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

Use validated Medium MTP2 and microbatch256 combination

Browse Source
This commit is contained in:
Mikei386
2026-09-20 23:23:57 +02:00
parent b352e29c89
commit fbe8c6f1e9
3 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
compose.yaml
+2 -2
View File
@@ -206,7 +206,7 @@ services:
- --batch-size - --batch-size
- "${MEDIUM_BATCH_SIZE:-2048}" - "${MEDIUM_BATCH_SIZE:-2048}"
- --ubatch-size - --ubatch-size
- "${MEDIUM_UBATCH_SIZE:-512}" - "${MEDIUM_UBATCH_SIZE:-256}"
- --parallel - --parallel
- "${MEDIUM_PARALLEL_SLOTS:-1}" - "${MEDIUM_PARALLEL_SLOTS:-1}"
- --kv-unified - --kv-unified
@@ -247,7 +247,7 @@ services:
- --spec-type - --spec-type
- draft-mtp - draft-mtp
- --spec-draft-n-max - --spec-draft-n-max
- "3" - "2"
- --spec-draft-type-k - --spec-draft-type-k
- f16 - f16
- --spec-draft-type-v - --spec-draft-type-v
config/profile-matrix.json
+1 -1
View File
@@ -24,7 +24,7 @@
"model_family": "Qwen3.8-27B IQ4 XS Pure", "model_family": "Qwen3.8-27B IQ4 XS Pure",
"gpu_split": "85:15", "gpu_split": "85:15",
"vision": true, "vision": true,
"mtp": 3, "mtp": 2,
"description": "Ausgewogenes Standardprofil für Alltag und lange agentische Aufgaben." "description": "Ausgewogenes Standardprofil für Alltag und lange agentische Aufgaben."
}, },
{ {
docs/STANDARD_PROFILE_MATRIX.md
+1 -1
View File
@@ -7,7 +7,7 @@ Standardprofil: **medium** · globales Ausgabelimit: **8192 Token**
| Profil | API-Alias | Gesamtkontext | Slots | Modell | GPU-Verteilung | Vision | MTP | | Profil | API-Alias | Gesamtkontext | Slots | Modell | GPU-Verteilung | Vision | MTP |
|---|---|---:|---:|---|---|---|---:| |---|---|---:|---:|---|---|---|---:|
| fast | `qwen-fast` | 76,800 | 1 | Qwen3.8-27B IQ4 Mix | 5080 only | ja | 2 | | fast | `qwen-fast` | 76,800 | 1 | Qwen3.8-27B IQ4 Mix | 5080 only | ja | 2 |
| medium | `qwen-medium` | 160,000 | 1 | Qwen3.8-27B IQ4 XS Pure | 85:15 | ja | 3 | | medium | `qwen-medium` | 160,000 | 1 | Qwen3.8-27B IQ4 XS Pure | 85:15 | ja | 2 |
| large | `qwen-large` | 192,000 | 1 | Qwen3.8-27B IQ4 XS Pure | 86:14 | ja | 2 | | large | `qwen-large` | 192,000 | 1 | Qwen3.8-27B IQ4 XS Pure | 86:14 | ja | 2 |
| ultra | `qwen-ultra` | 262,144 | 1 | Qwen3.8-27B IQ4 XS Pure | 80:20 | nein | 2 | | ultra | `qwen-ultra` | 262,144 | 1 | Qwen3.8-27B IQ4 XS Pure | 80:20 | nein | 2 |
| uncensored | `qwen-uncensored` | 80,000 | 1 | Qwen3.8-27B Abliterated Q4_K_M | 90:10 | ja | 2 | | uncensored | `qwen-uncensored` | 80,000 | 1 | Qwen3.8-27B Abliterated Q4_K_M | 90:10 | ja | 2 |