michael/AI-Profile-Router
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Bound Hermes context compression latency

Browse Source
This commit is contained in:
Mikei386 committed 2026-08-26 06:54:57 +02:00
1 parent a702353134
commit 400de54ff9
2 files changed
+8 -5
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

platform/hermes/config.yaml
+6 -4
View File
@@ -78,8 +78,8 @@ agent:
max_subagents: 8 max_subagents: 8
# Keep ample room for long agent work. Compression starts at 82% of whichever # Keep ample room for long agent work. Compression starts at 82% of whichever
# router profile is selected and retains a useful 35% instead of collapsing a # router profile is selected. Keep the result compact enough that a local
# large conversation to a tiny summary. # model does not spend many minutes generating the handoff.
compression: compression:
enabled: true enabled: true
# Fold one older assistant/tool exchange into the rolling summary every five # Fold one older assistant/tool exchange into the rolling summary every five
@@ -90,14 +90,16 @@ compression:
micro_compact_defrag_threshold_tokens: 2000 micro_compact_defrag_threshold_tokens: 2000
progress_notices: true progress_notices: true
threshold: 0.82 threshold: 0.82
target_ratio: 0.35 target_ratio: 0.15
tail_mode: "lean" tail_mode: "lean"
protect_last_n: 20 protect_last_n: 20
protect_first_n: 0 protect_first_n: 0
proactive_prune_tokens: 50000 proactive_prune_tokens: 50000
proactive_prune_min_result_chars: 4000 proactive_prune_min_result_chars: 4000
proactive_prune_min_reclaim_tokens: 4096 proactive_prune_min_reclaim_tokens: 4096
context_total_ceiling_seconds: 600 # A failed local summarizer must not block an interactive client for ten
# minutes. Continue without dropping messages after two minutes.
context_total_ceiling_seconds: 120
auxiliary: auxiliary:
# Session names are cosmetic and used to create a second concurrent LLM # Session names are cosmetic and used to create a second concurrent LLM
platform/hermes/install-profiles.sh
+2 -1
View File
@@ -33,7 +33,8 @@ create_profile() {
docker exec "$HERMES_CONTAINER" hermes -p "$name" config set auxiliary.title_generation.enabled false docker exec "$HERMES_CONTAINER" hermes -p "$name" config set auxiliary.title_generation.enabled false
docker exec "$HERMES_CONTAINER" hermes -p "$name" config unset compression.threshold_tokens || true docker exec "$HERMES_CONTAINER" hermes -p "$name" config unset compression.threshold_tokens || true
docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.threshold 0.82 docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.threshold 0.82
docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.target_ratio 0.35 docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.target_ratio 0.15
docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.context_total_ceiling_seconds 120
docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.protect_last_n 20 docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.protect_last_n 20
docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.protect_first_n 0 docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.protect_first_n 0
docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.micro_compact true docker exec "$HERMES_CONTAINER" hermes -p "$name" config set compression.micro_compact true