Instructions to use cmcmaster/il_7b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cmcmaster/il_7b with Transformers:

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

pipe = pipeline("text-generation", model="cmcmaster/il_7b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("cmcmaster/il_7b")
model = AutoModelForCausalLM.from_pretrained("cmcmaster/il_7b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use cmcmaster/il_7b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cmcmaster/il_7b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cmcmaster/il_7b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/cmcmaster/il_7b

SGLang

How to use cmcmaster/il_7b 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 "cmcmaster/il_7b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cmcmaster/il_7b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "cmcmaster/il_7b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cmcmaster/il_7b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use cmcmaster/il_7b with Docker Model Runner:
```
docker model run hf.co/cmcmaster/il_7b
```

cmcmaster commited on Mar 8, 2024

Commit

f0b89b7

verified ·

1 Parent(s): 69f1e6a

Update README.md

Browse files

Files changed (1) hide show

README.md +1 -24

README.md CHANGED Viewed

@@ -47,27 +47,4 @@ This model was merged using the SLERP merge method.
 The merge was made from two unreleased models:
 - rheumistral-sft was trained from the original mistral checkpoint in two stages: 1) "continued pretraining" on a large, curated dataset of rheumatology and immunology texts; 2) supervised finetuning on a combination of synthetic and human generated QA pairs and chat logs
-- biorheumistral-sft was trained the same way as rheumistral-sft, only it started from the [BioMistral-7B](https://huggingface.co/BioMistral/BioMistral-7B) checkpoint.
-### Configuration
-The following YAML configuration was used to produce this model:
-```yaml
-slices:
-  - sources:
-      - model: rheumistral-sft
-        layer_range: [0, 32]
-      - model: biorheumistral-sft
-        layer_range: [0, 32]
-merge_method: slerp
-base_model: /mnt/hdd/projects/rheum_llm/alignment-handbook/rheumistral-sft-merged-final
-parameters:
-  t:
-    - filter: self_attn
-      value: [0, 0.5, 0.3, 0.7, 1]
-    - filter: mlp
-      value: [1, 0.5, 0.7, 0.3, 0]
-    - value: 0.5
-dtype: bfloat16
-```

 The merge was made from two unreleased models:
 - rheumistral-sft was trained from the original mistral checkpoint in two stages: 1) "continued pretraining" on a large, curated dataset of rheumatology and immunology texts; 2) supervised finetuning on a combination of synthetic and human generated QA pairs and chat logs
+- biorheumistral-sft was trained the same way as rheumistral-sft, only it started from the [BioMistral-7B](https://huggingface.co/BioMistral/BioMistral-7B) checkpoint.