Instructions to use OpenNLPLab/TransNormerLLM2-3B-300B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OpenNLPLab/TransNormerLLM2-3B-300B with Transformers:

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

pipe = pipeline("text-generation", model="OpenNLPLab/TransNormerLLM2-3B-300B", trust_remote_code=True)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("OpenNLPLab/TransNormerLLM2-3B-300B", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use OpenNLPLab/TransNormerLLM2-3B-300B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OpenNLPLab/TransNormerLLM2-3B-300B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OpenNLPLab/TransNormerLLM2-3B-300B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/OpenNLPLab/TransNormerLLM2-3B-300B

SGLang

How to use OpenNLPLab/TransNormerLLM2-3B-300B 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 "OpenNLPLab/TransNormerLLM2-3B-300B" \
    --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": "OpenNLPLab/TransNormerLLM2-3B-300B",
		"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 "OpenNLPLab/TransNormerLLM2-3B-300B" \
        --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": "OpenNLPLab/TransNormerLLM2-3B-300B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use OpenNLPLab/TransNormerLLM2-3B-300B with Docker Model Runner:
```
docker model run hf.co/OpenNLPLab/TransNormerLLM2-3B-300B
```

OpenNLPLab commited on Jan 24, 2024

Commit

42b9a76

verified ·

1 Parent(s): 88b9dbc

Fix 3B config error

Browse files

Files changed (1) hide show

config.json +7 -5

config.json CHANGED Viewed

@@ -12,9 +12,9 @@
   "vocab_size": 64000,
   "use_cache": true,
   "init_std": 0.02,
-  "decoder_embed_dim": 4096,
-  "decoder_layers": 30,
-  "decoder_attention_heads": 32,
   "no_scale_embedding": true,
   "add_bos_token": false,
   "norm_type": "simplermsnorm",
@@ -48,11 +48,13 @@
     0,
     0,
     0,
     0
   ],
-  "hidden_dim": 4096,
   "linear_act_fun": "swish",
-  "glu_dim": 11008,
   "bias": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.32.0"

   "vocab_size": 64000,
   "use_cache": true,
   "init_std": 0.02,
+  "decoder_embed_dim": 2560,
+  "decoder_layers": 32,
+  "decoder_attention_heads": 20,
   "no_scale_embedding": true,
   "add_bos_token": false,
   "norm_type": "simplermsnorm",
     0,
     0,
     0,
+    0,
+    0,
     0
   ],
+  "hidden_dim": 2560,
   "linear_act_fun": "swish",
+  "glu_dim": 6912,
   "bias": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.32.0"