Instructions to use Mehakm/llama-3.2-3b-resume-parser with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Mehakm/llama-3.2-3b-resume-parser with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-3.2-3B-Instruct")
model = PeftModel.from_pretrained(base_model, "Mehakm/llama-3.2-3b-resume-parser")

Transformers

How to use Mehakm/llama-3.2-3b-resume-parser with Transformers:

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

pipe = pipeline("text-generation", model="Mehakm/llama-3.2-3b-resume-parser")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Mehakm/llama-3.2-3b-resume-parser", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Mehakm/llama-3.2-3b-resume-parser with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Mehakm/llama-3.2-3b-resume-parser"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Mehakm/llama-3.2-3b-resume-parser",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Mehakm/llama-3.2-3b-resume-parser

SGLang

How to use Mehakm/llama-3.2-3b-resume-parser 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 "Mehakm/llama-3.2-3b-resume-parser" \
    --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": "Mehakm/llama-3.2-3b-resume-parser",
		"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 "Mehakm/llama-3.2-3b-resume-parser" \
        --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": "Mehakm/llama-3.2-3b-resume-parser",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Mehakm/llama-3.2-3b-resume-parser with Docker Model Runner:
```
docker model run hf.co/Mehakm/llama-3.2-3b-resume-parser
```

Model Card for Model ID

Model Card for Mehakm/llama-3.2-3b-resume-parser

This model is a fine-tuned QLoRA adapter of meta-llama/Llama-3.2-3B-Instruct designed to extract deterministic, structured 15-field JSON data from unstructured plain-text and PDF resumes.

Model Details

Model Description

Developed by: Mehak (Mehakm)
Model type: Causal Language Model (QLoRA PEFT Adapter)
Language(s) (NLP): English (en)
License: Apache-2.0 / Llama 3.2 Community License
Finetuned from model: meta-llama/Llama-3.2-3B-Instruct

Model Sources

Repository: https://huggingface.co/Mehakm/llama-3.2-3b-resume-parser
Demo: https://huggingface.co/spaces/Mehakm/resume-json-parser

Uses

Direct Use

This model is intended to be used for automated resume parsing, candidate profiling, and Job Description (JD) skill-gap analysis. Given raw resume text formatted with ChatML tags, it outputs a structured JSON object containing 15 standardized recruitment attributes (current_title, previous_titles, current_company, previous_companies, years_experience, seniority, primary_domain, industries, core_skills, secondary_skills, tools, leadership_experience, key_achievements, location, summary).

Out-of-Scope Use

This model is not intended to make automated hiring or rejection decisions without human review, nor is it designed for general open-ended conversational chat outside of structured resume information extraction.

Bias, Risks, and Limitations

Formatting Sensitivity: Very long resumes exceeding the 1024-token context window may have trailing sections truncated if not chunked or paired with bracket auto-repair.
Recommendations: Downstream applications should pair generation with use_cache=True and a lightweight JSON bracket-repair utility to handle edge-case truncation.

How to Get Started with the Model

Use the code below to get started with the model:

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base_model_id = "meta-llama/Llama-3.2-3B-Instruct"
adapter_id = "Mehakm/llama-3.2-3b-resume-parser"

tokenizer = AutoTokenizer.from_pretrained(base_model_id)
base_model = AutoModelForCausalLM.from_pretrained(
    base_model_id,
    torch_dtype=torch.float16,
    device_map="auto"
)
model = PeftModel.from_pretrained(base_model, adapter_id)

Downloads last month: 34

Model tree for Mehakm/llama-3.2-3b-resume-parser

Base model

meta-llama/Llama-3.2-3B-Instruct

Adapter

(877)

this model

Mehakm
/

llama-3.2-3b-resume-parser