Instructions to use midorin-Linux/gpt-oss-20b-Coding-Distill with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use midorin-Linux/gpt-oss-20b-Coding-Distill with Transformers:

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

pipe = pipeline("text-generation", model="midorin-Linux/gpt-oss-20b-Coding-Distill")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("midorin-Linux/gpt-oss-20b-Coding-Distill")
model = AutoModelForCausalLM.from_pretrained("midorin-Linux/gpt-oss-20b-Coding-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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use midorin-Linux/gpt-oss-20b-Coding-Distill with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "midorin-Linux/gpt-oss-20b-Coding-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": "midorin-Linux/gpt-oss-20b-Coding-Distill",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/midorin-Linux/gpt-oss-20b-Coding-Distill

SGLang

How to use midorin-Linux/gpt-oss-20b-Coding-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 "midorin-Linux/gpt-oss-20b-Coding-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": "midorin-Linux/gpt-oss-20b-Coding-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 "midorin-Linux/gpt-oss-20b-Coding-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": "midorin-Linux/gpt-oss-20b-Coding-Distill",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Studio new

How to use midorin-Linux/gpt-oss-20b-Coding-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 midorin-Linux/gpt-oss-20b-Coding-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 midorin-Linux/gpt-oss-20b-Coding-Distill to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for midorin-Linux/gpt-oss-20b-Coding-Distill to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="midorin-Linux/gpt-oss-20b-Coding-Distill",
    max_seq_length=2048,
)

Docker Model Runner
How to use midorin-Linux/gpt-oss-20b-Coding-Distill with Docker Model Runner:
```
docker model run hf.co/midorin-Linux/gpt-oss-20b-Coding-Distill
```

midorin-Linux commited on Jan 10

Commit

21fc28a

verified ·

1 Parent(s): faa939a

Update README.md

Browse files

Files changed (1) hide show

README.md +0 -27

README.md CHANGED Viewed

@@ -38,30 +38,3 @@ Traditional fine-tuning often suffers from:
 - **Catastrophic forgetting** when training on sequential datasets
 - **Imbalanced capabilities** from single-source training
 - **Style inconsistencies** across different task types
-Our multi-phase approach with strategic layer freezing, replay buffers, and EWC regularization addresses these challenges systematically.
-## Architecture
-```text
-GPT-OSS-20B Base Model
-│
-├─── Phase 1: Foundation Training
-│    ├─ Data: GPT-5.2-codex-max (1000) + Claude 4.5 Opus (250) + Claude 4.5 Sonnet (250)
-│    ├─ Layers: MLP + Attention
-│    └─ Goal: Establish coding + reasoning foundation
-│
-├─── Phase 1.5: Knowledge Consolidation
-│    ├─ Data: Mixed replay of Phase 1 data
-│    ├─ Layers: MLP + Attention
-│    └─ Goal: Prevent early forgetting
-│
-├─── Phase 2: Specialization Training
-│    ├─ Data: Claude Sonnet (250) + GPT-5.2 high (250) + Replay (150)
-│    ├─ Layers: MLP + Adapter
-│    └─ Goal: Integrate balanced reasoning + maintain coding
-│
-└─── Phase 2.5: Gradual Unfreezing
-     ├─ Data: Full mixed dataset
-     ├─ Layers: Upper Attention layers + MLP + Adapter
-     └─ Goal: Fine-tune attention patterns if needed
-```

 - **Catastrophic forgetting** when training on sequential datasets
 - **Imbalanced capabilities** from single-source training
 - **Style inconsistencies** across different task types