Instructions to use BabyChou/Yi-VL-34B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use BabyChou/Yi-VL-34B with Transformers:

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

pipe = pipeline("text-generation", model="BabyChou/Yi-VL-34B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForCausalLM

processor = AutoProcessor.from_pretrained("BabyChou/Yi-VL-34B")
model = AutoModelForCausalLM.from_pretrained("BabyChou/Yi-VL-34B", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
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 Settings

vLLM

How to use BabyChou/Yi-VL-34B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "BabyChou/Yi-VL-34B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "BabyChou/Yi-VL-34B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/BabyChou/Yi-VL-34B

SGLang

How to use BabyChou/Yi-VL-34B 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 "BabyChou/Yi-VL-34B" \
    --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": "BabyChou/Yi-VL-34B",
		"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 "BabyChou/Yi-VL-34B" \
        --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": "BabyChou/Yi-VL-34B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use BabyChou/Yi-VL-34B with Docker Model Runner:
```
docker model run hf.co/BabyChou/Yi-VL-34B
```

Yi-VL-34B / config.json

BabyChou

Upload folder using huggingface_hub

92c8e0c verified over 2 years ago

Raw

History Blame

1.67 kB

	{
	"_name_or_path": "/home/ec2-user/model_weights/Yi-VL-34B",
	"architectures": [
	"YiVLForCausalLM"
	],
	"attention_bias": false,
	"bos_token_id": 1,
	"eos_token_id": 2,
	"freeze_mm_mlp_adapter": false,
	"hidden_act": "silu",
	"hidden_size": 7168,
	"ignore_index": -100,
	"image_aspect_ratio": "pad",
	"image_grid_pinpoints": null,
	"image_token_index": 64002,
	"initializer_range": 0.02,
	"intermediate_size": 20480,
	"max_position_embeddings": 4096,
	"mm_hidden_size": 1280,
	"mm_projector_type": "mlp2x_gelu_Norm",
	"mm_use_im_patch_token": false,
	"mm_use_im_start_end": false,
	"mm_vision_select_feature": "patch",
	"mm_vision_select_layer": -2,
	"mm_vision_tower": "./vit/clip-vit-H-14-laion2B-s32B-b79K-yi-vl-34B-448",
	"model_type": "llava",
	"moe_layer": null,
	"moe_mode": "easy",
	"num_attention_heads": 56,
	"num_hidden_layers": 60,
	"num_key_value_heads": 8,
	"pretraining_tp": 1,
	"projector_hidden_act": "gelu",
	"rms_norm_eps": 1e-05,
	"rope_scaling": null,
	"rope_theta": 5000000.0,
	"text_config": {
	"model_type": "llama"
	},
	"tie_word_embeddings": false,
	"torch_dtype": "bfloat16",
	"transformers_version": "4.36.1",
	"tune_mm_mlp_adapter": false,
	"tune_vision_tower": true,
	"use_cache": false,
	"use_mm_proj": true,
	"vision_config": {
	"hidden_size": 1024,
	"image_size": 336,
	"intermediate_size": 4096,
	"model_type": "clip_vision_model",
	"num_attention_heads": 16,
	"num_hidden_layers": 24,
	"patch_size": 14,
	"projection_dim": 768,
	"vocab_size": 32000
	},
	"vision_feature_layer": -2,
	"vision_feature_select_strategy": "default",
	"vocab_size": 64000
	}