Instructions to use xap/Meta-Llama3-8B-Instruct-assessment with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use xap/Meta-Llama3-8B-Instruct-assessment with Transformers:

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

pipe = pipeline("text-generation", model="xap/Meta-Llama3-8B-Instruct-assessment")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("xap/Meta-Llama3-8B-Instruct-assessment")
model = AutoModelForCausalLM.from_pretrained("xap/Meta-Llama3-8B-Instruct-assessment")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use xap/Meta-Llama3-8B-Instruct-assessment with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "xap/Meta-Llama3-8B-Instruct-assessment"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "xap/Meta-Llama3-8B-Instruct-assessment",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/xap/Meta-Llama3-8B-Instruct-assessment

SGLang

How to use xap/Meta-Llama3-8B-Instruct-assessment 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 "xap/Meta-Llama3-8B-Instruct-assessment" \
    --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": "xap/Meta-Llama3-8B-Instruct-assessment",
		"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 "xap/Meta-Llama3-8B-Instruct-assessment" \
        --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": "xap/Meta-Llama3-8B-Instruct-assessment",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use xap/Meta-Llama3-8B-Instruct-assessment with Docker Model Runner:
```
docker model run hf.co/xap/Meta-Llama3-8B-Instruct-assessment
```

Model Card for Meta-Llama3-8B-Instruct-assessment

Model Details

Model Description

This is the model card of a Meta-Llama3-8B-Instruct-assessment model that has been developed by fine-tuning Meta-Llama3-8B-Instruct. The model is finetuned using LoRA. This model calculates the semantic similarity between student explanation and expert/standard explanation for a given line of code during source code comprehension.

Developed by: xap
License: llama3
Finetuned from model : meta-llama/Meta-Llama-3-8B-Instruct
Finetuned using dataset : SelfCode2.0

How to Get Started with the Model

The dataset or input for this model should be in the alpaca format. If the model is loaded without using PeftModel,it will only load the base pre-trained weights without the LoRA fine-tuning. To use the model for inference use the following code:

from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("xap/Meta-Llama3-8B-Instruct-assessment")
base_model = AutoModelForCausalLM.from_pretrained("xap/Meta-Llama3-8B-Instruct-assessment")

lora_config = LoraConfig.from_pretrained("xap/Meta-Llama3-8B-Instruct-assessment")
model = PeftModel.from_pretrained(
    base_model,
    "xap/Meta-Llama3-8B-Instruct-assessment",
    lora_config=lora_config,
)

inputt = "Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.  ### Instruction: For the given line of code, both the student and expert have provided the explanation for that line of code. Compute the semantic similarity between the student explanation and the expert explanation for the line of code.. ###  Input: for given line of code int[] values = {5, 8, 4, 78, 95, 12, 1, 0, 6, 35, 46};, the expert explanation is We declare an array of values to hold the numbers. and the student explanation is This line creates the integer array with the values. you need this to achieve the goal bc you need an array to look in ### Response: "
inputs = tokenizer(inputt,return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=5)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Citation

@misc {jeevan_2024,
    author       = { {jeevan} },
    title        = { Meta-Llama3-8B-Instruct-assessment (Revision 726a029) },
    year         = 2024,
    url          = { https://huggingface.co/xap/Meta-Llama3-8B-Instruct-assessment },
    doi          = { 10.57967/hf/2244 },
    publisher    = { Hugging Face }
}

Downloads last month: 1

Safetensors

Model size

8B params

Tensor type

F32