27b tier: state the measurement condition first, then the weights-vs-graph-pools memory reason
Browse files
README.md
CHANGED
|
@@ -122,7 +122,7 @@ Measured on one RTX 4090 (48 GB) for `0.8b`/`4b` and one NVIDIA A100-SXM4-80GB f
|
|
| 122 |
|
| 123 |
### 27b (A100 80 GB)
|
| 124 |
|
| 125 |
-
Measured with CUDA graphs disabled (`VERDICT_TORCH_GRAPHS=0`)
|
| 126 |
|
| 127 |
<table>
|
| 128 |
<thead>
|
|
|
|
| 122 |
|
| 123 |
### 27b (A100 80 GB)
|
| 124 |
|
| 125 |
+
Measured with CUDA graphs disabled (`VERDICT_TORCH_GRAPHS=0`): the bf16 weights alone take most of the 80 GB card, leaving no room for the graph pools.
|
| 126 |
|
| 127 |
<table>
|
| 128 |
<thead>
|