HuggingFace Repository

This tutorial is available as an IPython notebook at Malaya/example/huggingface-repository.

Starting Malaya 4.7.4, you can load Malaya models from https://huggingface.co/huseinzol05 to get better download speed.

Starting Malaya 4.7.5, by default Malaya will use HuggingFace as backend repository.

[1]:

import os
os.environ['MALAYA_USE_HUGGINGFACE'] = 'false'

[2]:

import malaya

[3]:

malaya.__version__

[3]:

'4.7.5'

Load model from Backblaze

If you found some models not exist in HuggingFace, you can try to use BackBlaze as alternative,

  1. First, set global MALAYA_USE_HUGGINGFACE = false,

os.environ['MALAYA_USE_HUGGINGFACE'] = 'false'
import malaya

  1. Simply pass use_huggingface=False in load model function parameter, for an example,

malaya.emotion.transformer(model = 'bert', use_huggingface = False)

[6]:

malaya.emotion.transformer(model = 'albert', use_huggingface = False)

101%|██████████| 49.0/48.6 [00:07<00:00, 6.37MB/s]
184%|██████████| 1.00/0.54 [00:01<00:00, 1.37s/MB]
135%|██████████| 1.00/0.74 [00:01<00:00, 1.22s/MB]

[6]:

<malaya.model.bert.MulticlassBERT at 0x16647ce90>