Instructions to use SynLayers/Bbox-caption-8b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SynLayers/Bbox-caption-8b with Transformers:

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

pipe = pipeline("image-text-to-text", model="SynLayers/Bbox-caption-8b")
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("SynLayers/Bbox-caption-8b")
model = AutoModelForImageTextToText.from_pretrained("SynLayers/Bbox-caption-8b")
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 SynLayers/Bbox-caption-8b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SynLayers/Bbox-caption-8b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SynLayers/Bbox-caption-8b",
		"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/SynLayers/Bbox-caption-8b

SGLang

How to use SynLayers/Bbox-caption-8b 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 "SynLayers/Bbox-caption-8b" \
    --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": "SynLayers/Bbox-caption-8b",
		"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 "SynLayers/Bbox-caption-8b" \
        --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": "SynLayers/Bbox-caption-8b",
		"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 SynLayers/Bbox-caption-8b with Docker Model Runner:
```
docker model run hf.co/SynLayers/Bbox-caption-8b
```

SynLayers commited on 7 days ago

Commit

65cf82e

verified ·

1 Parent(s): f87e805

Upload demo/real_world_pipeline.py with huggingface_hub

Browse files

Files changed (1) hide show

demo/real_world_pipeline.py +21 -4

demo/real_world_pipeline.py CHANGED Viewed

@@ -23,6 +23,7 @@ from demo.infer.run_caption_bbox_infer import (  # noqa: E402
     draw_boxes,
     infer_caption_bbox,
 )
 from demo.infer.vlm_bbox_inference import get_model_and_processor  # noqa: E402
 from infer.common_infer import initialize_pipeline  # noqa: E402
 from infer.infer import build_run_save_dir, get_real_boxes, load_adapter_image  # noqa: E402
@@ -133,37 +134,48 @@ def build_runtime_config(
     seed: int | None = None,
 ) -> dict:
     config = load_config(str(config_path))
     config["data_dir"] = str(image_dir.parent)
     config["image_dir"] = str(image_dir)
     config["test_jsonl"] = str(bbox_jsonl)
     config["save_dir"] = str(results_root)
     config["run_name"] = run_name
-    config["lora_ckpt"] = str(DEFAULT_DECOMP_CKPT_ROOT / "transformer")
-    config["layer_ckpt"] = str(DEFAULT_DECOMP_CKPT_ROOT)
-    config["adapter_lora_dir"] = str(DEFAULT_DECOMP_CKPT_ROOT / "adapter")
     env_overrides = {
         "pretrained_model_name_or_path": (
             os.environ.get("SYNLAYERS_BASE_MODEL")
             or resolve_existing_path(PROJECT_ROOT / "SynLayers_checkpoints" / "FLUX.1-dev")
             or "black-forest-labs/FLUX.1-dev"
         ),
         "pretrained_adapter_path": (
             os.environ.get("SYNLAYERS_ADAPTER_MODEL")
             or resolve_existing_path(
                 PROJECT_ROOT / "SynLayers_checkpoints" / "FLUX.1-dev-Controlnet-Inpainting-Alpha"
             )
         ),
         "transp_vae_path": (
             os.environ.get("SYNLAYERS_TRANSP_VAE")
             or resolve_existing_path(PROJECT_ROOT / "ckpt" / "trans_vae" / "0008000.pt")
         ),
         "pretrained_lora_dir": (
             os.environ.get("SYNLAYERS_PRETRAINED_LORA")
             or resolve_existing_path(PROJECT_ROOT / "ckpt" / "pre_trained_LoRA")
         ),
         "artplus_lora_dir": (
             os.environ.get("SYNLAYERS_ARTPLUS_LORA")
             or resolve_existing_path(PROJECT_ROOT / "ckpt" / "prism_ft_LoRA")
         ),
     }
@@ -338,7 +350,12 @@ def run_real_world_pipeline(
     if not image_path.exists():
         raise FileNotFoundError(f"Input image not found: {image_path}")
-    bbox_model = bbox_model or os.environ.get("SYNLAYERS_BBOX_MODEL", DEFAULT_BBOX_MODEL)
     config_path = Path(config_path or os.environ.get("SYNLAYERS_REAL_CONFIG", str(DEFAULT_REAL_CONFIG_PATH)))
     work_dir = Path(work_dir or os.environ.get("SYNLAYERS_DEMO_WORK_DIR", str(DEFAULT_WORK_DIR)))

