frob/glm-5.1:744b-a40b-ud-q4_K_XL

frob/ glm-5.1:744b-a40b-ud-q4_K_XL

3 Downloads Updated 11 hours ago

tools thinking

ollama run frob/glm-5.1:744b-a40b-ud-q4_K_XL

curl http://localhost:11434/api/chat \
  -d '{
    "model": "frob/glm-5.1:744b-a40b-ud-q4_K_XL",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='frob/glm-5.1:744b-a40b-ud-q4_K_XL',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'frob/glm-5.1:744b-a40b-ud-q4_K_XL',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 11 hours ago

11 hours ago

8c2556e51f3a · 466GB ·

model

archglm-dsa

·

parameters754B

·

quantizationQ4_K_M

466GB

license

MIT License Copyright (c) [year] [fullname] Permission is hereby granted, free of charge, to any per

1.1kB

template

{{- $lastUserIdx := -1 }} {{- range $i, $_ := .Messages }} {{- if eq .Role "user" }}{{- $lastUserIdx

1.5kB

params

{ "stop": [ "<|system|>", "<|user|>", "<|assistant|>" ] }

81B

Readme

Imported from hf.co/unsloth/GLM-5.1-GGUF.

Note that ollama does not yet support this model. To run it, ollama needs to be patched with #14864.

$ git clone https://github.com/ollama/ollama.git .
$ git checkout v0.20.0
$ curl -L https://github.com/ollama/ollama/pull/14864.diff | patch -p1
$ docker build -t ollama/ollama:0.20.0-14864 .

This model wants to use the tool format it was trained with and resists using the simpler JSON format that is easily encoded in an ollama template. For this reason, this model is not a good tool user. Better tool use will be enabled with an ollama PARSER.

$ ollama run frob/glm-5.1
>>> Why don't scientists trust atoms?
Thinking...
1.  **Analyze the Request:** The user is asking a classic riddle/joke: "Why don't 
scientists trust atoms?"
2.  **Identify the Intent:** The intent is humor/wordplay based on a well-known 
scientific joke.
3.  **Retrieve Knowledge:** Access the punchline for this specific joke. The 
standard punchline is "Because they make up everything!"
4.  **Formulate Response:** Deliver the punchline clearly and concisely, perhaps 
with a slight playful tone since it's a joke.
...done thinking.

Because they make up everything!

$ ollama run frob/glm-5.1 hello. --think=false
Hello! How can I help you today? 😊

Imported from [hf.co/unsloth/GLM-5.1-GGUF](https://huggingface.co/unsloth/GLM-5.1-GGUF).

Note that ollama does not yet support this model.  To run it, ollama needs to be patched with [#14864](https://github.com/ollama/ollama/pull/14864).

```console
$ git clone https://github.com/ollama/ollama.git .
$ git checkout v0.20.0
$ curl -L https://github.com/ollama/ollama/pull/14864.diff | patch -p1
$ docker build -t ollama/ollama:0.20.0-14864 .
```

This model wants to use the tool format it was trained with and resists using the simpler JSON format that is easily encoded in an ollama template.  For this reason, this model is *not* a good tool user.  Better tool use will be enabled with an ollama `PARSER`.

```console
$ ollama run frob/glm-5.1
>>> Why don't scientists trust atoms?
Thinking...
1.  **Analyze the Request:** The user is asking a classic riddle/joke: "Why don't 
scientists trust atoms?"
2.  **Identify the Intent:** The intent is humor/wordplay based on a well-known 
scientific joke.
3.  **Retrieve Knowledge:** Access the punchline for this specific joke. The 
standard punchline is "Because they make up everything!"
4.  **Formulate Response:** Deliver the punchline clearly and concisely, perhaps 
with a slight playful tone since it's a joke.
...done thinking.

Because they make up everything!
```

```console
$ ollama run frob/glm-5.1 hello. --think=false
Hello! How can I help you today? 😊
```

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)