Image-Text-to-Text
GGUF
German
English
llama.cpp
mtp
speculative-decoding
qwen3
multimodal
conversational
Instructions to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0 # Run inference directly in the terminal: llama cli -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0 # Run inference directly in the terminal: llama cli -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0 # Run inference directly in the terminal: ./llama-cli -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0 # Run inference directly in the terminal: ./build/bin/llama-cli -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
Use Docker
docker model run hf.co/Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
- LM Studio
- Jan
- vLLM
How to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Davidmg0815/Qwen3.8-27B-MTP-GGUF" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Davidmg0815/Qwen3.8-27B-MTP-GGUF", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker
docker model run hf.co/Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
- Ollama
How to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with Ollama:
ollama run hf.co/Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
- Unsloth Studio
How to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for Davidmg0815/Qwen3.8-27B-MTP-GGUF to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for Davidmg0815/Qwen3.8-27B-MTP-GGUF to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for Davidmg0815/Qwen3.8-27B-MTP-GGUF to start chatting
- Pi
How to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with Docker Model Runner:
docker model run hf.co/Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
- Lemonade
How to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
Run and chat with the model
lemonade run user.Qwen3.8-27B-MTP-GGUF-Q8_0
List all available models
lemonade list
- Hermes Agent
How to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use Davidmg0815/Qwen3.8-27B-MTP-GGUF with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "Davidmg0815/Qwen3.8-27B-MTP-GGUF:Q8_0" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
File size: 6,209 Bytes
9751df6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 | #!/usr/bin/env python3
"""Mathe-Diagramme fuer die HuggingFace-Seite.
Zwei Bilder:
1. MATH-500 je Variante mit 95-Prozent-Intervall — zeigt, dass die
Varianten sich nicht unterscheiden
2. Trefferquote nach Schwierigkeitsstufe — zeigt, dass der Benchmark
sehr wohl differenziert, nur eben nicht zwischen diesen Modellen
Bild 2 ist das eigentlich interessante: eine einzelne Prozentzahl verschweigt,
dass die Modelle auf Stufe 1 bei knapp 90 Prozent liegen und auf Stufe 5 bei
gut der Haelfte.
"""
import glob
import json
import math
import os
import sys
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
sys.path.insert(0, "/mnt/models/qwen38-prep")
from diagramm import HELL, DUNKEL, NAMEN, stil
PREP = "/mnt/models/qwen38-prep"
def lade():
d = {}
for p in glob.glob(os.path.join(PREP, "mathe-math500-neubewertet", "*.json")):
k = os.path.splitext(os.path.basename(p))[0]
if k in NAMEN or k in ("q36-35b-a3b", "gemma4-12b"):
d[k] = json.load(open(p))
return d
def kurzname(k, j):
if k in NAMEN:
return NAMEN[k][0]
return j["modell"].split(" (")[0]
def ist38(k):
return NAMEN[k][1] if k in NAMEN else False
def bild_quoten(d, f, ziel):
zeilen = sorted(
((kurzname(k, j), ist38(k), j["quote"], j["bestanden"], j["n"],
j.get("am_limit", 0))
for k, j in d.items()), key=lambda z: z[2])
if not zeilen:
return False
fig, ax = plt.subplots(figsize=(8.6, 0.55 * len(zeilen) + 2.0), dpi=170)
fig.patch.set_facecolor(f["surface"])
y = list(range(len(zeilen)))
ax.barh(y, [z[2] for z in zeilen], height=0.58, zorder=3,
color=[f["akzent"] if z[1] else f["grau"] for z in zeilen])
hoechst = max(z[2] for z in zeilen)
for i, z in enumerate(zeilen):
p = z[3] / z[4]
ci = 1.96 * math.sqrt(p * (1 - p) / z[4]) * 100
ax.plot([z[2] - ci, z[2] + ci], [i, i], color=f["ink2"],
linewidth=1.2, zorder=5)
for x in (z[2] - ci, z[2] + ci):
ax.plot([x, x], [i - 0.13, i + 0.13], color=f["ink2"],
linewidth=1.2, zorder=5)
ax.text(z[2] + ci + hoechst * 0.02, i, f"{z[2]:.1f} %", va="center",
ha="left", color=f["ink"], fontsize=10, fontweight="600")
ax.text(hoechst * 0.025, i, f"{z[3]}/{z[4]}", va="center", ha="left",
color=f["surface"], fontsize=8.5, zorder=6)
ax.set_yticks(y)
ax.set_yticklabels([z[0] for z in zeilen], color=f["ink"], fontsize=10)
ax.set_xlim(0, hoechst * 1.30)
stil(ax, f, "MATH-500 · gelöste Aufgaben",
"Anteil gelöst in Prozent · Strich = 95-Prozent-Intervall")
abgeschnitten = sum(z[5] for z in zeilen) / len(zeilen)
fig.text(0.012, 0.015,
f"Alle Intervalle überlappen. Im Schnitt {abgeschnitten:.0f} von "
f"{zeilen[0][4]} Antworten liefen ins Token-Limit und zählen als "
f"falsch — die absoluten Werte liegen daher zu niedrig.",
color=f["muted"], fontsize=8)
fig.tight_layout(rect=(0, 0.05, 1, 1))
fig.savefig(ziel, facecolor=f["surface"])
plt.close(fig)
return True
def bild_stufen(d, f, ziel):
"""Trefferquote nach Schwierigkeitsstufe. Hier trennt der Benchmark."""
