Instructions to use motionlabs/AMD-Llama-350M-Upgraded with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use motionlabs/AMD-Llama-350M-Upgraded with Transformers:

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

pipe = pipeline("text-generation", model="motionlabs/AMD-Llama-350M-Upgraded")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("motionlabs/AMD-Llama-350M-Upgraded")
model = AutoModelForCausalLM.from_pretrained("motionlabs/AMD-Llama-350M-Upgraded", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use motionlabs/AMD-Llama-350M-Upgraded with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "motionlabs/AMD-Llama-350M-Upgraded"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "motionlabs/AMD-Llama-350M-Upgraded",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/motionlabs/AMD-Llama-350M-Upgraded

SGLang

How to use motionlabs/AMD-Llama-350M-Upgraded 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 "motionlabs/AMD-Llama-350M-Upgraded" \
    --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": "motionlabs/AMD-Llama-350M-Upgraded",
		"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 "motionlabs/AMD-Llama-350M-Upgraded" \
        --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": "motionlabs/AMD-Llama-350M-Upgraded",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use motionlabs/AMD-Llama-350M-Upgraded with Docker Model Runner:
```
docker model run hf.co/motionlabs/AMD-Llama-350M-Upgraded
```

Oscar Wang commited on Oct 5, 2024

Commit

c168c7b

verified ·

1 Parent(s): 0c5e284

Update README.md

Browse files

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -12,13 +12,13 @@ tags:
 ## Model Description
-The **AMD Llama 350M Upgraded** is a transformer-based causal language model built on the Llama architecture, designed to generate human-like text. This model has been upgraded from the original AMD Llama 135M model to provide enhanced performance with an increased parameter count of 320 million. It is suitable for various natural language processing tasks, including text generation, completion, and conversational applications.
 ## Model Details
 - **Model Type**: Causal Language Model
 - **Architecture**: Llama
-- **Number of Parameters**: 322 million
 - **Input Size**: Variable-length input sequences
 - **Output Size**: Variable-length output sequences

 ## Model Description
+The **AMD Llama 350M Upgraded** is a transformer-based causal language model built on the Llama architecture, designed to generate human-like text. This model has been upgraded from the original AMD Llama 135M model to provide enhanced performance with an increased parameter count of 332 million. It is suitable for various natural language processing tasks, including text generation, completion, and conversational applications.
 ## Model Details
 - **Model Type**: Causal Language Model
 - **Architecture**: Llama
+- **Number of Parameters**: 332 million
 - **Input Size**: Variable-length input sequences
 - **Output Size**: Variable-length output sequences