Instructions to use my-ai-stack/Stack-2-9-finetuned with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use my-ai-stack/Stack-2-9-finetuned with Transformers:

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

pipe = pipeline("text-generation", model="my-ai-stack/Stack-2-9-finetuned")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("my-ai-stack/Stack-2-9-finetuned")
model = AutoModelForCausalLM.from_pretrained("my-ai-stack/Stack-2-9-finetuned")
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 Settings

vLLM

How to use my-ai-stack/Stack-2-9-finetuned 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-2-9-finetuned"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "my-ai-stack/Stack-2-9-finetuned",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/my-ai-stack/Stack-2-9-finetuned

SGLang

How to use my-ai-stack/Stack-2-9-finetuned 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-2-9-finetuned" \
    --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": "my-ai-stack/Stack-2-9-finetuned",
		"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 "my-ai-stack/Stack-2-9-finetuned" \
        --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": "my-ai-stack/Stack-2-9-finetuned",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use my-ai-stack/Stack-2-9-finetuned with Docker Model Runner:
```
docker model run hf.co/my-ai-stack/Stack-2-9-finetuned
```

walidsobhie-code Claude Opus 4.6 commited on Apr 10

Commit

a075b90

1 Parent(s): 2064035

fix: improve Colab notebook for training

Browse files

- Fix subprocess to use direct shell command for better output
- Add environment variables for GPU optimization
- Add variables for easier token/username configuration
- Improve dependency versions

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Files changed (1) hide show

notebooks/colab_128k_training.ipynb +11 -49

notebooks/colab_128k_training.ipynb CHANGED Viewed

@@ -22,17 +22,7 @@
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": [
-    "import os\n",
-    "os.chdir(\"/content\")\n",
-    "\n",
-    "# Clone repo\n",
-    "!git clone https://github.com/my-ai-stack/stack-2.9.git\n",
-    "\n",
-    "# Install dependencies\n",
-    "!pip install -q transformers peft datasets bitsandbytes accelerate huggingface_hub\n",
-    "!pip install -q scipy torch --upgrade"
-   ]
   },
   {
    "cell_type": "markdown",
@@ -46,12 +36,7 @@
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": [
-    "from huggingface_hub import login\n",
-    "\n",
-    "# Get your token at: https://huggingface.co/settings/tokens\n",
-    "login(token=\"YOUR_HF_TOKEN\")  # ← Replace with your token"
-   ]
   },
   {
    "cell_type": "markdown",
@@ -65,47 +50,24 @@
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": [
-    "# Uncomment to mount Drive\n",
-    "# from google.colab import drive\n",
-    "# drive.mount('/content/drive')\n",
-    "# OUTPUT_DIR = \"/content/drive/MyDrive/stack-2.9-128k-output\"\n",
-    "\n",
-    "# Otherwise saves to /content/stack-2.9/output/\n",
-    "OUTPUT_DIR = \"/content/stack-2.9/output/stack-2.9-128k\""
-   ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": [
-    "## Step 4: Run 128K Context Fine-tuning"
-   ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": [
-    "import subprocess\n",
-    "os.chdir(\"/content/stack-2.9\")\n",
-    "\n",
-    "result = subprocess.run([\n",
-    "    \"python3\", \"training/train_extended_context.py\",\n",
-    "    \"--model-path\", \"Qwen/Qwen2.5-Coder-1.5B\",\n",
-    "    \"--data-path\", \"training/training-data/tool_examples_combined.jsonl\",\n",
-    "    \"--output-dir\", OUTPUT_DIR,\n",
-    "    \"--context-length\", \"131072\",\n",
-    "    \"--lora-rank\", \"64\",\n",
-    "    \"--epochs\", \"3\",\n",
-    "    \"--push-to-hub\",\n",
-    "    \"--hub-model-id\", \"YOUR_USERNAME/stack-2.9-128k\"  # ← Replace with your username\n",
-    "], capture_output=True, text=True)\n",
-    "\n",
-    "print(\"STDOUT:\", result.stdout[-5000:] if result.stdout else \"None\")\n",
-    "print(\"STDERR:\", result.stderr[-2000:] if result.stderr else \"None\")"
-   ]
   },
   {
    "cell_type": "markdown",
@@ -143,4 +105,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 4
-}

