Instructions to use Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5 with Transformers:

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

pipe = pipeline("text-generation", model="Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5")
model = AutoModelForCausalLM.from_pretrained("Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5")
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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5

SGLang

How to use Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5 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 "Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5" \
    --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": "Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5",
		"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 "Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5" \
        --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": "Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Studio

How to use Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5 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 Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5 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 Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5 to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5",
    max_seq_length=2048,
)

Docker Model Runner
How to use Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5 with Docker Model Runner:
```
docker model run hf.co/Andy-ML-And-AI/HyperThinkCode-Qwen3-8B-v1.5
```

Andy-ML-And-AI commited on 29 days ago

Commit

c6ca504

verified ·

1 Parent(s): a1b4b3a

Update README.md

Browse files

Files changed (1) hide show

README.md +73 -6

README.md CHANGED Viewed

@@ -6,17 +6,84 @@ tags:
 - unsloth
 - qwen3
 - trl
 license: apache-2.0
 language:
 - en
 ---
-# Uploaded  model
-- **Developed by:** Andy-ML-And-AI
-- **License:** apache-2.0
-- **Finetuned from model :** unsloth/qwen3-8b-unsloth-bnb-4bit
-This qwen3 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth)
-[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)

 - unsloth
 - qwen3
 - trl
+- qlora
+- reasoning
+- code
 license: apache-2.0
 language:
 - en
+datasets:
+- Sashvat/HyperThink-X-Nvidia-Opencode-Reasoning-200K
+metrics:
+- humaneval
+- gsm8k
+library_name: adapter
+pipeline_tag: text-generation
 ---
+# CTD-Qwen3-8B (Code Till Death)
+CTD-Qwen3-8B is a LoRA fine-tune of the Qwen3-8B base model\[cite: 1\].
+---
+## 🛠 Experimental Setup
+- Base model: Qwen3-8B\[cite: 1\]
+- Hardware: dual Tesla T4 (16GB VRAM each)\[cite: 1\]
+- 4-bit QLoRA with rank = 16 and alpha = 16\[cite: 1\]
+- All linear layers:
+  - Attention: q, k, v, o
+  - MLP: gate, up, down
+- Training time: ~1 hour 17 minutes
+- Total steps: 50\[cite: 1\]
+---
+## 🧠 Dataset & Objective
+Training on a specific 30k subset of the
+**Sashvat/HyperThink-X-Nvidia-Opencode-Reasoning-200K** dataset\[cite: 1\].
+- Uses chat template with assistant response in the *thinking* field\[cite: 1\]
+- Objective: encourage *thinking over direct response*\[cite: 1\]
+- Sequence length limited to 4096 tokens (for code complexity + VRAM constraints)\[cite: 1\]
+---
+## 📉 Training Logs
+With only 50 steps, the loss shows expected variance given model + dataset complexity\[cite: 1\].
+| Step | Training Loss |
+|------|--------------|
+| 10   | 0.8177       |
+| 25   | 0.7358       |
+| 50   | 0.6785       |
+- Global batch size: 8 (1 device × 8 gradient steps)\[cite: 1\]
+---
+## 📊 Evaluation (Ongoing)
+Currently running benchmarks using the **lm-eval** library:
+- HumanEval (Coding)
+- GSM8K (Math)
+Comparisons are being made against the base model.
+These evaluations are for internal use within the *Andy Labs* organization.
+---
+## 🔁 Reproduction
+```python
+from unsloth import FastLanguageModel
+model, tokenizer = FastLanguageModel.from_pretrained(
+    model_name = "Andy-ML-And-AI/CTD-Qwen3-8B",
+    max_seq_length = 4096,
+    load_in_4bit = True,
+)