glm-edge-4b-chat-mlx-16bit

glm-edge-4b-chat-mlx-16bit

16-bit MLX conversion of THUDM/glm-edge-4b-chat optimized for Apple Silicon native GPU inference.

Converted by: SirSahOl
Source Model: THUDM/glm-edge-4b-chat
Framework: MLX by Apple
Quantization: 16-bit (Average 16.00 (unquantized bfloat16) bits per weight)
Format: safetensors
License: other

---

Model Details

• Architecture: GlmForCausalLM
• Parameters: 4.15B
• Context Length: 8,192 tokens
• Format: MLX (Apple Silicon native GPU format)
• Quantization: 16-bit (Average 16.00 (unquantized bfloat16) bits per weight)
• Active VRAM Footprint: ~9.3 GB (Minimum recommended: 16 GB Unified Memory)

---

Quick Start

Installation

pip install mlx-lm

Usage

CLI

# Chat interactively
mlx_lm.chat --model SirSahOl/glm-edge-4b-chat-mlx-16bit

# Generate text
mlx_lm.generate --model SirSahOl/glm-edge-4b-chat-mlx-16bit --prompt "Write a short poem about artificial intelligence."

Python API (with Chat Template)

from mlx_lm import load, generate

model, tokenizer = load("SirSahOl/glm-edge-4b-chat-mlx-16bit")

messages = [
    {"role": "user", "content": "Explain quantum superposition in simple terms."}
]
prompt = tokenizer.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)

response = generate(model, tokenizer, prompt=prompt, verbose=True)
print(response)

---

Performance Benchmarks

Apple Silicon Hardware Sizing Matrix

Estimated decoding throughput, time-to-first-token (TTFT), and active unified memory footprint across Apple Silicon tiers:

Estimates based on Apple Silicon unified memory bandwidth and active parameter footprint. Real-world speeds may vary with context length.

---

Multi-Quantization Comparison

Evaluate your hardware budget and choose the optimal precision:

---

Who Should Use This?

General guidance:

• Use 4-bit if you want to run this model alongside IDEs, browsers, and background development tools.
• Use 8-bit if you have 16GB+ unified memory and require superior reasoning and code accuracy.
• Use 16-bit for research, benchmarking, evaluation, or high-end workstation deployments.

---

Other Quantization Variants

---

LM Studio & Local Inference Setup Guide

To prevent runaway loops and ensure correct conversational turn-taking, configure custom stop strings in your local inference runtime.

Custom Stop Strings Configuration

Ensure the following sequence tokens are configured as strict stop boundaries:

1. <|user|>
2. <|observation|>
3. <|endoftext|>

Option A: Automatic Preset (Recommended for LM Studio)

Create a custom prompt preset JSON file named GLM-Edge.json in your LM Studio config directory:

• macOS / Linux: ~/.lmstudio/config-presets/GLM-Edge.json
• Windows: %USERPROFILE%\.lmstudio\config-presets\GLM-Edge.json

{
  "name": "GLM-Edge",
  "inference_params": {
    "pre_prompt": "You are a helpful, respectful, and honest assistant.",
    "input_prefix": "<|user|>\n",
    "input_suffix": "",
    "pre_prompt_prefix": "<|system|>\n",
    "pre_prompt_suffix": "\n",
    "antiprompt": [
      "<|user|>",
      "<|observation|>",
      "<|endoftext|>"
    ],
    "stopStrings": [
      "<|user|>",
      "<|observation|>",
      "<|endoftext|>"
    ],
    "temperature": 0.7,
    "max_tokens": 2048
  }
}

Restart LM Studio, load SirSahOl/glm-edge-4b-chat-mlx-16bit, and select "GLM-Edge" from the Prompt Template dropdown.

Option B: Manual LM Studio Configuration

1. Load SirSahOl/glm-edge-4b-chat-mlx-16bit in LM Studio.
2. In the right-hand Advanced Configuration / Inference Parameters panel:
   • Stop Strings (antiprompt / stopStrings): Add <|user|>, <|observation|>, <|endoftext|>
   • System Prefix: <|system|>\n
   • System Suffix: \n
   • User Prefix: <|user|>\n
   • Assistant Suffix: \n<|assistant|>\n

Option C: Ollama Modelfile Setup

Create a Modelfile to run this model in Ollama:

FROM SirSahOl/glm-edge-4b-chat-mlx-16bit

PARAMETER stop "<|user|>"
PARAMETER stop "<|observation|>"
PARAMETER stop "<|endoftext|>"
PARAMETER temperature 0.7

TEMPLATE """{{ if .System }}<|im_start|>system
{{ .System }}<|im_end|>
{{ end }}{{ if .Prompt }}<|im_start|>user
{{ .Prompt }}<|im_end|>
{{ end }}<|im_start|>assistant
"""

Create and run with Ollama:

ollama create glm-edge -f Modelfile
ollama run glm-edge

---

Conversion Details

Reproduction

To reproduce this conversion:

pip install mlx-lm==0.31.3
python3 -m mlx_lm.convert --hf-path /root/.cache/huggingface/hub/models--THUDM--glm-edge-4b-chat/snapshots/a1817f2ab339ecdd8497c4d752ea71a65299f29a --mlx-path /opt/mlx-foundry/output/glm-edge-4b-chat-mlx-16bit

---

Limitations & Known Issues

• 4-bit group-wise quantization introduces minor precision loss compared to unquantized weights; for deep mathematical derivations or precision-critical reasoning, test the 8-bit or 16-bit variants.
• High context sequences (>32K tokens) require sufficient unified memory headroom; ensure unified memory is not overcommitted.
• This is a weight-only MLX conversion designed specifically for Apple Silicon GPUs (M1/M2/M3/M4 series).

---

License

This model conversion inherits the license of the source model: other.

See the original model card for full license details.

---

Changelog

---

Converted with MLX Foundry — a professional pipeline for converting models to Apple MLX format.