perf: refactor GLM KV cache and attention, add end-to-end timing inst… #98
Artifacts
Produced during runtime
| Name | Size | Digest | |
|---|---|---|---|
|
deepseek-ocr-macos
|
17.1 MB |
sha256:93b2f9ec8a496a4b608dd8b82f581acb5f052e9f75c5923b8c6512ac9206edc4
|
|
|
deepseek-ocr-windows
|
15.6 MB |
sha256:bc63c474a32519b8b417973d65cbea7729938e4130185b3b77aafcb56e59b13b
|
|
|
deepseek-ocr-windows-upx
|
10.3 MB |
sha256:48a85b8095eca65b9ab1e4f6f67b976185512db2a41059979249451558565bca
|
|