Instructions to use shaneMattner/Qwen3.6-35B-A3B-RFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use shaneMattner/Qwen3.6-35B-A3B-RFT with Transformers:

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

pipe = pipeline("text-generation", model="shaneMattner/Qwen3.6-35B-A3B-RFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForCausalLM

processor = AutoProcessor.from_pretrained("shaneMattner/Qwen3.6-35B-A3B-RFT")
model = AutoModelForCausalLM.from_pretrained("shaneMattner/Qwen3.6-35B-A3B-RFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use shaneMattner/Qwen3.6-35B-A3B-RFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "shaneMattner/Qwen3.6-35B-A3B-RFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "shaneMattner/Qwen3.6-35B-A3B-RFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/shaneMattner/Qwen3.6-35B-A3B-RFT

SGLang

How to use shaneMattner/Qwen3.6-35B-A3B-RFT 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 "shaneMattner/Qwen3.6-35B-A3B-RFT" \
    --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": "shaneMattner/Qwen3.6-35B-A3B-RFT",
		"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 "shaneMattner/Qwen3.6-35B-A3B-RFT" \
        --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": "shaneMattner/Qwen3.6-35B-A3B-RFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use shaneMattner/Qwen3.6-35B-A3B-RFT with Docker Model Runner:
```
docker model run hf.co/shaneMattner/Qwen3.6-35B-A3B-RFT
```

shaneMattner commited on 1 day ago

Commit

11f9b93

verified ·

1 Parent(s): 80e0957

Upload README.md with huggingface_hub

Browse files

Files changed (1) hide show

README.md +12 -14

README.md CHANGED Viewed

@@ -18,7 +18,7 @@ tags:
 base_model: Qwen/Qwen3.6-35B-A3B
 pipeline_tag: text-generation
 model-index:
-  - name: Qwen3.6-35B-A3B-SSD
     results:
       - task:
           type: text-generation
@@ -34,13 +34,11 @@ model-index:
             value: 0.985
 ---
-# Qwen3.6-35B-A3B-SSD
 A fine-tuned version of [Qwen/Qwen3.6-35B-A3B](https://huggingface.co/Qwen/Qwen3.6-35B-A3B) using **Rejection Fine-Tuning (RFT) on self-generated data**, inspired by the [Simple Self-Distillation (SSD)](https://arxiv.org/abs/2604.01193) paper. The LoRA adapter has been merged into the base weights -- this is a standard bf16 model ready for direct use or quantization.
-> **Note on the repo name**: The repo is named "SSD" because the project started as an SSD replication, but our method deviates from pure SSD in a key way (see below). We kept the name for continuity.
-## What We Actually Did (RFT, Not Pure SSD)
 Our method is **inspired by** the SSD paper ("Embarrassingly Simple Self-Distillation Improves Code Generation", arxiv 2604.01193) but differs in a critical way:
@@ -148,12 +146,12 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
 model = AutoModelForCausalLM.from_pretrained(
-    "shaneMattner/Qwen3.6-35B-A3B-SSD",
     torch_dtype=torch.bfloat16,
     device_map="auto",
     attn_implementation="eager",
 )
-tokenizer = AutoTokenizer.from_pretrained("shaneMattner/Qwen3.6-35B-A3B-SSD")
 messages = [
     {"role": "user", "content": "Write a Python function to merge two sorted lists into one sorted list."}
@@ -173,7 +171,7 @@ pip install mlx-lm
 ```python
 from mlx_lm import load, generate
-model, tokenizer = load("shaneMattner/Qwen3.6-35B-A3B-SSD")
 response = generate(
     model,
     tokenizer,
@@ -188,8 +186,8 @@ Or quantize first for faster inference:
 ```bash
 # Convert to 6-bit MLX format
 python -m mlx_lm.convert \
-    --hf-path shaneMattner/Qwen3.6-35B-A3B-SSD \
-    --mlx-path Qwen3.6-35B-A3B-SSD-6bit \
     -q --q-bits 6
 ```
@@ -201,10 +199,10 @@ Convert to GGUF for use with llama.cpp, Ollama, or other GGUF-compatible tools:
 ```bash
 # Clone llama.cpp and convert
-python convert_hf_to_gguf.py shaneMattner/Qwen3.6-35B-A3B-SSD --outtype bf16
 # Quantize to desired format
-./llama-quantize Qwen3.6-35B-A3B-SSD-bf16.gguf Qwen3.6-35B-A3B-SSD-Q4_K_M.gguf Q4_K_M
 ```
 ## Limitations
@@ -229,10 +227,10 @@ If you use this model, please cite:
 ```bibtex
 @misc{mattner2026qwen36rft,
-  title={Qwen3.6-35B-A3B-SSD: Rejection Fine-Tuned Qwen3.6 for Coding},
   author={Shane Mattner},
   year={2026},
-  url={https://huggingface.co/shaneMattner/Qwen3.6-35B-A3B-SSD}
 }
 ```

 base_model: Qwen/Qwen3.6-35B-A3B
 pipeline_tag: text-generation
 model-index:
+  - name: Qwen3.6-35B-A3B-RFT
     results:
       - task:
           type: text-generation
             value: 0.985
 ---
+# Qwen3.6-35B-A3B-RFT
 A fine-tuned version of [Qwen/Qwen3.6-35B-A3B](https://huggingface.co/Qwen/Qwen3.6-35B-A3B) using **Rejection Fine-Tuning (RFT) on self-generated data**, inspired by the [Simple Self-Distillation (SSD)](https://arxiv.org/abs/2604.01193) paper. The LoRA adapter has been merged into the base weights -- this is a standard bf16 model ready for direct use or quantization.
+## Method (RFT, Not Pure SSD)
 Our method is **inspired by** the SSD paper ("Embarrassingly Simple Self-Distillation Improves Code Generation", arxiv 2604.01193) but differs in a critical way:
 import torch
 model = AutoModelForCausalLM.from_pretrained(
+    "shaneMattner/Qwen3.6-35B-A3B-RFT",
     torch_dtype=torch.bfloat16,
     device_map="auto",
     attn_implementation="eager",
 )
+tokenizer = AutoTokenizer.from_pretrained("shaneMattner/Qwen3.6-35B-A3B-RFT")
 messages = [
     {"role": "user", "content": "Write a Python function to merge two sorted lists into one sorted list."}
 ```python
 from mlx_lm import load, generate
+model, tokenizer = load("shaneMattner/Qwen3.6-35B-A3B-RFT")
 response = generate(
     model,
     tokenizer,
 ```bash
 # Convert to 6-bit MLX format
 python -m mlx_lm.convert \
+    --hf-path shaneMattner/Qwen3.6-35B-A3B-RFT \
+    --mlx-path Qwen3.6-35B-A3B-RFT-6bit \
     -q --q-bits 6
 ```
 ```bash
 # Clone llama.cpp and convert
+python convert_hf_to_gguf.py shaneMattner/Qwen3.6-35B-A3B-RFT --outtype bf16
 # Quantize to desired format
+./llama-quantize Qwen3.6-35B-A3B-RFT-bf16.gguf Qwen3.6-35B-A3B-RFT-Q4_K_M.gguf Q4_K_M
 ```
 ## Limitations
 ```bibtex
 @misc{mattner2026qwen36rft,
+  title={Qwen3.6-35B-A3B-RFT: Rejection Fine-Tuned Qwen3.6 for Coding},
   author={Shane Mattner},
   year={2026},
+  url={https://huggingface.co/shaneMattner/Qwen3.6-35B-A3B-RFT}
 }
 ```