docs: format Method pipeline as tree

README.md

@@ -17,20 +17,20 @@ tags:
 
 # BashkirRoBERTa
 
-> A Bashkir masked language model for fill-mask suggestions and further fine-tuning.
+> A Bashkir masked language model for fill-mask suggestions and further fine-tuning.
 
 ## Overview
 
-BashkirRoBERTa predicts masked SentencePiece tokens from context. It can provide
-fill-mask suggestions and serve as an encoder for further fine-tuning, including
-the related person-name model. The release includes Transformers weights and FP32,
-FP16 and INT8 ONNX graphs. Its custom Pre-LayerNorm architecture requires the
-included model code when loaded with Transformers.
+BashkirRoBERTa predicts masked SentencePiece tokens from context. It can provide
+fill-mask suggestions and serve as an encoder for further fine-tuning, including
+the related person-name model. The release includes Transformers weights and FP32,
+FP16 and INT8 ONNX graphs. Its custom Pre-LayerNorm architecture requires the
+included model code when loaded with Transformers.
 
 | At a glance | |
 | --- | --- |
 | Task | Masked language modelling / fill-mask |
-| Default artifact | `model.safetensors` (Transformers); `onnx/model_int8.onnx` (CPU ONNX) |
+| Default artifact | `model.safetensors` (Transformers); `onnx/model_int8.onnx` (CPU ONNX) |
 | Source | A monolingual Bashkir-language dataset |
 | Version / license | v1 / Apache-2.0 |
 
@@ -40,10 +40,10 @@ included model code when loaded with Transformers.
 
 | File | Purpose | Size |
 | --- | --- | ---: |
-| **`model.safetensors`** | Transformers weights and fine-tuning starting point | 200.2 MB |
-| `onnx/model_fp32.onnx` | Full-precision ONNX reference for CPU inference | 242.2 MB |
-| `onnx/model_fp16.onnx` | FP16 ONNX model for GPU / DirectML | 121.2 MB |
-| **`onnx/model_int8.onnx`** | Compact INT8 ONNX graph for CPU use | 60.9 MB |
+| **`model.safetensors`** | Transformers weights and fine-tuning starting point | 200.2 MB |
+| `onnx/model_fp32.onnx` | Full-precision ONNX reference for CPU inference | 242.2 MB |
+| `onnx/model_fp16.onnx` | FP16 ONNX model for GPU / DirectML | 121.2 MB |
+| **`onnx/model_int8.onnx`** | Compact INT8 ONNX graph for CPU use | 60.9 MB |
 | `spm_bashkir_bert_16k.model` | SentencePiece tokenizer | — |
 | `config.json` | Model configuration (`auto_map` for custom code) | — |
 | `configuration_bashkir_roberta.py`, `modeling_bashkir_roberta.py`, `tokenization_bashkir_roberta.py` | Custom Pre-LayerNorm implementation | — |
@@ -82,45 +82,48 @@ Outputs from the INT8 ONNX model on CPU:
 
 ## Method
 
-The model was pretrained with dynamic masked-language modelling on a monolingual
-Bashkir-language dataset assembled from encyclopedic, periodical and literary
-sources. The source texts are not distributed in this repository.
-
-```text
-Bashkir text → SentencePiece tokens → masked-language pretraining
-    → Transformers checkpoint → FP32 / FP16 / INT8 ONNX exports
-```
-
-### Benchmark
+The model was pretrained with dynamic masked-language modelling on a monolingual
+Bashkir-language dataset assembled from encyclopedic, periodical and literary
+sources. The source texts are not distributed in this repository.
+
+```text
+raw Bashkir text (monolingual corpus)
+  ├── 1. SentencePiece unigram tokenization (vocab size: 32,000)
+  ├── 2. dynamic masked-language pretraining (15% subword masking rate)
+  ├── 3. export PyTorch Transformers checkpoint
+  └── 4. export optimized ONNX graphs: FP32, FP16 & dynamic INT8
+```
+
+### Benchmark
 
