Instructions to use my-ai-stack/Stack-4.0-Qwen-3B-Merged with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use my-ai-stack/Stack-4.0-Qwen-3B-Merged with Transformers:

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

pipe = pipeline("text-generation", model="my-ai-stack/Stack-4.0-Qwen-3B-Merged")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("my-ai-stack/Stack-4.0-Qwen-3B-Merged")
model = AutoModelForCausalLM.from_pretrained("my-ai-stack/Stack-4.0-Qwen-3B-Merged", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use my-ai-stack/Stack-4.0-Qwen-3B-Merged with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "my-ai-stack/Stack-4.0-Qwen-3B-Merged"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "my-ai-stack/Stack-4.0-Qwen-3B-Merged",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/my-ai-stack/Stack-4.0-Qwen-3B-Merged

SGLang

How to use my-ai-stack/Stack-4.0-Qwen-3B-Merged 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 "my-ai-stack/Stack-4.0-Qwen-3B-Merged" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "my-ai-stack/Stack-4.0-Qwen-3B-Merged",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "my-ai-stack/Stack-4.0-Qwen-3B-Merged" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "my-ai-stack/Stack-4.0-Qwen-3B-Merged",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use my-ai-stack/Stack-4.0-Qwen-3B-Merged with Docker Model Runner:
```
docker model run hf.co/my-ai-stack/Stack-4.0-Qwen-3B-Merged
```

Walid Sobhi commited on Apr 26

Commit

5fe8bed

verified ·

1 Parent(s): 14c6994

Upload benchmark_results.json with huggingface_hub

Browse files

Files changed (1) hide show

benchmark_results.json +45 -0

benchmark_results.json ADDED Viewed

	@@ -0,0 +1,45 @@

+{
+  "model": "my-ai-stack/Stack-4.0-Qwen-3B-Merged",
+  "date": "2026-04-26",
+  "hardware": "GCP Tesla V100 16GB",
+  "training": {
+    "final_loss": 0.1411,
+    "total_steps": 1000,
+    "effective_batch_size": 16,
+    "learning_rate": 2e-4,
+    "method": "QLoRA",
+    "trainable_params": "7.3M / 3.1B (0.24%)",
+    "training_time": "~10 hours",
+    "cost": "$23 GCP spot instance"
+  },
+  "benchmarks": {
+    "hellaswag": {
+      "acc_norm": 0.74,
+      "acc": 0.52,
+      "n_samples": 50,
+      "note": "50-sample eval (lm_eval --limit 50)"
+    },
+    "arc_challenge": {
+      "acc_norm": 0.52,
+      "acc": 0.48,
+      "n_samples": 50,
+      "note": "50-sample eval"
+    }
+  },
+  "comparison_vs_stack3": {
+    "hellaswag_acc_norm": {
+      "stack_3_0_7b": 0.5961,
+      "stack_4_0_3b": 0.74,
+      "delta": "+14.4%"
+    },
+    "arc_challenge_acc_norm": {
+      "stack_3_0_7b": 0.8328,
+      "stack_4_0_3b": 0.52,
+      "note": "3B model expected lower than 7B"
+    }
+  },
+  "coding_sample": {
+    "score": "10/10",
+    "note": "Internal sample of 10 coding problems — all produced valid code"
+  }
+}