stufen = sorted({s for j in d.values() for s in j.get("nach_level", {})})
if not stufen:
return False
fig, ax = plt.subplots(figsize=(8.6, 5.2), dpi=170)
fig.patch.set_facecolor(f["surface"])
# Farbe traegt die Generation, Strichart und Markerform die Variante.
# Farbe allein reicht hier nicht: mehrere Qwen3.8-Varianten teilen sich
# denselben Blauton und waeren in einem Liniendiagramm nicht trennbar.
striche = ["-", "--", ":", "-."]
marker = ["o", "s", "^", "D"]
zaehler = {True: 0, False: 0}
for k, j in sorted(d.items()):
nl = j.get("nach_level", {})
xs, ys = [], []
for s in stufen:
if s in nl and nl[s][1]:
xs.append(int(s))
ys.append(nl[s][0] / nl[s][1] * 100)
if not xs:
continue
g = ist38(k)
i = zaehler[g] % 4
zaehler[g] += 1
ax.plot(xs, ys, marker=marker[i], markersize=7, linewidth=2,
linestyle=striche[i],
color=f["akzent"] if g else f["grau"],
markeredgecolor=f["surface"], markeredgewidth=1.5,
label=kurzname(k, j), zorder=3)
ax.set_facecolor(f["surface"])
ax.set_title("MATH-500 nach Schwierigkeitsstufe", color=f["ink"],
fontsize=12.5, pad=12, loc="left", fontweight="600")
ax.set_xlabel("Schwierigkeitsstufe (1 = leicht, 5 = Wettbewerbsniveau)",
color=f["ink2"], fontsize=9.5)
ax.set_ylabel("gelöst in Prozent", color=f["ink2"], fontsize=9.5)
ax.set_xticks([int(s) for s in stufen])
ax.tick_params(colors=f["muted"], labelsize=9.5, length=0)
for s in ("top", "right"):
ax.spines[s].set_visible(False)
for s in ("left", "bottom"):
ax.spines[s].set_color(f["achse"])
ax.grid(True, color=f["grid"], linewidth=1)
ax.set_axisbelow(True)
ax.set_ylim(0, 100)
leg = ax.legend(frameon=False, fontsize=9, loc="lower left")
for t in leg.get_texts():
t.set_color(f["ink2"])
fig.text(0.012, 0.015,
"Der Benchmark trennt sauber über die Stufen — nur eben nicht "
"zwischen diesen Varianten. Blau: Qwen3.8, Grau: übrige",
color=f["muted"], fontsize=8)
fig.tight_layout(rect=(0, 0.045, 1, 1))
fig.savefig(ziel, facecolor=f["surface"])
plt.close(fig)
return True
def main():
d = lade()
print(f"MATH-500: {len(d)} Modelle")
if not d:
return 1
ziel = os.path.join(PREP, "bilder")
os.makedirs(ziel, exist_ok=True)
for name, f in (("hell", HELL), ("dunkel", DUNKEL)):
for bau, datei in ((bild_quoten, f"mathe-{name}.png"),
(bild_stufen, f"mathe-stufen-{name}.png")):
if bau(d, f, os.path.join(ziel, datei)):
print(f" {datei}")
return 0
if __name__ == "__main__":
sys.exit(main())
|