Instructions to use tokyotech-llm/Swallow-70b-NVE-instruct-hf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tokyotech-llm/Swallow-70b-NVE-instruct-hf with Transformers:

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

pipe = pipeline("text-generation", model="tokyotech-llm/Swallow-70b-NVE-instruct-hf")

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("tokyotech-llm/Swallow-70b-NVE-instruct-hf")
model = AutoModelForMultimodalLM.from_pretrained("tokyotech-llm/Swallow-70b-NVE-instruct-hf")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tokyotech-llm/Swallow-70b-NVE-instruct-hf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tokyotech-llm/Swallow-70b-NVE-instruct-hf"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tokyotech-llm/Swallow-70b-NVE-instruct-hf",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/tokyotech-llm/Swallow-70b-NVE-instruct-hf

SGLang

How to use tokyotech-llm/Swallow-70b-NVE-instruct-hf 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 "tokyotech-llm/Swallow-70b-NVE-instruct-hf" \
    --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": "tokyotech-llm/Swallow-70b-NVE-instruct-hf",
		"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 "tokyotech-llm/Swallow-70b-NVE-instruct-hf" \
        --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": "tokyotech-llm/Swallow-70b-NVE-instruct-hf",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use tokyotech-llm/Swallow-70b-NVE-instruct-hf with Docker Model Runner:
```
docker model run hf.co/tokyotech-llm/Swallow-70b-NVE-instruct-hf
```

Taishi-N324 commited on Jan 26, 2024

Commit

d2e04fa

verified ·

1 Parent(s): 57f4c6f

Update config.json

Browse files

Files changed (1) hide show

config.json +2 -2

config.json CHANGED Viewed

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "/bb/llm/gaf51275/llama/from_megatron_hf_checkpoints/hf_checkpoints/Llama-2-70b/iter_0025000",
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -23,7 +23,7 @@
   "rope_theta": 10000.0,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.35.0",
   "use_cache": true,
   "vocab_size": 32000
 }

 {
+  "_name_or_path": "tokyotech-llm/Swallow-70b-NVE-instruct-hf",
   "architectures": [
     "LlamaForCausalLM"
   ],
   "rope_theta": 10000.0,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
+  "transformers_version": "4.33.2",
   "use_cache": true,
   "vocab_size": 32000
 }