Chess Challenge submission by azizbacha

README.md

@@ -0,0 +1,26 @@
+---
+library_name: transformers
+tags:
+- chess
+- llm-course
+- chess-challenge
+license: mit
+---
+
+# chess_learning-v1
+
+Chess model submitted to the LLM Course Chess Challenge.
+
+## Submission Info
+
+- **Submitted by**: [azizbacha](https://huggingface.co/azizbacha)
+- **Parameters**: 952,896
+- **Organization**: LLM-course
+
+## Model Details
+
+- **Architecture**: Chess Transformer (GPT-style)
+- **Vocab size**: 84
+- **Embedding dim**: 128
+- **Layers**: 5
+- **Heads**: 8

config.json

@@ -0,0 +1,20 @@
+{
+  "architectures": [
+    "ChessForCausalLM"
+  ],
+  "bos_token_id": 1,
+  "dropout": 0.1,
+  "dtype": "float32",
+  "eos_token_id": 2,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "chess_transformer",
+  "n_ctx": 256,
+  "n_embd": 128,
+  "n_head": 8,
+  "n_inner": 448,
+  "n_layer": 5,
+  "pad_token_id": 0,
+  "tie_weights": true,
+  "transformers_version": "4.57.6",
+  "vocab_size": 84
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec7a2c6810ce0bd13b988d09d90d96ef845f1292f87d880aac538ddb48977831
+size 3817008

special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "[BOS]",
+  "eos_token": "[EOS]",
+  "pad_token": "[PAD]",
+  "unk_token": "[UNK]"
+}

