Instructions to use HuggingFaceTB/SmolVLM-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use HuggingFaceTB/SmolVLM-Base with Transformers:

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

pipe = pipeline("image-text-to-text", model="HuggingFaceTB/SmolVLM-Base")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("HuggingFaceTB/SmolVLM-Base")
model = AutoModelForImageTextToText.from_pretrained("HuggingFaceTB/SmolVLM-Base")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use HuggingFaceTB/SmolVLM-Base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "HuggingFaceTB/SmolVLM-Base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HuggingFaceTB/SmolVLM-Base",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/HuggingFaceTB/SmolVLM-Base

SGLang

How to use HuggingFaceTB/SmolVLM-Base 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 "HuggingFaceTB/SmolVLM-Base" \
    --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": "HuggingFaceTB/SmolVLM-Base",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "HuggingFaceTB/SmolVLM-Base" \
        --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": "HuggingFaceTB/SmolVLM-Base",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use HuggingFaceTB/SmolVLM-Base with Docker Model Runner:
```
docker model run hf.co/HuggingFaceTB/SmolVLM-Base
```

Not able to calculate the "eval_loss" when passed the validation set in the trainer class

by dutta18 - opened Dec 4, 2024

Discussion

dutta18

Dec 4, 2024

I guess the model doesn't return the loss in the forward pass, thats why not able to calculate the "eval_loss" with a val set during the finetuning process. I used the official finetuning notebook provided to finetune the model.

Here is my implementation I tried with a toy dataset:

training_args = TrainingArguments(
num_train_epochs=4,
per_device_train_batch_size=2,
gradient_accumulation_steps=2,
warmup_steps=50,
learning_rate=1e-4,
weight_decay=0.01,
logging_steps=2,
save_strategy="steps",
save_steps=2,
save_total_limit=1,
optim="paged_adamw_8bit", # for 8-bit, keep this, else adamw_hf
bf16=True, # underlying precision for 8bit
output_dir=f"./{model_name}-vqav2",
report_to="tensorboard",
remove_unused_columns=False,
eval_strategy="steps", # Enables evaluation
eval_steps=2, # Frequency of evaluation
load_best_model_at_end=True, # Load the best model at the end
metric_for_best_model="eval_loss", # Monitor this metric
)

trainer = Trainer(
model=model,
args=training_args,
data_collator=collate_fn,
train_dataset=train_set,
eval_dataset=val_set,
)

trainer.train()

When the eval step is reached it goes to calculate the loss I get the below error:
KeyError: "The metric_for_best_model training argument is set to 'eval_loss', which is not found in the evaluation metrics. The available evaluation metrics are: []. Please ensure that the compute_metrics function returns a dictionary that includes 'eval_loss' or consider changing the metric_for_best_model via the TrainingArguments."

Please help me fix this. Or is it something to do with the model itself.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment