I want to quan the m2m-100 model to do the translation work, but i don’t know how to do
1 Like
For example, like this.
# "pip install bitsandbytes" in advance
from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer
from transformers import BitsAndBytesConfig # <= added
import torch # <= added
nf4_config = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_quant_type="nf4", bnb_4bit_use_double_quant=True, bnb_4bit_compute_dtype=torch.bfloat16) # <= added
hi_text = "जीवन एक चॉकलेट बॉक्स की तरह है।"
chinese_text = "生活就像一盒巧克力。"
model = M2M100ForConditionalGeneration.from_pretrained("facebook/m2m100_418M", quantization_config=nf4_config) # <= modified
tokenizer = M2M100Tokenizer.from_pretrained("facebook/m2m100_418M")
# translate Hindi to French
tokenizer.src_lang = "hi"
encoded_hi = tokenizer(hi_text, return_tensors="pt")
generated_tokens = model.generate(**encoded_hi, forced_bos_token_id=tokenizer.get_lang_id("fr"))
tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
# => "La vie est comme une boîte de chocolat."
# translate Chinese to English
tokenizer.src_lang = "zh"
encoded_zh = tokenizer(chinese_text, return_tensors="pt")
generated_tokens = model.generate(**encoded_zh, forced_bos_token_id=tokenizer.get_lang_id("en"))
tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
# => "Life is like a box of chocolate."
1 Like
This topic was automatically closed 12 hours after the last reply. New replies are no longer allowed.