Instructions to use yujiepan/hy-v3-tiny-random with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use yujiepan/hy-v3-tiny-random with Transformers:

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

pipe = pipeline("text-generation", model="yujiepan/hy-v3-tiny-random")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("yujiepan/hy-v3-tiny-random")
model = AutoModelForCausalLM.from_pretrained("yujiepan/hy-v3-tiny-random")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use yujiepan/hy-v3-tiny-random with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "yujiepan/hy-v3-tiny-random"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "yujiepan/hy-v3-tiny-random",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/yujiepan/hy-v3-tiny-random

SGLang

How to use yujiepan/hy-v3-tiny-random 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 "yujiepan/hy-v3-tiny-random" \
    --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": "yujiepan/hy-v3-tiny-random",
		"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 "yujiepan/hy-v3-tiny-random" \
        --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": "yujiepan/hy-v3-tiny-random",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use yujiepan/hy-v3-tiny-random with Docker Model Runner:
```
docker model run hf.co/yujiepan/hy-v3-tiny-random
```

yujiepan commited on 14 days ago

Commit

811e4b9

verified ·

1 Parent(s): 80d69e7

Upload folder using huggingface_hub

Browse files

Files changed (2) hide show

README.md +35 -2
model.safetensors +2 -2

README.md CHANGED Viewed

@@ -13,6 +13,38 @@ This tiny model is intended for debugging. It is randomly initialized using the
 ### Example usage:
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
@@ -85,8 +117,8 @@ config = AutoConfig.from_pretrained(
 )
 print(config)
 torch.set_default_dtype(torch.bfloat16)
 model = AutoModelForCausalLM.from_config(config, trust_remote_code=True).eval().cpu()
-torch.set_default_dtype(torch.float32)
 if file_exists(filename="generation_config.json", repo_id=source_model_id, repo_type='model'):
     model.generation_config = GenerationConfig.from_pretrained(
         source_model_id, trust_remote_code=True,
@@ -111,13 +143,14 @@ with torch.no_grad():
         print(name, p.shape, p.dtype, f'{p.numel() / n_params * 100: .2f}%')
 # expert bias is in float32
-for i in range(config.first_k_dense_replace, config.num_hidden_layers, 1):
     model.model.layers[i].mlp.e_score_correction_bias = nn.Parameter(torch.randn_like(
         model.model.layers[i].mlp.e_score_correction_bias
     ).float() * 0.002)
 model.save_pretrained(save_folder)
 print(model)
 ```
 </details>

 ### Example usage:
+- vLLM
+```bash
+# Multi-token prediction is supported
+model_id=yujiepan/hy-v3-tiny-random
+vllm serve $model_id \
+  --tensor-parallel-size 2 \
+  --speculative-config.method mtp \
+  --speculative-config.num_speculative_tokens 1 \
+  --tool-call-parser hy_v3 \
+  --reasoning-parser hy_v3 \
+  --enable-auto-tool-choice
+```
+- SGLang
+```bash
+# Multi-token prediction is supported
+model_id=yujiepan/hy-v3-tiny-random
+python3 -m sglang.launch_server \
+  --model tencent/Hy3-preview \
+  --tp 2 \
+  --tool-call-parser hunyuan \
+  --reasoning-parser hunyuan \
+  --speculative-num-steps 1 \
+  --speculative-eagle-topk 1 \
+  --speculative-num-draft-tokens 2 \
+  --speculative-algorithm EAGLE
+```
+- Transformers
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
 )
 print(config)
 torch.set_default_dtype(torch.bfloat16)
+set_seed(42)
 model = AutoModelForCausalLM.from_config(config, trust_remote_code=True).eval().cpu()
 if file_exists(filename="generation_config.json", repo_id=source_model_id, repo_type='model'):
     model.generation_config = GenerationConfig.from_pretrained(
         source_model_id, trust_remote_code=True,
         print(name, p.shape, p.dtype, f'{p.numel() / n_params * 100: .2f}%')
 # expert bias is in float32
+for i in range(config.first_k_dense_replace, config.num_hidden_layers + 1, 1):
     model.model.layers[i].mlp.e_score_correction_bias = nn.Parameter(torch.randn_like(
         model.model.layers[i].mlp.e_score_correction_bias
     ).float() * 0.002)
 model.save_pretrained(save_folder)
 print(model)
+torch.set_default_dtype(torch.float32)
 ```
 </details>

model.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:30b39e65fb3d102afe15a892641b41b7ff8eb4a5f01f5a199dfb0bda397c5401
-size 5401256

 version https://git-lfs.github.com/spec/v1
+oid sha256:d14912ca0a2c0a4487ff365b54caf64ef7885fc06059c225bc0668e7013055f8
+size 5401352