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

Libraries

How to use ChrisToukmaji/llama_luganda_LAFT with Transformers:

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

pipe = pipeline("text-generation", model="ChrisToukmaji/llama_luganda_LAFT", device_map="auto")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ChrisToukmaji/llama_luganda_LAFT")
model = AutoModelForCausalLM.from_pretrained("ChrisToukmaji/llama_luganda_LAFT", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ChrisToukmaji/llama_luganda_LAFT with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/ChrisToukmaji/llama_luganda_LAFT

SGLang

How to use ChrisToukmaji/llama_luganda_LAFT 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 "ChrisToukmaji/llama_luganda_LAFT" \
    --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": "ChrisToukmaji/llama_luganda_LAFT",
		"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 "ChrisToukmaji/llama_luganda_LAFT" \
        --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": "ChrisToukmaji/llama_luganda_LAFT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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

Paper and Citation

Paper: Few-Shot Cross-Lingual Transfer for Prompting Large Language Models in Low-Resource Languages

@misc{toukmaji2024fewshot,
      title={Few-Shot Cross-Lingual Transfer for Prompting Large Language Models in Low-Resource Languages}, 
      author={Christopher Toukmaji},
      year={2024},
      eprint={2403.06018},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

luganda_finetuned_model

This model is a fine-tuned version of HF_llama on the common_voice lg dataset. It achieves the following results on the evaluation set:

Loss: 3.4695
Accuracy: 0.2928

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 1
eval_batch_size: 1
seed: 42
distributed_type: multi-GPU
num_devices: 4
total_train_batch_size: 4
total_eval_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.03
num_epochs: 3.0

Training results

Framework versions

Transformers 4.29.2
Pytorch 2.0.1+cu117
Datasets 2.12.0
Tokenizers 0.13.3

Downloads last month: 6

Dataset used to train ChrisToukmaji/llama_luganda_LAFT

Paper for ChrisToukmaji/llama_luganda_LAFT

Few-Shot Cross-Lingual Transfer for Prompting Large Language Models in Low-Resource Languages

Paper • 2403.06018 • Published Mar 9, 2024

Evaluation results

Accuracy on common_voice lg
validation set self-reported

0.293