docs(frank-small): record the CACHE_RAM=2048 cap in the deployed flags
This commit is contained in:
1 parent
2b046a21d0
commit
9e0c918b6e
1 file changed
+1
-1
@@ -133,7 +133,7 @@ prefill ~254k tokens / decode at ~253k context:
|
|||||||
Repo truth: `scripts/frank-small.env` + `scripts/serve-frank-small.sh`.
|
Repo truth: `scripts/frank-small.env` + `scripts/serve-frank-small.sh`.
|
||||||
Key values: `MODEL_DIR=/home/dank/models/Gemma4-12B-QAT`,
|
Key values: `MODEL_DIR=/home/dank/models/Gemma4-12B-QAT`,
|
||||||
`MODEL_FILE=Gemma4-12B-QAT-Uncensored-HauhauCS-Balanced-Q4_K_M.gguf`,
|
`MODEL_FILE=Gemma4-12B-QAT-Uncensored-HauhauCS-Balanced-Q4_K_M.gguf`,
|
||||||
`DRAFT_FILE=mtp-gemma-4-12B-it.gguf`, `CACHE_TYPE_K=V=q8_0`,
|
`DRAFT_FILE=mtp-gemma-4-12B-it.gguf`, `CACHE_TYPE_K=V=q8_0`, `CACHE_RAM=2048`,
|
||||||
`SPEC_TYPE=draft-mtp,ngram-mod`, `SPEC_DRAFT_N_MAX=3`, `POLL_LEVEL=50`,
|
`SPEC_TYPE=draft-mtp,ngram-mod`, `SPEC_DRAFT_N_MAX=3`, `POLL_LEVEL=50`,
|
||||||
`CTX_SIZE=1048576` with `PARALLEL=4` (four full 262,144-token slots),
|
`CTX_SIZE=1048576` with `PARALLEL=4` (four full 262,144-token slots),
|
||||||
`BATCH_SIZE=8192`, `UBATCH_SIZE=1024`, `NGL=999`, `FLASH_ATTN=on`,
|
`BATCH_SIZE=8192`, `UBATCH_SIZE=1024`, `NGL=999`, `FLASH_ATTN=on`,
|
||||||
|
|||||||
Reference in new issue
Block a user