Instructions to use Digsm003/model_sft_lora_fv with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Digsm003/model_sft_lora_fv with Transformers:

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

pipe = pipeline("text-generation", model="Digsm003/model_sft_lora_fv")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

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

tokenizer = AutoTokenizer.from_pretrained("Digsm003/model_sft_lora_fv")
model = AutoModelForCausalLM.from_pretrained("Digsm003/model_sft_lora_fv", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Digsm003/model_sft_lora_fv with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Digsm003/model_sft_lora_fv"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Digsm003/model_sft_lora_fv",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Digsm003/model_sft_lora_fv

SGLang

How to use Digsm003/model_sft_lora_fv 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 "Digsm003/model_sft_lora_fv" \
    --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": "Digsm003/model_sft_lora_fv",
		"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 "Digsm003/model_sft_lora_fv" \
        --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": "Digsm003/model_sft_lora_fv",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Digsm003/model_sft_lora_fv with Docker Model Runner:
```
docker model run hf.co/Digsm003/model_sft_lora_fv
```

model_sft_lora_fv

File size: 1,130 Bytes

159e68f

{
  "inject_layer": 9,
  "lambda": 1.0,
  "fv_file": "function_vector.pt",
  "top_heads": [
    {
      "layer": 24,
      "head": 3,
      "aie": 0.004586321767419577
    },
    {
      "layer": 20,
      "head": 9,
      "aie": 0.0027352971956133842
    },
    {
      "layer": 0,
      "head": 6,
      "aie": 0.0017912477487698197
    },
    {
      "layer": 17,
      "head": 7,
      "aie": 0.0017713563283905387
    },
    {
      "layer": 20,
      "head": 8,
      "aie": 0.0016093882732093334
    },
    {
      "layer": 21,
      "head": 3,
      "aie": 0.0014112978242337704
    },
    {
      "layer": 18,
      "head": 10,
      "aie": 0.001295118359848857
    },
    {
      "layer": 17,
      "head": 3,
      "aie": 0.0012491021770983934
    },
    {
      "layer": 15,
      "head": 1,
      "aie": 0.00092243158724159
    },
    {
      "layer": 19,
      "head": 11,
      "aie": 0.0007095862529240549
    }
  ],
  "model_name": "Qwen/Qwen2.5-1.5B-Instruct",
  "description": "Apply FV by: hook = model.model.layers[inject_layer].register_forward_hook(make_fv_hook(torch.load('function_vector.pt'), lambda))"
}