docs: format Method pipeline as tree

README.md

@@ -72,8 +72,11 @@ form was never seen during training.
 ## Method
 
 ```text
-monolingual Bashkir text → Unicode NFC tokenization → skip-gram 300d (char n-grams 3–6)
-    → prune (top-50k words, top-100k subwords) → PQ FP16 → mini .bin
+monolingual Bashkir text (clean corpus)
+  ├── 1. Unicode NFC tokenization preserving 9 Bashkir letters
+  ├── 2. train skip-gram 300d with subword character n-grams (3–6)
+  ├── 3. prune vocabulary to top-50k words and top-100k subwords
+  └── 4. product quantization (PQ FP16) export to mini .bin
 ```
 
 The model was trained from a monolingual Bashkir-language dataset.
@@ -135,7 +138,7 @@ print(model.similarity("атай", "әсәй"))
 ## License
 
 The model and release code are provided under the
-[Apache-2.0 license](https://huggingface.co/failed09/bashkir-fasttext/blob/main/LICENSE). The model
+[Apache-2.0 license](https://huggingface.co/failed09/bashkir-fasttext/blob/main/LICENSE). The model
 is derived from upstream and project corpora; this release does not claim
 ownership of source texts or other training materials. Upstream rights and
 licensing conditions remain applicable.