     draw_boxes,
     infer_caption_bbox,
 )
+from demo.hf_repo_assets import build_repo_asset_overrides  # noqa: E402
 from demo.infer.vlm_bbox_inference import get_model_and_processor  # noqa: E402
 from infer.common_infer import initialize_pipeline  # noqa: E402
 from infer.infer import build_run_save_dir, get_real_boxes, load_adapter_image  # noqa: E402
     seed: int | None = None,
 ) -> dict:
     config = load_config(str(config_path))
+    repo_overrides = build_repo_asset_overrides(os.environ.get("SYNLAYERS_MODEL_REPO"))
+    decomp_ckpt_root = Path(
+        os.environ.get("SYNLAYERS_DECOMP_CKPT_ROOT")
+        or repo_overrides.get("decomp_ckpt_root")
+        or DEFAULT_DECOMP_CKPT_ROOT
+    )
     config["data_dir"] = str(image_dir.parent)
     config["image_dir"] = str(image_dir)
     config["test_jsonl"] = str(bbox_jsonl)
     config["save_dir"] = str(results_root)
     config["run_name"] = run_name
+    config["lora_ckpt"] = str(decomp_ckpt_root / "transformer")
+    config["layer_ckpt"] = str(decomp_ckpt_root)
+    config["adapter_lora_dir"] = str(decomp_ckpt_root / "adapter")
     env_overrides = {
         "pretrained_model_name_or_path": (
             os.environ.get("SYNLAYERS_BASE_MODEL")
+            or repo_overrides.get("pretrained_model_name_or_path")
             or resolve_existing_path(PROJECT_ROOT / "SynLayers_checkpoints" / "FLUX.1-dev")
             or "black-forest-labs/FLUX.1-dev"
         ),
         "pretrained_adapter_path": (
             os.environ.get("SYNLAYERS_ADAPTER_MODEL")
+            or repo_overrides.get("pretrained_adapter_path")
             or resolve_existing_path(
                 PROJECT_ROOT / "SynLayers_checkpoints" / "FLUX.1-dev-Controlnet-Inpainting-Alpha"
             )
         ),
         "transp_vae_path": (
             os.environ.get("SYNLAYERS_TRANSP_VAE")
+            or repo_overrides.get("transp_vae_path")
             or resolve_existing_path(PROJECT_ROOT / "ckpt" / "trans_vae" / "0008000.pt")
         ),
         "pretrained_lora_dir": (
             os.environ.get("SYNLAYERS_PRETRAINED_LORA")
+            or repo_overrides.get("pretrained_lora_dir")
             or resolve_existing_path(PROJECT_ROOT / "ckpt" / "pre_trained_LoRA")
         ),
         "artplus_lora_dir": (
             os.environ.get("SYNLAYERS_ARTPLUS_LORA")
+            or repo_overrides.get("artplus_lora_dir")
             or resolve_existing_path(PROJECT_ROOT / "ckpt" / "prism_ft_LoRA")
         ),
     }
     if not image_path.exists():
         raise FileNotFoundError(f"Input image not found: {image_path}")
+    bbox_model = (
+        bbox_model
+        or os.environ.get("SYNLAYERS_BBOX_MODEL")
+        or os.environ.get("SYNLAYERS_MODEL_REPO")
+        or DEFAULT_BBOX_MODEL
+    )
     config_path = Path(config_path or os.environ.get("SYNLAYERS_REAL_CONFIG", str(DEFAULT_REAL_CONFIG_PATH)))
     work_dir = Path(work_dir or os.environ.get("SYNLAYERS_DEMO_WORK_DIR", str(DEFAULT_WORK_DIR)))