Instructions to use tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned with Transformers:

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

pipe = pipeline("image-text-to-text", model="tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned")

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned

SGLang

How to use tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned 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 "tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned with Docker Model Runner:
```
docker model run hf.co/tech4humans/deepseek-vl2-tiny-swift-r2-rslora-bf16-merged-tuned
```

samuellimabraz commited on Feb 24, 2025

Commit

c0f2068

verified ·

1 Parent(s): a32f872

Upload processor_config.json

Browse files

Files changed (1) hide show

processor_config.json +116 -0

processor_config.json ADDED Viewed

	@@ -0,0 +1,116 @@

+{
+  "add_special_token": false,
+  "candidate_resolutions": [
+    [
+      384,
+      384
+    ],
+    [
+      384,
+      768
+    ],
+    [
+      768,
+      384
+    ],
+    [
+      384,
+      1152
+    ],
+    [
+      1152,
+      384
+    ],
+    [
+      384,
+      1536
+    ],
+    [
+      1536,
+      384
+    ],
+    [
+      768,
+      768
+    ],
+    [
+      384,
+      1920
+    ],
+    [
+      1920,
+      384
+    ],
+    [
+      384,
+      2304
+    ],
+    [
+      2304,
+      384
+    ],
+    [
+      768,
+      1152
+    ],
+    [
+      1152,
+      768
+    ],
+    [
+      384,
+      2688
+    ],
+    [
+      2688,
+      384
+    ],
+    [
+      384,
+      3072
+    ],
+    [
+      3072,
+      384
+    ],
+    [
+      768,
+      1536
+    ],
+    [
+      1536,
+      768
+    ],
+    [
+      384,
+      3456
+    ],
+    [
+      3456,
+      384
+    ],
+    [
+      1152,
+      1152
+    ]
+  ],
+  "downsample_ratio": 2,
+  "ignore_id": -100,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_token": "<image>",
+  "mask_prompt": false,
+  "normalize": true,
+  "pad_token": "<\uff5c\u2581pad\u2581\uff5c>",
+  "patch_size": 14,
+  "processor_class": "DeepseekVLV2Processor",
+  "sft_format": "deepseek"
+}