Instructions to use OpenGVLab/Mono-InternVL-2B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OpenGVLab/Mono-InternVL-2B with Transformers:

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

pipe = pipeline("image-text-to-text", model="OpenGVLab/Mono-InternVL-2B", trust_remote_code=True)
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 AutoModel
model = AutoModel.from_pretrained("OpenGVLab/Mono-InternVL-2B", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use OpenGVLab/Mono-InternVL-2B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OpenGVLab/Mono-InternVL-2B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OpenGVLab/Mono-InternVL-2B",
		"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/OpenGVLab/Mono-InternVL-2B

SGLang

How to use OpenGVLab/Mono-InternVL-2B 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 "OpenGVLab/Mono-InternVL-2B" \
    --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": "OpenGVLab/Mono-InternVL-2B",
		"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 "OpenGVLab/Mono-InternVL-2B" \
        --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": "OpenGVLab/Mono-InternVL-2B",
		"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 OpenGVLab/Mono-InternVL-2B with Docker Model Runner:
```
docker model run hf.co/OpenGVLab/Mono-InternVL-2B
```

favor123 commited on Oct 11, 2024

Commit

40e5dc4

verified ·

1 Parent(s): 82e212b

Update modeling_internlm2_ve.py

Browse files

Files changed (1) hide show

modeling_internlm2_ve.py +11 -13

modeling_internlm2_ve.py CHANGED Viewed

@@ -689,20 +689,18 @@ class InternLM2DecoderLayer(nn.Module):
         hidden_states = self.ffn_norm(hidden_states)
         if past_key_value is None:
-            """
-            *************************
-            maybe faster
-            ***************************
-            """
-            ##############################################################################################################
-            # dim=hidden_states.shape[-1]
-            # visual_token_mask=visual_token_mask.repeat(1,1,dim)
-            # if visual_token_mask.any():
-            #     hidden_states[visual_token_mask] = self.feed_forward_ve(hidden_states[visual_token_mask].reshape(-1,dim)).reshape(-1)
-            # if (~visual_token_mask).any():
-            #     hidden_states[~visual_token_mask] = self.feed_forward(hidden_states[~visual_token_mask].reshape(-1,dim)).reshape(-1)
             ##############################################################################################################
-            hidden_states = self.feed_forward(hidden_states)*(1.-visual_token_mask)+ self.feed_forward_ve(hidden_states)*visual_token_mask
         else:
             hidden_states = self.feed_forward(hidden_states)

         hidden_states = self.ffn_norm(hidden_states)
         if past_key_value is None:
+            ##########################################--modified by luogen--##############################################
+            if self.training:
+                hidden_states = self.feed_forward(hidden_states)*(1.-visual_token_mask)+ self.feed_forward_ve(hidden_states)*visual_token_mask
+            else:
+                dim=hidden_states.shape[-1]
+                visual_token_mask=visual_token_mask.repeat(1,1,dim).bool()
+                non_visual_token_mask=~visual_token_mask
+                if visual_token_mask.any():
+                    hidden_states[visual_token_mask] = self.feed_forward_ve(hidden_states[visual_token_mask].reshape(-1,dim)).reshape(-1)
+                if (non_visual_token_mask).any():
+                    hidden_states[non_visual_token_mask] = self.feed_forward(hidden_states[non_visual_token_mask].reshape(-1,dim)).reshape(-1)
             ##############################################################################################################
         else:
             hidden_states = self.feed_forward(hidden_states)