gortium/compose
1
1
Fork 0
Code Issues Pull Requests 18 Actions Packages Projects Releases Wiki Activity

fix: honcho embedding config - fix base_url nesting, switch to bge-m3, add deriver to CMD
Some checks failed
Build Hermes agent / build (pull_request) Has been cancelled
Details
Build ollama (gfx906) / build (pull_request) Has been cancelled
Details

Browse Source
This commit is contained in:
Hermes
2026-05-23 18:54:35 -04:00
parent 08778db685
commit 63b6cd3461
2 changed files with 12 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ai/honcho/Dockerfile
View File

@@ -71,4 +71,4 @@ COPY honcho-nginx.conf /etc/nginx/conf.d/default.conf
EXPOSE 80
CMD ["sh", "-c", "nginx -g 'daemon off;' & fastapi run --host 127.0.0.1 --port 8000 src/main.py"]
CMD ["sh", "-c", "nginx -g 'daemon off;' & fastapi run --host 127.0.0.1 --port 8000 src/main.py & python3 -m src.deriver & wait -n"]

29
ai/honcho/config.toml
View File

@@ -29,15 +29,15 @@ URL = "redis://honcho-redis:6379/0"
[llm]
DEFAULT_MAX_TOKENS = 4096
# Embeddings via Ollama
# Embeddings via Ollama — bge-m3 provides 1024-dim
[embedding]
VECTOR_DIMENSIONS = 1024
MAX_INPUT_TOKENS = 8192
[embedding.model_config]
transport = "openai"
model = "nomic-embed-text"
base_url = "http://ollama:11434/v1"
model = "bge-m3"
overrides = {base_url = "http://ollama:11434/v1", api_key = "ollama"}
# --- Deriver ---
[deriver]
@@ -47,10 +47,9 @@ POLLING_SLEEP_INTERVAL_SECONDS = 5.0
FLUSH_ENABLED = true
[deriver.model_config]
overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
transport = "openai"
model = "deepseek-v4-flash"
base_url = "https://opencode.ai/zen/go/v1"
api_key_env = "HONCHO_OPENAI_API_KEY"
# --- Dialectic ---
[dialectic]
@@ -61,42 +60,37 @@ SESSION_HISTORY_MAX_TOKENS = 8192
MAX_TOOL_ITERATIONS = 1
MAX_OUTPUT_TOKENS = 512
[dialectic.levels.minimal.model_config]
overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
transport = "openai"
model = "deepseek-v4-flash"
base_url = "https://opencode.ai/zen/go/v1"
api_key_env = "HONCHO_OPENAI_API_KEY"
[dialectic.levels.low]
MAX_TOOL_ITERATIONS = 3
[dialectic.levels.low.model_config]
overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
transport = "openai"
model = "deepseek-v4-flash"
base_url = "https://opencode.ai/zen/go/v1"
api_key_env = "HONCHO_OPENAI_API_KEY"
[dialectic.levels.medium]
MAX_TOOL_ITERATIONS = 2
[dialectic.levels.medium.model_config]
overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
transport = "openai"
model = "deepseek-v4-flash"
base_url = "https://opencode.ai/zen/go/v1"
api_key_env = "HONCHO_OPENAI_API_KEY"
[dialectic.levels.high]
MAX_TOOL_ITERATIONS = 4
[dialectic.levels.high.model_config]
overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
transport = "openai"
model = "deepseek-v4-flash"
base_url = "https://opencode.ai/zen/go/v1"
api_key_env = "HONCHO_OPENAI_API_KEY"
[dialectic.levels.max]
MAX_TOOL_ITERATIONS = 10
[dialectic.levels.max.model_config]
overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
transport = "openai"
model = "deepseek-v4-flash"
base_url = "https://opencode.ai/zen/go/v1"
api_key_env = "HONCHO_OPENAI_API_KEY"
# --- Summary ---
[summary]
@@ -105,10 +99,9 @@ MESSAGES_PER_SHORT_SUMMARY = 20
MESSAGES_PER_LONG_SUMMARY = 60
[summary.model_config]
overrides = {base_url = "https://opencode.ai/zen/go/v1", api_key_env = "HONCHO_OPENAI_API_KEY"}
transport = "openai"
model = "deepseek-v4-flash"
base_url = "https://opencode.ai/zen/go/v1"
api_key_env = "HONCHO_OPENAI_API_KEY"
# --- Dream ---
[dream]
@@ -121,4 +114,4 @@ ENABLED = true
# --- Vector Store ---
[vector_store]
TYPE = "pgvector"
DIMENSIONS = 1024
# DIMENSIONS is deprecated — EMBEDDING.VECTOR_DIMENSIONS is authoritative