Instructions to use vogel61/GLM-5.3-oQ4e with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use vogel61/GLM-5.3-oQ4e with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("vogel61/GLM-5.3-oQ4e")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use vogel61/GLM-5.3-oQ4e with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "vogel61/GLM-5.3-oQ4e"

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "vogel61/GLM-5.3-oQ4e"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use vogel61/GLM-5.3-oQ4e with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "vogel61/GLM-5.3-oQ4e"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "vogel61/GLM-5.3-oQ4e"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "vogel61/GLM-5.3-oQ4e",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use vogel61/GLM-5.3-oQ4e with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "vogel61/GLM-5.3-oQ4e"

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 vogel61/GLM-5.3-oQ4e

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use vogel61/GLM-5.3-oQ4e with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "vogel61/GLM-5.3-oQ4e"

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 "vogel61/GLM-5.3-oQ4e" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

what's the speed of prefill / generation on M3U 512?

by Janumacky - opened 18 days ago

Discussion

Janumacky

18 days ago

what's the speed of prefill / generation on M3U 512?
thanks in advance

vogel61

Owner 15 days ago

omlx
Benchmark Model: GLM-5.3-oQ4e-mtp
Engine: Auto
Context: Code (Python)

Single Request Results

Test TTFT(ms) TPOT(ms) pp TPS tg TPS E2E(s) Throughput Peak Mem
pp1024/tg128 5207.3 55.98 196.6 tok/s 18.0 tok/s 12.321 93.5 tok/s 404.92 GB
pp4096/tg128 18749.0 71.61 218.5 tok/s 14.1 tok/s 27.847 151.7 tok/s 406.76 GB
pp8192/tg128 41426.3 76.39 197.7 tok/s 13.2 tok/s 51.133 162.7 tok/s 408.07 GB
pp16384/tg128 88869.0 68.51 184.4 tok/s 14.7 tok/s 97.577 169.2 tok/s 408.88 GB
pp32768/tg128 185967.2 74.07 176.2 tok/s 13.6 tok/s 195.382 168.4 tok/s 410.39 GB
pp65536/tg128 381613.4 73.08 171.7 tok/s 13.8 tok/s 390.902 168.0 tok/s 413.16 GB
pp131072/tg128 801708.5 79.08 163.5 tok/s 12.7 tok/s 811.758 161.6 tok/s 422.16 GB

Continuous Batching
pp1024 / tg128

Batch tg TPS Speedup pp TPS pp TPS/req TTFT(ms) E2E(s)
1x 18.0 tok/s 1.00x 196.6 tok/s 196.6 tok/s 5207.3 12.321
2x 27.2 tok/s 1.51x 87.4 tok/s 43.7 tok/s 15363.6 32.869
4x 36.2 tok/s 2.01x 77.0 tok/s 19.3 tok/s 37009.1 67.316
8x 52.5 tok/s 2.92x 88.3 tok/s 11.0 tok/s 63282.2 112.310

Janumacky

12 days ago

thanks!

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment

what's the speed of prefill / generation on M3U 512?

oMLX - LLM inference, optimized for your Machttps://github.com/jundot/omlxBenchmark Model: GLM-5.3-oQ4e-mtpEngine: AutoContext: Code (Python)

Single Request Results

Continuous Batchingpp1024 / tg128

oMLX - LLM inference, optimized for your Mac
https://github.com/jundot/omlx
Benchmark Model: GLM-5.3-oQ4e-mtp
Engine: Auto
Context: Code (Python)

Continuous Batching
pp1024 / tg128