Instructions to use fuzhao/openmoe_large_tmp with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use fuzhao/openmoe_large_tmp with Transformers:

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

pipe = pipeline("text-generation", model="fuzhao/openmoe_large_tmp", device_map="auto")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("fuzhao/openmoe_large_tmp")
model = AutoModelForCausalLM.from_pretrained("fuzhao/openmoe_large_tmp", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use fuzhao/openmoe_large_tmp with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/fuzhao/openmoe_large_tmp

SGLang

How to use fuzhao/openmoe_large_tmp 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 "fuzhao/openmoe_large_tmp" \
    --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": "fuzhao/openmoe_large_tmp",
		"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 "fuzhao/openmoe_large_tmp" \
        --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": "fuzhao/openmoe_large_tmp",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use fuzhao/openmoe_large_tmp with Docker Model Runner:
```
docker model run hf.co/fuzhao/openmoe_large_tmp
```

fuzhao commited on Sep 8, 2023

Commit

a6d19bf

1 Parent(s): 838471b

First model version

Browse files

Files changed (1) hide show

config.json +38 -0

config.json ADDED Viewed

	@@ -0,0 +1,38 @@

+{
+  "architectures": [
+    "OpenMoeForCausalLM"
+  ],
+  "bos_token_id": 1,
+  "capacity_factor_eval": 2.0,
+  "capacity_factor_train": 1.25,
+  "drop_tks": true,
+  "dropout_rate": 0.0,
+  "eos_token_id": 2,
+  "expert_parallel": null,
+  "gated": true,
+  "head_dim": 128,
+  "hidden_act": "swiglu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "layer_norm_epsilon": 1e-06,
+  "max_position_embeddings": 2048,
+  "min_capacity": 4,
+  "model_type": "llama",
+  "moe_layer_interval": 6,
+  "noisy_policy": null,
+  "num_attention_heads": 24,
+  "num_experts": 32,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 24,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "tie_word_embeddings": false,
+  "topk": 2,
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "use_cache": true,
+  "vocab_size": 256384
+}