Instructions to use Echelon-AI/marathi-llama3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Echelon-AI/marathi-llama3 with Transformers:

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

pipe = pipeline("text-generation", model="Echelon-AI/marathi-llama3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Echelon-AI/marathi-llama3")
model = AutoModelForCausalLM.from_pretrained("Echelon-AI/marathi-llama3")
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 Echelon-AI/marathi-llama3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Echelon-AI/marathi-llama3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Echelon-AI/marathi-llama3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Echelon-AI/marathi-llama3

SGLang

How to use Echelon-AI/marathi-llama3 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 "Echelon-AI/marathi-llama3" \
    --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": "Echelon-AI/marathi-llama3",
		"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 "Echelon-AI/marathi-llama3" \
        --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": "Echelon-AI/marathi-llama3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Echelon-AI/marathi-llama3 with Docker Model Runner:
```
docker model run hf.co/Echelon-AI/marathi-llama3
```

Meta llama3 8B trained on marathi alpaca cleaned for 1.5 epochs, On A100 40GB

Model Overview

Marathi-Llama3 is a fine-tuned version of the Llama3 model, tailored specifically for the Marathi language. This model leverages the power of the Llama3 architecture to provide accurate and nuanced responses in Marathi, opening up advanced AI capabilities to Marathi-speaking communities.

from transformers import AutoModelForCausalLM, AutoTokenizer

# Load the model and tokenizer
model_name = "Echelon-AI/marathi-llama3"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

# Generate text
input_text = "कृपया मला मराठी भाषेत एक गोष्ट सांगा."
inputs = tokenizer(input_text, return_tensors="pt")
outputs = model.generate(**inputs, max_length=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

GGUF available at GGUF

Downloads last month: 2

Safetensors

Model size

8B params

Tensor type

BF16

Echelon-AI
/

marathi-llama3

Model Overview

Dataset used to train Echelon-AI/marathi-llama3