dank/serverAppConfigs
1
0 Fork 0
Code Issues 4 Pull Requests Actions Packages Projects Releases Wiki Activity

Flip production to FASTGREEDY=1 (operator decision 2026-08-22)

Browse Source 
Reasoning off, argmax fastpath on, deterministic sampler. The fast
greedy path rejects non-greedy requests with 400; chat clients must
send temperature 0.
This commit is contained in:
Frank committed 2026-08-22 13:26:52 -07:00
1 parent 7f617b9ee6
commit fe71f7516d
1 file changed
+5 -2
Download Patch File Download Diff File Expand all files Collapse all files
scripts/frank-default-large.env
+5 -2
View File
@@ -83,10 +83,13 @@ SPEC_DRAFT_P_SPLIT=0.10
# preserve_thinking=true (preserve thinking tokens across turns). # preserve_thinking=true (preserve thinking tokens across turns).
# #
# FASTGREEDY mode: set to 1 for fast-greedy benchmark (card's default profile) # FASTGREEDY mode: set to 1 for fast-greedy benchmark (card's default profile)
# FASTGREEDY=0 (default): production chat config (reasoning on, argmax off) # FASTGREEDY=0: production chat config (reasoning on, argmax off)
# FASTGREEDY=1: reasoning off, LLAMA_TARGET_GREEDY_ARGMAX_FASTPATH=1, deterministic # FASTGREEDY=1: reasoning off, LLAMA_TARGET_GREEDY_ARGMAX_FASTPATH=1, deterministic
# sampler. Mirrors the card's 'fast greedy (default)' benchmark config. # sampler. Mirrors the card's 'fast greedy (default)' benchmark config.
FASTGREEDY=0 # 2026-08-22: flipped to 1 for real (operator decision). NOTE: the fast
# greedy path rejects non-greedy requests (temp != 0, top_k > 1, penalties,
# reasoning budgets) with 400. Chat clients must send temperature 0.
FASTGREEDY=1
# For fast-greedy bench: set FASTGREEDY=1 above. This flips to: reasoning off, # For fast-greedy bench: set FASTGREEDY=1 above. This flips to: reasoning off,
# argmax fastpath=1, sampling defaults. Deploy with # argmax fastpath=1, sampling defaults. Deploy with