tokenizer.py

@@ -0,0 +1,358 @@
+"""
+Custom Chess Tokenizer for the Chess Challenge.
+
+This tokenizer treats each move as a single token using the extended UCI notation
+from the Lichess dataset (e.g., WPe2e4, BNg8f6).
+
+The dataset format uses:
+- W/B prefix for White/Black
+- Piece letter: P=Pawn, N=Knight, B=Bishop, R=Rook, Q=Queen, K=King
+- Source and destination squares (e.g., e2e4)
+- Special suffixes: (x)=capture, (+)=check, (+*)=checkmate, (o)/(O)=castling
+"""
+
+from __future__ import annotations
+
+import json
+import os
+import re
+from pathlib import Path
+from typing import Dict, List, Optional, Tuple
+
+from transformers import PreTrainedTokenizer
+
+
+
+# Parse "WPe2e4(x+*)" etc.
+_MOVE_RE = re.compile(
+    r"^(?P<side>[WB])"
+    r"(?P<piece>[PNBRQK])"
+    r"(?P<src>[a-h][1-8])"
+    r"(?P<dst>[a-h][1-8])"
+    r"(?P<suffix>.*)$"
+)
+
+
+
+# Promotions like "=Q" or "=q"
+_PROMO_RE = re.compile(r"=([QRBNqrbn])")
+
+
+def _parse_suffix(suffix: str) -> Tuple[bool, bool, bool, Optional[str], Optional[str]]:
+    """
+    Returns:
+      is_capture, is_check, is_mate, castle_kind, promo_piece
+
+    castle_kind: "k" (kingside) or "q" (queenside) or None
+    promo_piece: one of "q","r","b","n" or None
+    """
+    if not suffix:
+        return False, False, False, None, None
+
+    # Normalize 
+    suf = suffix.strip()
+
+    is_capture = "x" in suf
+    is_check = "+" in suf
+
+    # Mate indicator
+    # We'll treat any "*" as mate.
+    is_mate = "*" in suf
+
+    # Castling: dataset uses (o)/(O) in the move string for king moves
+    castle_kind = None
+    if "(O)" in suf:
+        castle_kind = "q"
+    elif "(o)" in suf:
+        castle_kind = "k"
+
+    promo_piece = None
+    m = _PROMO_RE.search(suf)
+    if m:
+        promo_piece = m.group(1).lower()
+
+    return is_capture, is_check, is_mate, castle_kind, promo_piece
+
+
+def _reindex_vocab(vocab: Dict[str, int]) -> Dict[str, int]:
+    # sort by old id for stability
+    items = sorted(vocab.items(), key=lambda kv: kv[1])
+    return {tok: new_id for new_id, (tok, _) in enumerate(items)}
+
+
+
+class ChessTokenizer(PreTrainedTokenizer):
+    """
+    A custom tokenizer for chess moves using extended UCI notation.
+    
+    This tokenizer maps each possible chess move to a unique token ID.
+    The vocabulary is built from the training dataset to ensure all moves
+    encountered during training have a corresponding token.
+    
+    Example:
+        >>> tokenizer = ChessTokenizer()
+        >>> tokenizer.encode("WPe2e4 BPe7e5")
+        [1, 42, 87, 2]  # [BOS, e2e4, e7e5, EOS]
+    """
+    
+    model_input_names = ["input_ids", "attention_mask"]
+    vocab_files_names = {"vocab_file": "vocab.json"}
+    
+    # Special tokens
+    PAD_TOKEN = "[PAD]"
+    BOS_TOKEN = "[BOS]"
+    EOS_TOKEN = "[EOS]"
+    UNK_TOKEN = "[UNK]"
+
+
+     # Component tokens
+    SIDE_TOKENS = ("[W]", "[B]")
+    PIECE_TOKENS = ("[P]", "[N]", "[B]", "[R]", "[Q]", "[K]")
+    # flags
+    FLAG_TOKENS = (
+        "[x]",       # capture
+        "[+]",       # check
+        "[#]",       # mate
+        "[O-O]",     # kingside castle marker (not required by evaluator)
+        "[O-O-O]",   # queenside castle marker
+        # promotions
+        "[=q]", "[=r]", "[=b]", "[=n]",
+    )   
+    def __init__(
+        self,
+        vocab_file: Optional[str] = None,
+        vocab: Optional[Dict[str, int]] = None,
+        **kwargs,
+    ):
+        """
+        Initialize the chess tokenizer.
+        
+        Args:
+            vocab_file: Path to a JSON file containing the vocabulary mapping.
+            vocab: Dictionary mapping tokens to IDs (alternative to vocab_file).
+            **kwargs: Additional arguments passed to PreTrainedTokenizer.
+        """
+        # Initialize special tokens
+        self._pad_token = self.PAD_TOKEN
+        self._bos_token = self.BOS_TOKEN
+        self._eos_token = self.EOS_TOKEN
+        self._unk_token = self.UNK_TOKEN
+
+        # Remove any duplicate special-token entries passed through kwargs
+        # to avoid "multiple values for keyword" errors when loading from disk.
+        kwargs.pop("pad_token", None)
+        kwargs.pop("bos_token", None)
+        kwargs.pop("eos_token", None)
+        kwargs.pop("unk_token", None)
+        
+        # Load or create vocabulary
+        if vocab is not None:
+            self._vocab = vocab
+        elif vocab_file is not None and os.path.exists(vocab_file):
+            with open(vocab_file, "r", encoding="utf-8") as f:
+                self._vocab = json.load(f)
+        else:
+            # Create a minimal vocabulary with just special tokens
+            # The full vocabulary should be built from the dataset
+            self._vocab = self._create_default_vocab()
+        
+        self._vocab = _reindex_vocab(self._vocab)
+        
+        # Create reverse mapping
+        self._ids_to_tokens = {v: k for k, v in self._vocab.items()}
+        
+        # Call parent init AFTER setting up vocab
+        super().__init__(
+            pad_token=self._pad_token,
+            bos_token=self._bos_token,
+            eos_token=self._eos_token,
+            unk_token=self._unk_token,
+            **kwargs,
+        )
+    
+    def _create_default_vocab(self) -> Dict[str, int]:
+        """
+        Create a minimal default vocabulary with just special tokens.
+        
+        For the full vocabulary, use `build_vocab_from_dataset()`.
+        This minimal vocab is just a placeholder - you should build from data.
+        """
+        tokens: List[str] = [self.PAD_TOKEN, self.BOS_TOKEN, self.EOS_TOKEN, self.UNK_TOKEN]
