P3: llama-swap capabilities deklariert + Plugin lernt Tool-Namen (Hermes-0.18-API)
- deploy-Config: capabilities (in/out/tools/context) je Modell — /v1/models informiert Clients korrekt; context = nutzbarer Kontext pro Request (c/parallel) - mc2-memory: sync_turn nutzt messages aus Hermes >=0.18 — nur Tool-NAMEN (Ergebnisse bleiben draussen: untrusted/Poisoning-Vektor); deployt, Postcheck gruen - Mem0 v3-Algorithmus auf der Box verifiziert (BM25/Entity/Hybrid in 2.0.8 aktiv) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
@@ -13,24 +13,44 @@ models:
|
|||||||
aliases:
|
aliases:
|
||||||
- hermes
|
- hermes
|
||||||
- fast
|
- fast
|
||||||
|
# context = nutzbarer Kontext PRO REQUEST (c geteilt durch parallel-Slots).
|
||||||
|
capabilities:
|
||||||
|
in: [text]
|
||||||
|
out: [text]
|
||||||
|
tools: true
|
||||||
|
context: 65536
|
||||||
Qwen3.5-122B-A10B:
|
Qwen3.5-122B-A10B:
|
||||||
cmd: |
|
cmd: |
|
||||||
llama-server -m /srv/models/Qwen3.5-122B-A10B-GGUF/Q4_K_M/Qwen3.5-122B-A10B-Q4_K_M-00001-of-00003.gguf --host 127.0.0.1 --port ${PORT} -c 32768 -ngl 999 -fa on --no-mmap --jinja --cache-reuse 256 -cram 16384
|
llama-server -m /srv/models/Qwen3.5-122B-A10B-GGUF/Q4_K_M/Qwen3.5-122B-A10B-Q4_K_M-00001-of-00003.gguf --host 127.0.0.1 --port ${PORT} -c 32768 -ngl 999 -fa on --no-mmap --jinja --cache-reuse 256 -cram 16384
|
||||||
ttl: 600
|
ttl: 600
|
||||||
aliases:
|
aliases:
|
||||||
- heavy
|
- heavy
|
||||||
|
capabilities:
|
||||||
|
in: [text]
|
||||||
|
out: [text]
|
||||||
|
tools: true
|
||||||
|
context: 32768
|
||||||
Qwen3-Coder-Next:
|
Qwen3-Coder-Next:
|
||||||
cmd: |
|
cmd: |
|
||||||
llama-server -m /srv/models/Qwen3-Coder-Next-GGUF/Qwen3-Coder-Next-Q4_K_M.gguf --host 127.0.0.1 --port ${PORT} -c 131072 -ngl 999 -fa on --no-mmap --jinja --cache-reuse 256 -cram 16384
|
llama-server -m /srv/models/Qwen3-Coder-Next-GGUF/Qwen3-Coder-Next-Q4_K_M.gguf --host 127.0.0.1 --port ${PORT} -c 131072 -ngl 999 -fa on --no-mmap --jinja --cache-reuse 256 -cram 16384
|
||||||
ttl: 600
|
ttl: 600
|
||||||
aliases:
|
aliases:
|
||||||
- coder
|
- coder
|
||||||
|
capabilities:
|
||||||
|
in: [text]
|
||||||
|
out: [text]
|
||||||
|
tools: true
|
||||||
|
context: 131072
|
||||||
Qwen3-VL-8B-Instruct:
|
Qwen3-VL-8B-Instruct:
|
||||||
cmd: |
|
cmd: |
|
||||||
llama-server -m /srv/models/Qwen3-VL-8B-Instruct-GGUF/Qwen3VL-8B-Instruct-Q4_K_M.gguf --host 127.0.0.1 --port ${PORT} -c 32768 -ngl 999 -fa on --no-mmap --mmproj /srv/models/Qwen3-VL-8B-Instruct-GGUF/mmproj-Qwen3VL-8B-Instruct-F16.gguf --jinja
|
llama-server -m /srv/models/Qwen3-VL-8B-Instruct-GGUF/Qwen3VL-8B-Instruct-Q4_K_M.gguf --host 127.0.0.1 --port ${PORT} -c 32768 -ngl 999 -fa on --no-mmap --mmproj /srv/models/Qwen3-VL-8B-Instruct-GGUF/mmproj-Qwen3VL-8B-Instruct-F16.gguf --jinja
|
||||||
ttl: 300
|
ttl: 300
|
||||||
aliases:
|