 On a held-out Bashkir encyclopedic evaluation set the project reports **24.7%
 top-1** and **54.0% top-5** accuracy for masked subword prediction. These are
-diagnostic MLM results, not a general-purpose language-understanding score: a mask
-may represent a whole word or a SentencePiece subword fragment.
-The FP32 ONNX graph was checked against the source PyTorch model on multiple
-sequence lengths and batch sizes, with matching top-token predictions.
+diagnostic MLM results, not a general-purpose language-understanding score: a mask
+may represent a whole word or a SentencePiece subword fragment.
+The FP32 ONNX graph was checked against the source PyTorch model on multiple
+sequence lengths and batch sizes, with matching top-token predictions.
 
 ## Quality and Use
 
-Use the Transformers checkpoint for fill-mask experiments or fine-tuning and the
-INT8 ONNX graph for compact CPU inference. Choose FP32 ONNX when a full-precision
-ONNX reference is needed. Predictions are token suggestions that need review in
-context, especially when the mask represents only part of a word.
+Use the Transformers checkpoint for fill-mask experiments or fine-tuning and the
+INT8 ONNX graph for compact CPU inference. Choose FP32 ONNX when a full-precision
+ONNX reference is needed. Predictions are token suggestions that need review in
+context, especially when the mask represents only part of a word.
 
 ### Limitations
 
 - Diagnostic MLM accuracy only; not fine-tuned for any downstream task.
 - A mask may correspond to a partial subword, not always a full word.
 - Predictions reflect the training corpus and may prefer frequent or encyclopedic phrasing.
-- No training texts are redistributed; provenance or removal requests go through the maintainer.
-
-## Related Resources
-
-- [BashkirRoBERTa NER](https://huggingface.co/failed09/bashkir-roberta-ner)
-  fine-tunes this encoder to find person names and return their text spans. Use it
-  when the task is person-name extraction rather than masked-token prediction.
-
-## Usage
+- No training texts are redistributed; provenance or removal requests go through the maintainer.
+
+## Related Resources
+
+- [BashkirRoBERTa NER](https://huggingface.co/failed09/bashkir-roberta-ner)
+  fine-tunes this encoder to find person names and return their text spans. Use it
+  when the task is person-name extraction rather than masked-token prediction.
+
+## Usage
 
 ```bash
 pip install transformers torch huggingface_hub
@@ -143,7 +146,7 @@ prediction_id = logits[0, mask_index].argmax().item()
 print(tokenizer.decode([prediction_id]))  # яратам
 ```
 
-ONNX Runtime for CPU and edge deployment:
+ONNX Runtime for CPU and edge deployment:
 
 ```python
 import numpy as np
@@ -161,16 +164,16 @@ tokens = [2] + sp.encode("Мин башҡорт телен ") + [6] + sp.encode(
 mask_idx = tokens.index(6)
 logits = session.run(None, {"input_ids": np.array([tokens], dtype=np.int64)})[0][0, mask_idx]
 top_tokens = np.argsort(logits)[::-1][:5]
-print([sp.decode([int(t)]) for t in top_tokens])  # ['яратам', 'беләм', 'өйрәнә', ...]
-```
-
-For full-precision ONNX inference, change the downloaded filename to
-`onnx/model_fp32.onnx`; the input and output names are the same.
+print([sp.decode([int(t)]) for t in top_tokens])  # ['яратам', 'беләм', 'өйрәнә', ...]
+```
+
+For full-precision ONNX inference, change the downloaded filename to
+`onnx/model_fp32.onnx`; the input and output names are the same.
 
 ## License
 
 The model weights, tokenizer and release code are distributed under the
-[Apache-2.0 license](https://huggingface.co/failed09/bashkir-roberta/blob/main/LICENSE). Training texts are not redistributed; their
+[Apache-2.0 license](https://huggingface.co/failed09/bashkir-roberta/blob/main/LICENSE). Training texts are not redistributed; their
 rights remain with their respective owners. For provenance or removal requests,
 contact the maintainer through the Hub.