Instructions to use goldfish-models/san_deva_full with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use goldfish-models/san_deva_full with Transformers:

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

pipe = pipeline("text-generation", model="goldfish-models/san_deva_full")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("goldfish-models/san_deva_full")
model = AutoModelForCausalLM.from_pretrained("goldfish-models/san_deva_full", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use goldfish-models/san_deva_full with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/goldfish-models/san_deva_full

SGLang

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

Docker Model Runner
How to use goldfish-models/san_deva_full with Docker Model Runner:
```
docker model run hf.co/goldfish-models/san_deva_full
```

san_deva_full / README.md

goldfish-models

Upload README.md with huggingface_hub

769fdb9 verified almost 2 years ago

preview code

Raw

History Blame Contribute Delete

3.32 kB


	---
	license: apache-2.0
	language:
	- san
	datasets:
	- allenai/MADLAD-400
	- allenai/nllb
	- oscar-corpus/OSCAR-2109
	- cis-lmu/Glot500
	library_name: transformers
	pipeline_tag: text-generation
	tags:
	- goldfish
	- arxiv:2408.10441
	---

	# san_deva_full

	Goldfish is a suite of monolingual language models trained for 350 languages.
	This model is the <b>Sanskrit</b> (Devanagari script) model trained on 119MB of data (all our data in the language), after accounting for an estimated byte premium of 2.54; content-matched text in Sanskrit takes on average 2.54x as many UTF-8 bytes to encode as English.
	The Goldfish models are trained primarily for comparability across languages and for low-resource languages; Goldfish performance for high-resource languages is not designed to be comparable with modern large language models (LLMs).

	Note: This language is available in Goldfish with other scripts (writing systems). See: san_latn.

	Note: san_deva is a [macrolanguage](https://iso639-3.sil.org/code_tables/639/data) code. None of its contained individual languages are included in Goldfish (for script deva).

	All training and hyperparameter details are in our paper, [Goldfish: Monolingual Language Models for 350 Languages (Chang et al., 2024)](https://www.arxiv.org/abs/2408.10441).

	Training code and sample usage: https://github.com/tylerachang/goldfish

	Sample usage also in this Google Colab: [link](https://colab.research.google.com/drive/1rHFpnQsyXJ32ONwCosWZ7frjOYjbGCXG?usp=sharing)

	## Model details:

	To access all Goldfish model details programmatically, see https://github.com/tylerachang/goldfish/blob/main/model_details.json.
	All models are trained with a [CLS] (same as [BOS]) token prepended, and a [SEP] (same as [EOS]) token separating sequences.
	For best results, make sure that [CLS] is prepended to your input sequence (see sample usage linked above)!
	Details for this model specifically:

	* Architecture: gpt2
	* Parameters: 124770816
	* Maximum sequence length: 512 tokens
	* Training text data (raw): 303.45MB
	* Training text data (byte premium scaled): 119.345MB
	* Training tokens: 31856128 (x10 epochs)
	* Vocabulary size: 50000
	* Compute cost: 1.62598373425152e+17 FLOPs or ~15.4 NVIDIA A6000 GPU hours

	Training datasets (percentages prior to deduplication):
	* 39.60135%: [MADLAD-400 (CommonCrawl)](https://huggingface.co/datasets/allenai/MADLAD-400)
	* 37.07458%: [NLLB (CommonCrawl and ParaCrawl)](https://huggingface.co/datasets/allenai/nllb)
	* 9.73656%: [Wikipedia 2023/08](https://dumps.wikimedia.org/)
	* 7.30509%: [OSCAR 2021/09](https://huggingface.co/datasets/oscar-corpus/OSCAR-2109)
	* 5.49767%: [Glot500](https://huggingface.co/datasets/cis-lmu/Glot500), including [CCNet](https://github.com/facebookresearch/cc_net), [Hindialect](https://lindat.mff.cuni.cz/repository/xmlui/handle/11234/1-4839), [Wortschatz Leipzig Data](https://wortschatz.uni-leipzig.de/en/download), [OSCAR](https://oscar-project.org/)
	* 0.78475%: [eBible](https://ebible.org/find/)


	## Citation

	If you use this model, please cite:

	```
	@article{chang-etal-2024-goldfish,
	title={Goldfish: Monolingual Language Models for 350 Languages},
	author={Chang, Tyler A. and Arnett, Catherine and Tu, Zhuowen and Bergen, Benjamin K.},
	journal={Preprint},
	year={2024},
	url={https://www.arxiv.org/abs/2408.10441},
	}
	```