Model reference · open weights
fasttext-fa-300 is an open-weight embedding model from hezarai. fasttext-fa-300 (BF16) weighs 69 MB; the smallest configuration that runs it is RTX 3060 12 GB.
For self-hosting, compare this build's requirements below with the GPU you plan to rent.
Summary of the hezarai/fasttext-fa-300 model card, 2026-10-01
What it is
| Released by | hezarai |
|---|---|
| Released | 2023-08-06 |
| VRAM | 69 MB for the weights |
What it runs on
| Card | Runs | Memory |
|---|---|---|
| RTX 3060 12 GB | fits | 11.6 GB |
| RTX 4060 Ti 16 GB | fits | 15.4 GB |
| RTX 3090 24 GB | fits | 23.4 GB |
| RTX 4090 24 GB | fits | 23.4 GB |
| RTX 5090 32 GB | fits | 31.0 GB |
| L40S 48 GB | fits | 44.0 GB |
| A100 80 GB | fits | 78.2 GB |
| H100 80 GB | fits | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | fits | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
From the model card
This is the original fasttext embedding model for Persian from here loaded and converted using Gensim and exported to Hezar compatible format. For more info, see here.
In order to use this model in Hezar you can simply use this piece of code:
pip install hezar
from hezar.embeddings import Embedding
fasttext = Embedding.load("hezarai/fasttext-fa-300")
# Get embedding vector
vector = fasttext("هزار")
# Find the word that doesn't match with the rest
doesnt_match = fasttext.doesnt_match(["خانه", "اتاق", "ماشین"])
# Find the top-n most similar words to the given word
most_similar = fasttext.most_similar("هزار", top_n=5)
# Find the cosine similarity value between two words
similarity = fasttext.similarity("مهندس", "دکتر")
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.
How it works