Instructions to use Ebumping/Qwen3-32B-Fable-Distill with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Ebumping/Qwen3-32B-Fable-Distill with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("unsloth/qwen3-32b-bnb-4bit")
model = PeftModel.from_pretrained(base_model, "Ebumping/Qwen3-32B-Fable-Distill")

Transformers

How to use Ebumping/Qwen3-32B-Fable-Distill with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Ebumping/Qwen3-32B-Fable-Distill")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("Ebumping/Qwen3-32B-Fable-Distill")
model = AutoModelForMultimodalLM.from_pretrained("Ebumping/Qwen3-32B-Fable-Distill")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

llama-cpp-python

How to use Ebumping/Qwen3-32B-Fable-Distill with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Ebumping/Qwen3-32B-Fable-Distill",
	filename="Qwen3-32B-Fable-Distill.Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use Ebumping/Qwen3-32B-Fable-Distill with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

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 Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

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 Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

Use Docker

docker model run hf.co/Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

LM Studio
Jan

vLLM

How to use Ebumping/Qwen3-32B-Fable-Distill with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Ebumping/Qwen3-32B-Fable-Distill"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ebumping/Qwen3-32B-Fable-Distill",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

SGLang

How to use Ebumping/Qwen3-32B-Fable-Distill with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Ebumping/Qwen3-32B-Fable-Distill" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ebumping/Qwen3-32B-Fable-Distill",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Ebumping/Qwen3-32B-Fable-Distill" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ebumping/Qwen3-32B-Fable-Distill",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use Ebumping/Qwen3-32B-Fable-Distill with Ollama:
```
ollama run hf.co/Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M
```

Unsloth Studio

How to use Ebumping/Qwen3-32B-Fable-Distill 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 Ebumping/Qwen3-32B-Fable-Distill 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 Ebumping/Qwen3-32B-Fable-Distill to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Ebumping/Qwen3-32B-Fable-Distill to start chatting

How to use Ebumping/Qwen3-32B-Fable-Distill with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Ebumping/Qwen3-32B-Fable-Distill with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

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 Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use Ebumping/Qwen3-32B-Fable-Distill with Docker Model Runner:
```
docker model run hf.co/Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M
```

Lemonade

How to use Ebumping/Qwen3-32B-Fable-Distill with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Ebumping/Qwen3-32B-Fable-Distill:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3-32B-Fable-Distill-Q4_K_M

List all available models

lemonade list

Could YOU?

by usermma - opened 4 days ago

Discussion

usermma

4 days ago

Share the dataset of which you did make this model on it

Ebumping

Owner 3 days ago

•

edited 3 days ago

Mostly Glint and a few friends -p results.

I flattened them together and lobotomized the models reasoning away but I am currently getting traces from anyone who wants to contribute to this, I will stand up a drive with all of the data but I would like to urge anyone willing to drop their traces in as well. Going to include a script you can run to pull them from local if u ran fable in CC. Let's do something really special.

80-90% of the data of this (v0.1) is the Glint reasoning traces. -https://huggingface.co/datasets/Glint-Research/Fable-5-traces

Have some ideas moving forward but I haven't even done any serious benchmarking yet, just asked it goofy questions and liked the results.

usermma

3 days ago

in the meantime, i did try convert this model into MLX, but faced this issue:

TypeError: ModelArgs.init() missing 1 required positional argument: 'rope_theta'

Could you fix?

Ebumping

Owner 3 days ago

https://drive.google.com/drive/folders/1H7hkbYIn_OJ3YCA1OezKtRSt8QbF6jp-?usp=drive_link

Request the permissions you would like, if you select editor I will expect you to make a folder for your contributions like how I did with mine.

I will approve everyone, but this just helps me keep this stuff sorted and attributable.

Put some up if you've got em and I'll be putting more up myself, have a ton of headless execution ones. Would be really great to get more varied response data, more traces could maybe get us some of the deep tool chain behavior.

Ebumping

Owner 3 days ago

and I will work on that

usermma

3 days ago

https://drive.google.com/drive/folders/1H7hkbYIn_OJ3YCA1OezKtRSt8QbF6jp-?usp=drive_link

Request the permissions you would like, if you select editor I will expect you to make a folder for your contributions like how I did with mine.

I will approve everyone, but this just helps me keep this stuff sorted and attributable.

Put some up if you've got em and I'll be putting more up myself, have a ton of headless execution ones. Would be really great to get more varied response data, more traces could maybe get us some of the deep tool chain behavior.

sorry, i may not help in any closed source projects, anything seems to be hidden or closed, i may not, sense i dont have a resources, such alike a api key for good fast ai to get good quality responses, which could help in the model that is gonna be open but not its dataset...

why if this is just a hobby project, or you are kind of a lazy, why not open source it instead of let it alone without no making real progress in the digital intelligence world...?

Ebumping

Owner 3 days ago

literally just reading them before I post them, laziness, working on more than just this. It will all be open soon man I've never open sourced anything in my life have mercy

usermma

3 days ago

i just dont trust giving my email, you could actually show me them in HF, by private it, and give me access to it, if it didnt work, just create an example org in HF and put me there, and make the dataset private just in there, and even if messy, i will read it, just do it, ("still refusing to giving my email address")

Ebumping

Owner 3 days ago

I totally getcha working on it

usermma

3 days ago

awaiting...

Ebumping

Owner 3 days ago

•

edited 3 days ago

collected_shard_*.jsonl
include all of the reasoning traces that were used to supplement the Glint data set in v0.1 training

usermma

3 days ago

i have searched all the files in the Glint data you did provided of
"https://huggingface.co/datasets/Glint-Research/Fable-5-traces"

i found nothing of files starting with "collected_shard_"

usermma

3 days ago

collected_shard_*.jsonl
include all of the reasoning traces that were used to supplement the Glint data set in v0.1 training

Ebumping

Owner 3 days ago

this repo

Ebumping

Owner 3 days ago

•

edited 3 days ago

Plus Glint

Ebumping

Owner 3 days ago

--added
"rope_theta": 1000000
to config.json, this should fix the MLX issue you were having earlier as well

usermma

3 days ago

so this is the secret recipe you use that you dont want to get it open sourced?

and what if i join, what do i get?
do i get a api key for a ai so i can generate a datasets or whats your plan?
and is it a long-term plan or just a hobby?

usermma

3 days ago

or at least if i did generate a dataset using free ai such as qwen, would you train a models on it?

Ebumping

Owner 3 days ago

No interest in traces from anything other than fable being for this model but its not a secret, the way I trained it is straight up on the model card.

GLM5.1 will literally walk you through it if you all it "how to train qwopus" on a rented gpu, it's the same methodology the community has established and documented, just with fable output as data instead of opus.

usermma

3 days ago

this model may be your interest... WeiboAI/VibeThinker-3B

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