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

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
```

Stack-2-9-finetuned

File size: 2,299 Bytes

b6ae7b8
 
 
 
 
 
 
 
bfc7d04
b6ae7b8
bfc7d04
 
 
 
b6ae7b8
 
bfc7d04
b6ae7b8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bfc7d04
 
b6ae7b8
bfc7d04
 
b6ae7b8
bfc7d04
b6ae7b8
 
 
bfc7d04
 
 
b6ae7b8
 
 
 
 
 
 
 
 
bfc7d04
b6ae7b8
 
 
bfc7d04
b6ae7b8
bfc7d04
b6ae7b8
bfc7d04

# Stack 2.9 Training Configuration

# Model Configuration
model:
  name: "Qwen/Qwen2.5-Coder-32B"
  trust_remote_code: true
  torch_dtype: "bfloat16"

# Data Configuration - supports multiple training files
data:
  train_files:
    - "/Users/walidsobhi/.openclaw/workspace/stack-2.9/training-data/final/train.jsonl"
  val_file: "/Users/walidsobhi/.openclaw/workspace/stack-2.9/training-data/final/val.jsonl"
  test_file: "/Users/walidsobhi/.openclaw/workspace/stack-2.9/training-data/final/test.jsonl"
  train_dir: "/Users/walidsobhi/.openclaw/workspace/stack-2.9-training/data/train"
  eval_dir: "/Users/walidsobhi/.openclaw/workspace/stack-2.9-training/data/eval"
  max_length: 4096  # Reduced for practical training on MPS/consumer GPUs
  train_split: 0.9
  test_split: 0.1

# LoRA Configuration
lora:
  r: 64
  alpha: 128
  dropout: 0.05
  target_modules:
    - "q_proj"
    - "k_proj"
    - "v_proj"
    - "o_proj"
    - "gate_proj"
    - "up_proj"
    - "down_proj"
  bias: "none"
  task_type: "CAUSAL_LM"

# Training Configuration
training:
  num_epochs: 3
  batch_size: 1
  gradient_accumulation: 16
  learning_rate: 1.0e-4
  warmup_steps: 100
  weight_decay: 0.01
  max_grad_norm: 1.0
  logging_steps: 10
  eval_steps: 500
  save_steps: 1000
  save_total_limit: 3
  fp16: false
  bf16: true
  gradient_checkpointing: true
  optim: "adamw_torch"

# Output Configuration
output:
  lora_dir: "/Users/walidsobhi/.openclaw/workspace/stack-2.9-training/output/stack-2.9-32b-lora"
  merged_dir: "/Users/walidsobhi/.openclaw/workspace/stack-2.9-training/output/stack-2.9-32b-merged"
  awq_dir: "/Users/walidsobhi/.openclaw/workspace/stack-2.9-training/output/stack-2.9-32b-awq"

# Quantization Configuration  
quantization:
  enabled: true
  bits: 4
  group_size: 128

# Logging Configuration
logging:
  report_to: "none"  # Set to "wandb" to enable Weights & Biases tracking
  wandb_project: "stack-2.9-training"
  run_name: null

# Hardware Configuration
hardware:
  device: "cuda"  # Change to "mps" for Mac, "cpu" for CPU
  num_gpus: 1
  use_4bit: true  # Enable 4-bit for 32B model on limited VRAM
  use_8bit: false

# Merge Configuration (for after training completes)
merge:
  enabled: true
  output_dir: "/Users/walidsobhi/.openclaw/workspace/stack-2.9-training/output/stack-2.9-32b-merged"