Text Generation
Transformers
Safetensors
English
qwen3_moe
text-generation-inference
unsloth
hybrid-thinking
coding-assistant
conversational
Instructions to use Daemontatox/FerrisMind with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Daemontatox/FerrisMind with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Daemontatox/FerrisMind") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("Daemontatox/FerrisMind") model = AutoModelForCausalLM.from_pretrained("Daemontatox/FerrisMind", device_map="auto") 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=256) print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Daemontatox/FerrisMind with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Daemontatox/FerrisMind" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Daemontatox/FerrisMind", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/Daemontatox/FerrisMind
- SGLang
How to use Daemontatox/FerrisMind 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 "Daemontatox/FerrisMind" \ --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": "Daemontatox/FerrisMind", "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 "Daemontatox/FerrisMind" \ --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": "Daemontatox/FerrisMind", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Unsloth Desktop
- Docker Model Runner
How to use Daemontatox/FerrisMind with Docker Model Runner:
docker model run hf.co/Daemontatox/FerrisMind
Update README.md
Browse files
README.md
CHANGED
|
@@ -6,6 +6,8 @@ tags:
|
|
| 6 |
- transformers
|
| 7 |
- unsloth
|
| 8 |
- qwen3_moe
|
|
|
|
|
|
|
| 9 |
license: apache-2.0
|
| 10 |
language:
|
| 11 |
- en
|
|
@@ -25,10 +27,10 @@ new_version: Daemontatox/FerrisMind
|
|
| 25 |
- **Year released:** 2025
|
| 26 |
- **License:** apache-2.0
|
| 27 |
- **Base model:** [unsloth/qwen3-coder-30b-a3b-instruct](https://huggingface.co/unsloth/qwen3-coder-30b-a3b-instruct)
|
| 28 |
-
- **Model type:** Instruction-tuned large language model for code generation
|
| 29 |
|
| 30 |
## Model Summary
|
| 31 |
-
FerrisMind is a finetuned variant of Qwen3 Coder Flash, specialized for **Rust programming**.
|
| 32 |
It is optimized for:
|
| 33 |
- Idiomatic Rust generation
|
| 34 |
- High-performance and memory-safe code practices
|
|
@@ -47,8 +49,8 @@ It is optimized for:
|
|
| 47 |
|
| 48 |
## Training
|
| 49 |
- **Finetuned from:** unsloth/qwen3-coder-30b-a3b-instruct
|
| 50 |
-
- **Objective:** Specialization in Rust code generation and idiomatic best practices
|
| 51 |
-
- **Methods:** Instruction tuning and domain-specific data
|
| 52 |
|
| 53 |
## Limitations
|
| 54 |
- May generate non-compiling Rust code in complex cases
|
|
@@ -66,37 +68,4 @@ async fn main() -> io::Result<()> {
|
|
| 66 |
file.read_to_string(&mut contents).await?;
|
| 67 |
println!("File content: {}", contents);
|
| 68 |
Ok(())
|
| 69 |
-
}
|
| 70 |
-
```
|
| 71 |
-
|
| 72 |
-
```rust
|
| 73 |
-
use std::thread;
|
| 74 |
-
use std::sync::mpsc;
|
| 75 |
-
|
| 76 |
-
fn main() {
|
| 77 |
-
let (tx, rx) = mpsc::channel();
|
| 78 |
-
|
| 79 |
-
for i in 0..5 {
|
| 80 |
-
let tx_clone = tx.clone();
|
| 81 |
-
thread::spawn(move || {
|
| 82 |
-
let message = format!("Message from thread {}", i);
|
| 83 |
-
tx_clone.send(message).unwrap();
|
| 84 |
-
});
|
| 85 |
-
}
|
| 86 |
-
|
| 87 |
-
drop(tx); // Close the original sender
|
| 88 |
-
|
| 89 |
-
for received in rx {
|
| 90 |
-
println!("Got: {}", received);
|
| 91 |
-
}
|
| 92 |
-
}
|
| 93 |
-
```
|
| 94 |
-
|
| 95 |
-
|
| 96 |
-
|
| 97 |
-
@misc{daemontatox2025ferrismind,
|
| 98 |
-
title={FerrisMind: Rust-specialized Qwen3 Coder Finetune},
|
| 99 |
-
author={Daemontatox},
|
| 100 |
-
year={2025},
|
| 101 |
-
howpublished={\url{https://huggingface.co/Daemontatox/FerrisMind}}
|
| 102 |
}
|
|
|
|
| 6 |
- transformers
|
| 7 |
- unsloth
|
| 8 |
- qwen3_moe
|
| 9 |
+
- hybrid-thinking
|
| 10 |
+
- coding-assistant
|
| 11 |
license: apache-2.0
|
| 12 |
language:
|
| 13 |
- en
|
|
|
|
| 27 |
- **Year released:** 2025
|
| 28 |
- **License:** apache-2.0
|
| 29 |
- **Base model:** [unsloth/qwen3-coder-30b-a3b-instruct](https://huggingface.co/unsloth/qwen3-coder-30b-a3b-instruct)
|
| 30 |
+
- **Model type:** Instruction-tuned large language model for code generation, specifically designed to mimic hybrid thinking and utilize it in coding instruct models.
|
| 31 |
|
| 32 |
## Model Summary
|
| 33 |
+
FerrisMind is a finetuned variant of Qwen3 Coder Flash, specialized for **Rust programming**. It was trained using GRPO in an attempt to mimic hybrid thinking and utilize it in coding instruct models.
|
| 34 |
It is optimized for:
|
| 35 |
- Idiomatic Rust generation
|
| 36 |
- High-performance and memory-safe code practices
|
|
|
|
| 49 |
|
| 50 |
## Training
|
| 51 |
- **Finetuned from:** unsloth/qwen3-coder-30b-a3b-instruct
|
| 52 |
+
- **Objective:** Specialization in Rust code generation and idiomatic best practices, mimicking hybrid thinking.
|
| 53 |
+
- **Methods:** Instruction tuning with GRPO and domain-specific data
|
| 54 |
|
| 55 |
## Limitations
|
| 56 |
- May generate non-compiling Rust code in complex cases
|
|
|
|
| 68 |
file.read_to_string(&mut contents).await?;
|
| 69 |
println!("File content: {}", contents);
|
| 70 |
Ok(())
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 71 |
}
|