Instructions to use microsoft/Phi-tiny-MoE-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use microsoft/Phi-tiny-MoE-instruct with Transformers:

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

pipe = pipeline("text-generation", model="microsoft/Phi-tiny-MoE-instruct", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-tiny-MoE-instruct", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("microsoft/Phi-tiny-MoE-instruct", trust_remote_code=True)
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 microsoft/Phi-tiny-MoE-instruct with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/microsoft/Phi-tiny-MoE-instruct

SGLang

How to use microsoft/Phi-tiny-MoE-instruct 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 "microsoft/Phi-tiny-MoE-instruct" \
    --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": "microsoft/Phi-tiny-MoE-instruct",
		"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 "microsoft/Phi-tiny-MoE-instruct" \
        --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": "microsoft/Phi-tiny-MoE-instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use microsoft/Phi-tiny-MoE-instruct with Docker Model Runner:
```
docker model run hf.co/microsoft/Phi-tiny-MoE-instruct
```

man2machine commited on Feb 23

Commit

3a421cd

1 Parent(s): 274f8d5

Update working

Browse files

Files changed (2) hide show

configuration_slimmoe.py +1 -1
modeling_slimmoe.py +9 -1

configuration_slimmoe.py CHANGED Viewed

@@ -111,7 +111,7 @@ class PhiMoEConfig(PretrainedConfig):
     >>> configuration = model.config
     ```"""
-    model_type = "phimoe"
     keys_to_ignore_at_inference = ["past_key_values"]
     def __init__(

     >>> configuration = model.config
     ```"""
+    model_type = "phimoe_slim"  # renamed from "phimoe" to bypass transformers >=4.46 conversion_mapping
     keys_to_ignore_at_inference = ["past_key_values"]
     def __init__(

modeling_slimmoe.py CHANGED Viewed

@@ -330,9 +330,17 @@ class PhiMoEAttention(nn.Module):
                 base=self.rope_theta,
             )
         else:
-            scaling_type = self.config.rope_scaling["type"]
             if scaling_type == "longrope":
                 self.rotary_emb = Phi3LongRoPEScaledRotaryEmbedding(self.head_dim, self.config)
             else:
                 raise ValueError(f"Unknown RoPE scaling type {scaling_type}")

                 base=self.rope_theta,
             )
         else:
+            # "type" key was renamed to "rope_type" in transformers >=4.46; handle both
+            scaling_type = self.config.rope_scaling.get("type") or self.config.rope_scaling.get("rope_type", "")
             if scaling_type == "longrope":
                 self.rotary_emb = Phi3LongRoPEScaledRotaryEmbedding(self.head_dim, self.config)
+            elif not scaling_type or scaling_type == "default":
+                # newer transformers injects {"rope_type": "default"} when rope_scaling is absent
+                self.rotary_emb = PhiMoERotaryEmbedding(
+                    self.head_dim,
+                    max_position_embeddings=self.max_position_embeddings,
+                    base=self.rope_theta,
+                )
             else:
                 raise ValueError(f"Unknown RoPE scaling type {scaling_type}")