Instructions to use femiari/Llama3MoE with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use femiari/Llama3MoE with Transformers:

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

pipe = pipeline("text-generation", model="femiari/Llama3MoE")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("femiari/Llama3MoE")
model = AutoModelForCausalLM.from_pretrained("femiari/Llama3MoE", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use femiari/Llama3MoE with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "femiari/Llama3MoE"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "femiari/Llama3MoE",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/femiari/Llama3MoE

SGLang

How to use femiari/Llama3MoE 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 "femiari/Llama3MoE" \
    --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": "femiari/Llama3MoE",
		"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 "femiari/Llama3MoE" \
        --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": "femiari/Llama3MoE",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use femiari/Llama3MoE with Docker Model Runner:
```
docker model run hf.co/femiari/Llama3MoE
```

femiari commited on Jun 27, 2024

Commit

ea144f2

verified ·

1 Parent(s): 3ecf165

Upload folder using huggingface_hub

Browse files

Files changed (2) hide show

README.md +68 -0
config.json +54 -0

README.md ADDED Viewed

	@@ -0,0 +1,68 @@

+---
+base_model:
+- Qwen/Qwen-7B
+- TideDra/Qwen-VL-Chat-DPO
+license: apache-2.0
+tags:
+- moe
+- frankenmoe
+- merge
+- mergekit
+- lazymergekit
+- Qwen/Qwen-7B
+- TideDra/Qwen-VL-Chat-DPO
+---
+# QwenMoEAriel
+QwenMoEAriel is a Mixture of Experts (MoE) made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
+* [Qwen/Qwen-7B](https://huggingface.co/Qwen/Qwen-7B)
+* [TideDra/Qwen-VL-Chat-DPO](https://huggingface.co/TideDra/Qwen-VL-Chat-DPO)
+## 🧩 Configuration
+```yaml
+base_model: Qwen/Qwen-7B
+gate_mode: cheap_embed
+experts:
+  - source_model: Qwen/Qwen-7B
+    positive_prompts:
+    - "chat"
+    - "assistant"
+    - "tell me"
+    - "explain"
+    - "I want"
+  - source_model: TideDra/Qwen-VL-Chat-DPO
+    positive_prompts:
+    - "code"
+    - "python"
+    - "javascript"
+    - "programming"
+    - "algorithm"
+shared_experts:
+  - source_model: Qwen/Qwen-7B
+```
+## 💻 Usage
+```python
+!pip install -qU transformers bitsandbytes accelerate
+from transformers import AutoTokenizer
+import transformers
+import torch
+model = "femiari/QwenMoEAriel"
+tokenizer = AutoTokenizer.from_pretrained(model)
+pipeline = transformers.pipeline(
+    "text-generation",
+    model=model,
+    model_kwargs={"torch_dtype": torch.float16, "load_in_4bit": True},
+)
+messages = [{"role": "user", "content": "Explain what a Mixture of Experts is in less than 100 words."}]
+prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
+print(outputs[0]["generated_text"])
+```

config.json ADDED Viewed

	@@ -0,0 +1,54 @@

+{
+  "_name_or_path": "Qwen/Qwen-7B",
+  "architectures": [
+    "MixtralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "attn_dropout_prob": 0.0,
+  "auto_map": {
+    "AutoConfig": "Qwen/Qwen-7B--configuration_qwen.QWenConfig",
+    "AutoModelForCausalLM": "Qwen/Qwen-7B--modeling_qwen.QWenLMHeadModel"
+  },
+  "bf16": false,
+  "bos_token_id": null,
+  "emb_dropout_prob": 0.0,
+  "eos_token_id": null,
+  "fp16": false,
+  "fp32": false,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 22016,
+  "kv_channels": 128,
+  "layer_norm_epsilon": 1e-06,
+  "max_position_embeddings": 32768,
+  "model_type": "mixtral",
+  "no_bias": true,
+  "num_attention_heads": 32,
+  "num_experts_per_tok": 2,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "num_local_experts": 2,
+  "onnx_safe": null,
+  "output_router_logits": false,
+  "rms_norm_eps": 1e-06,
+  "rope_theta": 10000.0,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 1.0,
+  "router_aux_loss_coef": 0.001,
+  "router_jitter_noise": 0.0,
+  "scale_attn_weights": true,
+  "seq_length": 8192,
+  "sliding_window": null,
+  "softmax_in_fp32": false,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "QWenTokenizer",
+  "transformers_version": "4.41.2",
+  "use_cache": true,
+  "use_cache_kernel": false,
+  "use_cache_quantization": false,
+  "use_dynamic_ntk": true,
+  "use_flash_attn": "auto",
+  "use_logn_attn": true,
+  "vocab_size": 151936
+}