
-
All four checkpoints passed per-layer parameter restoration checks against the current MoBiQuant implementation. GPU WikiText2 sweeps are in progress; this release does not claim that all paper results have been reproduced.
|
| 58 |
|
| 59 |
The implementation evaluates quantization using floating-point simulation. These files do not constitute a packed low-bit deployment format or establish inference speedups.
|
| 60 |
|
|
|
|
| 52 |
|
| 53 |
Replace `llama3-1b` with another directory above to download its parameters. See the code repository for evaluation instructions. The current directory-resume workflow additionally expects `router_stats.pth`; that file is not included in this release, so the downloaded pair alone is not a complete directory-resume bundle.
|
| 54 |
|
| 55 |
+
## Limitations
|
|
|
|
|
|
|
| 56 |
|
| 57 |
The implementation evaluates quantization using floating-point simulation. These files do not constitute a packed low-bit deployment format or establish inference speedups.
|
| 58 |
|