+        tokens += list(self.SIDE_TOKENS)
+        tokens += list(self.PIECE_TOKENS)
+
+        # Squares (64)
+        for file in "abcdefgh":
+            for rank in "12345678":
+                tokens.append(f"[{file}{rank}]")
+
+        tokens += list(self.FLAG_TOKENS)
+
+        return {tok: idx for idx, tok in enumerate(tokens)}
+    
+    @classmethod
+    def build_vocab_from_iterator(
+        cls,
+        iterator,
+        min_frequency: int = 1,
+    ) -> "ChessTokenizer":
+        return cls()
+    
+    @classmethod
+    def build_vocab_from_dataset(
+        cls,
+        dataset_name: str = "dlouapre/lichess_2025-01_1M",
+        split: str = "train",
+        column: str = "text",
+        min_frequency: int = 500,
+        max_samples: Optional[int] = 100000,
+    ) -> "ChessTokenizer":
+        return cls()
+    
+    @property
+    def vocab_size(self) -> int:
+        """Return the size of the vocabulary."""
+        return len(self._vocab)
+    
+    def get_vocab(self) -> Dict[str, int]:
+        """Return the vocabulary as a dictionary."""
+        return dict(self._vocab)
+    
+    def _tokenize(self, text: str) -> List[str]:
+        """
+        Tokenize a string of moves into a list of tokens.
+        
+        Args:
+            text: A string of space-separated moves.
+        
+        Returns:
+            List of move tokens.
+        """
+        text = (text or "").strip()
+        if not text:
+            return []
+
+        chunks = text.split()
+        out: List[str] = []
+
+        for chunk in chunks:
+            # If chunk is pure uci like "e2e4" or "e7e8q"
+            if re.fullmatch(r"[a-h][1-8][a-h][1-8][qrbn]?", chunk):
+                src = chunk[0:2]
+                dst = chunk[2:4]
+                out.append(f"[{src}]")
+                out.append(f"[{dst}]")
+                if len(chunk) == 5 and chunk[4] in "qrbn":
+                    out.append(f"[={chunk[4]}]")
+                continue
+
+            m = _MOVE_RE.match(chunk)
+            if not m:
+                out.append(self.UNK_TOKEN)
+                continue
+
+            side = "[W]" if m.group("side") == "W" else "[BL]"
+            piece = m.group("piece")
+            src = m.group("src")
+            dst = m.group("dst")
+            suffix = m.group("suffix") or ""
+
+            out.append(side)
+            out.append(f"[{piece}]")
+            out.append(f"[{src}]")
+            out.append(f"[{dst}]")
+
+            is_cap, is_chk, is_mate, castle_kind, promo = _parse_suffix(suffix)
+
+            # Castling markers (optional; evaluator doesn't need them)
+            if castle_kind == "k":
+                out.append("[O-O]")
+            elif castle_kind == "q":
+                out.append("[O-O-O]")
+
+            if is_cap:
+                out.append("[x]")
+            if is_mate:
+                out.append("[#]")
+            elif is_chk:
+                out.append("[+]")
+
+            if promo in ("q", "r", "b", "n"):
+                out.append(f"[={promo}]")
+
+        return out
+
+    
+    def _convert_token_to_id(self, token: str) -> int:
+        """Convert a token to its ID."""
+        return self._vocab.get(token, self._vocab.get(self.UNK_TOKEN, 0))
+    
+    def _convert_id_to_token(self, index: int) -> str:
+        """Convert an ID to its token."""
+        return self._ids_to_tokens.get(index, self.UNK_TOKEN)
+    
+    def convert_tokens_to_string(self, tokens: List[str]) -> str:
+        """Convert a list of tokens back to a string."""
+        # Filter out special tokens for cleaner output
+        special = {self.PAD_TOKEN, self.BOS_TOKEN, self.EOS_TOKEN, self.UNK_TOKEN}
+        return " ".join(t for t in tokens if t not in special)
+    
+    def save_vocabulary(
+        self,
+        save_directory: str,
+        filename_prefix: Optional[str] = None,
+    ) -> tuple:
+        """
+        Save the vocabulary to a JSON file.
+        
+        Args:
+            save_directory: Directory to save the vocabulary.
+            filename_prefix: Optional prefix for the filename.
+        
+        Returns:
+            Tuple containing the path to the saved vocabulary file.
+        """
+        if not os.path.isdir(save_directory):
+            os.makedirs(save_directory, exist_ok=True)
+        
+        vocab_file = os.path.join(
+            save_directory,
+            (filename_prefix + "-" if filename_prefix else "") + "vocab.json",
+        )
+        
+        with open(vocab_file, "w", encoding="utf-8") as f:
+            json.dump(self._vocab, f, ensure_ascii=False, indent=2)
+        
+        return (vocab_file,)
+
+
+def count_vocab_from_dataset(
+    dataset_name: str = "dlouapre/lichess_2025-01_1M",
+    split: str = "train",
+    column: str = "text",
+    max_samples: Optional[int] = 10000,
+) -> Dict[str, int]:
+    """
+    Count token frequencies in a dataset (useful for vocabulary analysis).
+    
+    Args:
+        dataset_name: Name of the dataset on Hugging Face Hub.
+        split: Dataset split to use.
+        column: Column containing the game strings.
+        max_samples: Maximum number of samples to process.
+    
+    Returns:
+        Dictionary mapping tokens to their frequencies.
+    """
+    from collections import Counter
+    from datasets import load_dataset
+
+    ds = load_dataset(dataset_name, split=split)
+    if max_samples is not None:
+        ds = ds.select(range(min(max_samples, len(ds))))
+
+    tok = ChessTokenizer()
+    counts = Counter()
+    for ex in ds:
+        counts.update(tok._tokenize(ex[column]))
+    return dict(counts)

