Instructions to use steerapi/Llama-2-7b-chat-hf-onnx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use steerapi/Llama-2-7b-chat-hf-onnx with Transformers:

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

pipe = pipeline("text-generation", model="steerapi/Llama-2-7b-chat-hf-onnx")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("steerapi/Llama-2-7b-chat-hf-onnx")
model = AutoModelForCausalLM.from_pretrained("steerapi/Llama-2-7b-chat-hf-onnx", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use steerapi/Llama-2-7b-chat-hf-onnx with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "steerapi/Llama-2-7b-chat-hf-onnx"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "steerapi/Llama-2-7b-chat-hf-onnx",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/steerapi/Llama-2-7b-chat-hf-onnx

SGLang

How to use steerapi/Llama-2-7b-chat-hf-onnx 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 "steerapi/Llama-2-7b-chat-hf-onnx" \
    --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": "steerapi/Llama-2-7b-chat-hf-onnx",
		"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 "steerapi/Llama-2-7b-chat-hf-onnx" \
        --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": "steerapi/Llama-2-7b-chat-hf-onnx",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use steerapi/Llama-2-7b-chat-hf-onnx with Docker Model Runner:
```
docker model run hf.co/steerapi/Llama-2-7b-chat-hf-onnx
```

steerapi commited on Aug 6, 2023

Commit

8d2a77c

1 Parent(s): bd15347

Upload folder using huggingface_hub

Browse files

Files changed (4) hide show

.gitattributes +1 -0
onnx/decoder_model_merged_quantized.onnx +2 -2
onnx/decoder_model_merged_quantized.onnx_data +3 -0
onnx/quantize_config.json +23 -23

.gitattributes CHANGED Viewed

@@ -45,3 +45,4 @@ onnx/q1/decoder_model_merged_quantized.onnx.data filter=lfs diff=lfs merge=lfs -
 onnx/decoder_model_merged_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text
 onnx/decoder_model_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text
 onnx/decoder_with_past_model_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text

 onnx/decoder_model_merged_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text
 onnx/decoder_model_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text
 onnx/decoder_with_past_model_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text
+onnx/decoder_model_merged_quantized.onnx_data filter=lfs diff=lfs merge=lfs -text

onnx/decoder_model_merged_quantized.onnx CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:289772dc7b8a483e82ca0232aab62d166923f9ffe916109e9dd73d16f5c95525
-size 12084802

 version https://git-lfs.github.com/spec/v1
+oid sha256:210c6485f34cc1c35daa38072b29cb540c7dc7c04326cfd5b81c1429368398e3
+size 19045741

onnx/decoder_model_merged_quantized.onnx_data ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:d312201a36ec7d97cac2fc87c664ef0ec16a681f55ec013b9e3ce383904411e4
+size 6739214336

onnx/quantize_config.json CHANGED Viewed

@@ -1,39 +1,39 @@
 {
-    "per_channel": false,
-    "reduce_range": false,
     "use_external_data_format": true,
     "per_model_config": {
         "decoder_model_merged": {
             "op_types": [
-                "Cast",
-                "Sigmoid",
                 "If",
                 "Shape",
-                "Gather",
-                "Squeeze",
-                "Pow",
                 "Sub",
-                "Identity",
-                "Unsqueeze",
-                "MatMul",
-                "ReduceMean",
-                "Constant",
                 "Equal",
-                "Sqrt",
-                "Expand",
                 "Add",
-                "Range",
-                "Less",
                 "Div",
-                "ConstantOfShape",
                 "Where",
-                "Transpose",
                 "Mul",
-                "Slice",
-                "Softmax",
-                "Neg",
-                "Concat",
-                "Reshape"
             ],
             "weight_type": "QInt8"
         }

 {
+    "per_channel": true,
+    "reduce_range": true,
     "use_external_data_format": true,
     "per_model_config": {
         "decoder_model_merged": {
             "op_types": [
+                "Slice",
+                "Sqrt",
+                "Less",
                 "If",
                 "Shape",
+                "Transpose",
+                "Range",
                 "Sub",
+                "Concat",
+                "Squeeze",
+                "Cast",
+                "Neg",
                 "Equal",
                 "Add",
+                "Pow",
+                "Reshape",
+                "Identity",
                 "Div",
+                "Constant",
+                "Gather",
+                "Softmax",
+                "Sigmoid",
+                "ReduceMean",
                 "Where",
                 "Mul",
+                "Expand",
+                "MatMul",
+                "Unsqueeze",
+                "ConstantOfShape"
             ],
             "weight_type": "QInt8"
         }