Alexander Korolev
e0eeffe3b1
lab-dflash2: use dequantized bf16 drafter variant (arch DFlashDraftModel for lab image; MXFP8 scales unsupported by its loader)
|
преди 1 месец | |
|---|---|---|
| .. | ||
| recipe.toml | преди 1 месец | |
|
Alexander Korolev
e0eeffe3b1
lab-dflash2: use dequantized bf16 drafter variant (arch DFlashDraftModel for lab image; MXFP8 scales unsupported by its loader)
|
преди 1 месец | |
|---|---|---|
| .. | ||
| recipe.toml | преди 1 месец | |