tokenizer_config.json

@@ -0,0 +1,50 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "[BOS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "[EOS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenizer.ChessTokenizer",
+      null
+    ]
+  },
+  "bos_token": "[BOS]",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "[EOS]",
+  "extra_special_tokens": {},
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "[PAD]",
+  "tokenizer_class": "ChessTokenizer",
+  "unk_token": "[UNK]"
+}

vocab.json

@@ -0,0 +1,86 @@
+{
+  "[PAD]": 0,
+  "[BOS]": 1,
+  "[EOS]": 2,
+  "[UNK]": 3,
+  "[W]": 4,
+  "[P]": 5,
+  "[N]": 6,
+  "[B]": 7,
+  "[R]": 8,
+  "[Q]": 9,
+  "[K]": 10,
+  "[a1]": 11,
+  "[a2]": 12,
+  "[a3]": 13,
+  "[a4]": 14,
+  "[a5]": 15,
+  "[a6]": 16,
+  "[a7]": 17,
+  "[a8]": 18,
+  "[b1]": 19,
+  "[b2]": 20,
+  "[b3]": 21,
+  "[b4]": 22,
+  "[b5]": 23,
+  "[b6]": 24,
+  "[b7]": 25,
+  "[b8]": 26,
+  "[c1]": 27,
+  "[c2]": 28,
+  "[c3]": 29,
+  "[c4]": 30,
+  "[c5]": 31,
+  "[c6]": 32,
+  "[c7]": 33,
+  "[c8]": 34,
+  "[d1]": 35,
+  "[d2]": 36,
+  "[d3]": 37,
+  "[d4]": 38,
+  "[d5]": 39,
+  "[d6]": 40,
+  "[d7]": 41,
+  "[d8]": 42,
+  "[e1]": 43,
+  "[e2]": 44,
+  "[e3]": 45,
+  "[e4]": 46,
+  "[e5]": 47,
+  "[e6]": 48,
+  "[e7]": 49,
+  "[e8]": 50,
+  "[f1]": 51,
+  "[f2]": 52,
+  "[f3]": 53,
+  "[f4]": 54,
+  "[f5]": 55,
+  "[f6]": 56,
+  "[f7]": 57,
+  "[f8]": 58,
+  "[g1]": 59,
+  "[g2]": 60,
+  "[g3]": 61,
+  "[g4]": 62,
+  "[g5]": 63,
+  "[g6]": 64,
+  "[g7]": 65,
+  "[g8]": 66,
+  "[h1]": 67,
+  "[h2]": 68,
+  "[h3]": 69,
+  "[h4]": 70,
+  "[h5]": 71,
+  "[h6]": 72,
+  "[h7]": 73,
+  "[h8]": 74,
+  "[x]": 75,
+  "[+]": 76,
+  "[#]": 77,
+  "[O-O]": 78,
+  "[O-O-O]": 79,
+  "[=q]": 80,
+  "[=r]": 81,
+  "[=b]": 82,
+  "[=n]": 83
+}