Update model card for Georgian

README.md

@@ -1,37 +1,62 @@
----
-language:
-- kat
-license: cc-by-nc-4.0
-tags: [trimmed, mbart, seq2seq, monolingual, translation]
-base_model: facebook/mbart-large-50
-datasets:
-  - Lumberjackk/fineweb-2-trimming
----
-
-# mbart-large-50-kat-16384
-
-Version de [facebook/mbart-large-50](https://huggingface.co/facebook/mbart-large-50) avec vocabulaire réduit
-pour **Georgian** (total 16,384 tokens).
-
-| | Original | Trimmed |
-|---|---|---|
-| Vocabulaire | 250,027 | 16,384 |
-| Paramètres | 610,879,488 | 371,601,408 |
-
-## Usage
-```python
-from transformers import MBartTokenizerFast, MBartForConditionalGeneration
-
-tokenizer = MBartTokenizerFast.from_pretrained("lbourdois/mbart-large-50-kat-16384")
-model     = MBartForConditionalGeneration.from_pretrained("lbourdois/mbart-large-50-kat-16384")
-
-lang_id = tokenizer.lang_code_to_id["ka_GE"]
-model.config.decoder_start_token_id            = lang_id
-model.generation_config.decoder_start_token_id = lang_id
-model.generation_config.forced_bos_token_id    = lang_id
-
-tokenizer.src_lang = "ka_GE"
-inputs = tokenizer("Hello world.", return_tensors="pt")
-out    = model.generate(inputs["input_ids"], forced_bos_token_id=lang_id, max_new_tokens=50)
-print(tokenizer.decode(out[0], skip_special_tokens=True))
-```
+---
+language: kat
+license: mit
+tags:
+  - trimmed
+library_name: transformers
+base_model: facebook/mbart-large-50
+base_model_relation: quantized
+datasets:
+  - lbourdois/fineweb-2-trimming
+---
+
+# mbart-large-50-kat-16384
+This model is a **39.17% smaller** version of [facebook/mbart-large-50](https://huggingface.co/facebook/mbart-large-50) optimized for Georgian language via vocabulary size reduction using the [trimming](https://huggingface.co/blog/lbourdois/introduction-to-trimming) method.  
+This trimmed model should perform similarly to the original model with only 16,384 tokens and a much smaller memory footprint. However, it may not perform well for other languages as tokens not commonly used in the selected languages were removed from the vocabulary.
+
+## Model Statistics
+| Metric | Original | Trimmed | Reduction |
+|--------|----------|---------|-----------|
+| **Vocabulary size** | 250,027 tokens | 16,384 tokens | **93.45%** |
+| **Model size** | 610,879,488 params | 371,601,408 params | **39.17%** |
+
+![image](https://raw.githubusercontent.com/lbourdois/blog/refs/heads/master/assets/images/Trimming/mbart-large-50-16384.png)
+
+## Mining Dataset Statistics
+- **Number of texts used for mining**: 200,000 texts  
+- **Dataset**: [lbourdois/fineweb-2-trimming](https://huggingface.co/datasets/lbourdois/fineweb-2-trimming)
+
+## Usage
+
+```python
+from transformers import AutoModel, AutoTokenizer
+
+model_name = "alphaedge-ai/mbart-large-50-kat-16384"
+model = AutoModel.from_pretrained(model_name)
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+```
+
+## Citations
+
+#### mBART
+```
+@misc{tang2020multilingualtranslationextensiblemultilingual,
+      title={Multilingual Translation with Extensible Multilingual Pretraining and Finetuning}, 
+      author={Yuqing Tang and Chau Tran and Xian Li and Peng-Jen Chen and Naman Goyal and Vishrav Chaudhary and Jiatao Gu and Angela Fan},
+      year={2020},
+      eprint={2008.00401},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL},
+      url={https://arxiv.org/abs/2008.00401}, 
+}
+```
+
+#### Trimming blog post
+```
+@misc{hf_blogpost_trimming,
+      title={Introduction to Trimming}, 
+      author={Loïck BOURDOIS and Tom AARSEN and Bram VANROY and Christopher AKIKI and Woojun JUNG and Manuel ROMERO and Prithiv SAKTHI},
+      year={2026},
+      url={https://huggingface.co/blog/lbourdois/introduction-to-trimming}, 
+}
+```