aliases:
|
||||||
- vision
|
- vision
|
||||||
|
capabilities:
|
||||||
|
in: [text, image]
|
||||||
|
out: [text]
|
||||||
|
context: 32768
|
||||||
Qwen3-Embedding-0.6B:
|
Qwen3-Embedding-0.6B:
|
||||||
cmd: |
|
cmd: |
|
||||||
llama-server -m /srv/models/Qwen3-Embedding-0.6B-GGUF/Qwen3-Embedding-0.6B-Q8_0.gguf --host 127.0.0.1 --port ${PORT} --embedding --pooling last -ngl 999 -fa on --no-mmap -c 8192
|
llama-server -m /srv/models/Qwen3-Embedding-0.6B-GGUF/Qwen3-Embedding-0.6B-Q8_0.gguf --host 127.0.0.1 --port ${PORT} --embedding --pooling last -ngl 999 -fa on --no-mmap -c 8192
|
||||||
@@ -43,12 +63,22 @@ models:
|
|||||||
ttl: 300
|
ttl: 300
|
||||||
aliases:
|
aliases:
|
||||||
- coder-lite
|
- coder-lite
|
||||||
|
capabilities:
|
||||||
|
in: [text]
|
||||||
|
out: [text]
|
||||||
|
tools: true
|
||||||
|
context: 131072
|
||||||
GLM-4.6V-Flash:
|
GLM-4.6V-Flash:
|
||||||
cmd: |
|
cmd: |
|
||||||
llama-server -m /srv/models/GLM-4.6V-Flash-GGUF/GLM-4.6V-Flash-Q4_K_M.gguf --host 127.0.0.1 --port ${PORT} -c 131072 -ngl 999 -fa on --no-mmap --mmproj /srv/models/GLM-4.6V-Flash-GGUF/mmproj-BF16.gguf --jinja
|
llama-server -m /srv/models/GLM-4.6V-Flash-GGUF/GLM-4.6V-Flash-Q4_K_M.gguf --host 127.0.0.1 --port ${PORT} -c 131072 -ngl 999 -fa on --no-mmap --mmproj /srv/models/GLM-4.6V-Flash-GGUF/mmproj-BF16.gguf --jinja
|
||||||
ttl: 300
|
ttl: 300
|
||||||
aliases:
|
aliases:
|
||||||
- scout
|
- scout
|
||||||
|
capabilities:
|
||||||
|
in: [text, image]
|
||||||
|
out: [text]
|
||||||
|
tools: true
|
||||||
|
context: 131072
|
||||||
groups:
|
groups:
|
||||||
brains:
|
brains:
|
||||||
swap: false
|
swap: false
|
||||||
|
|||||||
@@ -150,6 +150,18 @@ class MC2MemoryProvider(MemoryProvider):
|
|||||||
return
|
return
|
||||||
msgs: List[Dict[str, str]] = [{"role": "user", "content": u}]
|
msgs: List[Dict[str, str]] = [{"role": "user", "content": u}]
|
||||||
a = (assistant_content or "").strip()
|
a = (assistant_content or "").strip()
|
||||||
|
# Hermes ≥0.18 liefert optional den vollen Turn-Kontext (`messages`, inkl. Tool-Calls).
|
||||||
|
# Bewusst NUR die Tool-NAMEN mitlernen ("hat X per Tool Y geprüft") — Tool-ERGEBNISSE
|
||||||
|
# sind untrusted (Web-Inhalte!) und wären ein Poisoning-Vektor am Junk-Guard vorbei.
|
||||||
|
if messages:
|
||||||
|
tools = []
|
||||||
|
for m in messages:
|
||||||
|
for tc in (m.get("tool_calls") or []):
|
||||||
|
name = ((tc.get("function") or {}).get("name") or "").strip()
|
||||||
|
if name and name not in tools:
|
||||||
|
tools.append(name)
|
||||||
|
if tools:
|
||||||
|
a = (a + f"\n[genutzte Tools: {', '.join(tools[:6])}]").strip()
|
||||||
if a:
|
if a:
|
||||||
msgs.append({"role": "assistant", "content": a[:4000]})
|
msgs.append({"role": "assistant", "content": a[:4000]})
|
||||||
try:
|
try:
|
||||||
|
|||||||
Reference in New Issue
Block a user