    "execution_count": null,
    "metadata": {},
    "outputs": [],
+   "source": "import os\nos.chdir(\"/content\")\n\n# Clone repo\n!git clone https://github.com/my-ai-stack/stack-2.9.git\n\n# Install dependencies\n!pip install -q transformers>=4.40.0 peft datasets bitsandbytes accelerate huggingface_hub\n!pip install -q scipy torch --upgrade\n!pip install -q flash-attn --no-build-isolation  # Optional: for faster attention"
   },
   {
    "cell_type": "markdown",
    "execution_count": null,
    "metadata": {},
    "outputs": [],
+   "source": "from huggingface_hub import login\n\n# Get your token at: https://huggingface.co/settings/tokens\n# Replace with your actual token before running\nHF_TOKEN = \"YOUR_HF_TOKEN\"  # ← Replace with your HF token\nlogin(token=HF_TOKEN)\nprint(\"✓ Logged in to HuggingFace\")"
   },
   {
    "cell_type": "markdown",
    "execution_count": null,
    "metadata": {},
    "outputs": [],
+   "source": "# Check GPU and environment\nimport torch\nprint(f\"✓ PyTorch version: {torch.__version__}\")\nprint(f\"✓ CUDA available: {torch.cuda.is_available()}\")\nif torch.cuda.is_available():\n    print(f\"✓ GPU: {torch.cuda.get_device_name(0)}\")\n    print(f\"✓ VRAM: {torch.cuda.get_device_properties(0).total_memory / 1e9:.1f} GB\")\n\n# Mount Google Drive (optional)\n# from google.colab import drive\n# drive.mount('/content/drive')\n# OUTPUT_DIR = \"/content/drive/MyDrive/stack-2.9-128k-output\"\n\nOUTPUT_DIR = \"/content/stack-2.9/output/stack-2.9-128k\""
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Step 3.5: Verify GPU Setup",
+   "metadata": {}
   },
   {
    "cell_type": "markdown",
    "metadata": {},
+   "source": "import os\nos.chdir(\"/content/stack-2.9\")\n\n# Set environment for better GPU usage\nos.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\"\nos.environ[\"TRANSFORMERS_NO_ADVISORY_WARNINGS\"] = \"true\"\n\n# === UPDATE THESE VALUES BEFORE RUNNING ===\nYOUR_HF_USERNAME = \"your-username\"  # Replace with your HF username\n\n# Run training\n!python training/train_extended_context.py \\\n    --model-path Qwen/Qwen2.5-Coder-1.5B \\\n    --data-path training/training-data/tool_examples_combined.jsonl \\\n    --output-dir /content/stack-2.9/output/stack-2.9-128k \\\n    --context-length 131072 \\\n    --lora-rank 64 \\\n    --epochs 3 \\\n    --batch-size 1 \\\n    --grad-accum 16 \\\n    --push-to-hub \\\n    --hub-model-id YOUR_HF_USERNAME/stack-2.9-128k"
   },
   {
    "cell_type": "code",
    "execution_count": null,
    "metadata": {},
    "outputs": [],
+   "source": "import os\nimport sys\nos.chdir(\"/content/stack-2.9\")\n\n# Set environment for better GPU usage\nos.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\"\nos.environ[\"TRANSFORMERS_NO_ADVISORY_WARNINGS\"] = \"true\"\n\n# Training parameters - UPDATE THESE\nYOUR_HF_TOKEN = \"YOUR_HF_TOKEN\"  # Replace with your token\nYOUR_USERNAME = \"your-username\"   # Replace with your HF username\n\n# Run training directly for better output visibility\n!python training/train_extended_context.py \\\n    --model-path Qwen/Qwen2.5-Coder-1.5B \\\n    --data-path training/training-data/tool_examples_combined.jsonl \\\n    --output-dir /content/stack-2.9/output/stack-2.9-128k \\\n    --context-length 131072 \\\n    --lora-rank 64 \\\n    --epochs 3 \\\n    --batch-size 1 \\\n    --grad-accum 16 \\\n    --push-to-hub \\\n    --hub-model-id {YOUR_USERNAME}/stack-2.9-128k"
   },
   {
    "cell_type": "markdown",
  },
  "nbformat": 4,
  "nbformat_minor": 4
+}