Instructions to use sambanovasystems/codegen-16B-mono-toolbench with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sambanovasystems/codegen-16B-mono-toolbench with Transformers:

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

pipe = pipeline("text-generation", model="sambanovasystems/codegen-16B-mono-toolbench")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("sambanovasystems/codegen-16B-mono-toolbench")
model = AutoModelForCausalLM.from_pretrained("sambanovasystems/codegen-16B-mono-toolbench")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use sambanovasystems/codegen-16B-mono-toolbench with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "sambanovasystems/codegen-16B-mono-toolbench"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sambanovasystems/codegen-16B-mono-toolbench",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/sambanovasystems/codegen-16B-mono-toolbench

SGLang

How to use sambanovasystems/codegen-16B-mono-toolbench 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 "sambanovasystems/codegen-16B-mono-toolbench" \
    --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": "sambanovasystems/codegen-16B-mono-toolbench",
		"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 "sambanovasystems/codegen-16B-mono-toolbench" \
        --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": "sambanovasystems/codegen-16B-mono-toolbench",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use sambanovasystems/codegen-16B-mono-toolbench with Docker Model Runner:
```
docker model run hf.co/sambanovasystems/codegen-16B-mono-toolbench
```

bol20162021 commited on May 24, 2023

Commit

4137166

1 Parent(s): 1da9502

update

Browse files

Files changed (1) hide show

README.md +7 -19

README.md CHANGED Viewed

@@ -1,11 +1,11 @@
 ---
 license: bsd-3-clause
 ---
-# codgen-16B-action
 <!-- Provide a quick summary of what the model is/does. -->
-codgen-16B-action is a 16 billion parameter model used for api based action generation. It is instruction tuned from [codegen-16B-mono](https://huggingface.co/Salesforce/codegen-16B-mono) on api based action generation datasets.
 ## Model Details
@@ -45,7 +45,7 @@ This model is intended for commercial and research use.
 <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
-codgen-16B-action should NOT be used for purpose other than API based action generation.
 ### Recommendations
@@ -67,8 +67,8 @@ Users should be made aware of the risks, biases, limitations, and restrictions o
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
-tokenizer = AutoTokenizer.from_pretrained("sambanovasystems/codegen-16b-action")
-model = AutoModelForCausalLM.from_pretrained("sambanovasystems/codegen-16b-action", device_map="auto", torch_dtype="auto")
 ```
 ### Suggested Inference Parameters
@@ -108,7 +108,7 @@ Input text: 十七岁的风是什么颜色的?
 <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
-We trained codegen-16b-action on 4 80GB A100 gpu's. We started from [codegen-16B-mono](https://huggingface.co/Salesforce/codegen-16B-mono). We finetuned it on XXX dataset.
 All of the code used to prepare the datasets and the scripts to run training and inference are open-sourced and freely available at [githublink here](dummy link)
@@ -129,19 +129,7 @@ All of the code used to prepare the datasets and the scripts to run training and
 - Learning Rate Scheduler: Fixed LR
 - Weight decay: 0.1
-**Instruction-tuned Training on Dolly 2.0 and Oasst1**
-- Hardware: SambaNova Reconfigurable Dataflow Unit (RDU)
-- Optimizer: AdamW
-- Grad accumulation: 1
-- Epochs: 3
-- Global Batch size: 128
-- Batch tokens: 128 * 2048 = 262,144 tokens
-- Learning Rate: 1e-5
-- Learning Rate Scheduler: Cosine Schedule with Warmup
-- Warmup Steps: 0
-- End Learning Ratio: 0.1
-- Weight decay: 0.1
 </details>
@@ -150,7 +138,7 @@ All of the code used to prepare the datasets and the scripts to run training and
 ## Acknowledgment
-## Cite codegen-16b-action
 ```
 @software{bloomchat,
   title = {{BLOOMChat: a New Open Multilingual Chat LLM}},

 ---
 license: bsd-3-clause
 ---
+# codgen-16B-mono-toolbench
 <!-- Provide a quick summary of what the model is/does. -->
+codgen-16B-mono-toolbench is a 16 billion parameter model used for api based action generation. It is instruction tuned from [codegen-16B-mono](https://huggingface.co/Salesforce/codegen-16B-mono) on api based action generation datasets.
 ## Model Details
 <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+codgen-16B-mono-toolbench should NOT be used for purpose other than API based action generation.
 ### Recommendations
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
+tokenizer = AutoTokenizer.from_pretrained("sambanovasystems/codegen-16b-mono-toolbench")
+model = AutoModelForCausalLM.from_pretrained("sambanovasystems/codegen-16b-mono-toolbench", device_map="auto", torch_dtype="auto")
 ```
 ### Suggested Inference Parameters
 <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+We trained codegen-16b-mono-toolbench on 4 80GB A100 gpu's. We started from [codegen-16B-mono](https://huggingface.co/Salesforce/codegen-16B-mono). We finetuned it on XXX dataset.
 All of the code used to prepare the datasets and the scripts to run training and inference are open-sourced and freely available at [githublink here](dummy link)
 - Learning Rate Scheduler: Fixed LR
 - Weight decay: 0.1
 </details>
 ## Acknowledgment
+## Cite codegen-16b-mono-toolbench
 ```
 @software{bloomchat,
   title = {{BLOOMChat: a New Open Multilingual Chat LLM}},