Micro test upload (fixed v3)

README.md

@@ -0,0 +1,8 @@
+---
+library_name: transformers
+tags:
+- chess
+license: mit
+---
+# Micro Test
+This is just a test upload for MDaytek/Model-v3-mdaytek.

config.json

@@ -0,0 +1,14 @@
+{
+    "architectures": [
+        "ChessForCausalLM"
+    ],
+    "dtype": "float32",
+    "model_type": "chess_lm",
+    "n_ctx": 256,
+    "n_embd": 128,
+    "n_head": 4,
+    "n_layer": 4,
+    "n_positions": 256,
+    "transformers_version": "4.57.3",
+    "vocab_size": 509
+}

model.py

@@ -0,0 +1,68 @@
+
+import torch
+import torch.nn as nn
+from transformers import PreTrainedModel, PretrainedConfig
+
+class ChessConfig(PretrainedConfig):
+    model_type = "chess_lm"
+    def __init__(
+        self,
+        vocab_size=1200,
+        n_positions=256,
+        n_embd=128,
+        n_layer=4,
+        n_head=4,
+        n_ctx=256,
+        tie_word_embeddings=True,
+        **kwargs,
+    ):
+        self.vocab_size = vocab_size
+        self.n_positions = n_positions
+        self.n_embd = n_embd
+        self.n_layer = n_layer
+        self.n_head = n_head
+        self.n_ctx = n_ctx
+        self.tie_word_embeddings = tie_word_embeddings
+        super().__init__(**kwargs)
+
+class ChessForCausalLM(PreTrainedModel):
+    config_class = ChessConfig
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.config = config
+        self.token_embedding = nn.Embedding(config.vocab_size, config.n_embd)
+        self.position_embedding = nn.Embedding(config.n_positions, config.n_embd)
+        encoder_layer = nn.TransformerEncoderLayer(
+            d_model=config.n_embd, nhead=config.n_head, dim_feedforward=config.n_embd * 4,
+            batch_first=True, norm_first=True
+        )
+        self.blocks = nn.TransformerEncoder(encoder_layer, num_layers=config.n_layer)
+        self.ln_f = nn.LayerNorm(config.n_embd)
+        self.head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
+        if config.tie_word_embeddings:
+            self.head.weight = self.token_embedding.weight
+        self.post_init()
+
+    def get_input_embeddings(self): return self.token_embedding
+    def set_input_embeddings(self, value): self.token_embedding = value
+
+    def forward(self, input_ids, labels=None, **kwargs):
+        B, T = input_ids.shape
+        tok_emb = self.token_embedding(input_ids)
+        pos_emb = self.position_embedding(torch.arange(T, device=input_ids.device))
+        x = tok_emb + pos_emb
+        mask = torch.triu(torch.ones(T, T, device=input_ids.device) * float('-inf'), diagonal=1)
+        x = self.blocks(x, mask=mask, is_causal=True)
+        x = self.ln_f(x)
+        logits = self.head(x)
+        loss = None
+        if labels is not None:
+            shift_logits = logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            loss_fct = nn.CrossEntropyLoss()
+            loss = loss_fct(shift_logits.view(-1, self.config.vocab_size), shift_labels.view(-1))
+        return (loss, logits) if loss is not None else logits
+
+def print_parameter_budget(config):
+    print(f"Model params: Check")

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00546273a45b93e2b8d7f7d880cd22340ccb06f5239c718f8c6c3b6f5f540516
+size 3581859

tokenizer.py

@@ -0,0 +1,41 @@
+
+import json
+import os
+import torch
+
+class ChessTokenizer:
+    def __init__(self, vocab=None):
+        self.vocab = vocab if vocab else {}
+        self.id_to_token = {v: k for k, v in self.vocab.items()}
+        self.pad_token_id = self.vocab.get("[PAD]", 0)
+        self.bos_token_id = self.vocab.get("[BOS]", 1)
+        self.eos_token_id = self.vocab.get("[EOS]", 2)
+
+    @property
+    def vocab_size(self):
+        return len(self.vocab)
+
+    def _convert_token_to_id(self, token):
+        return self.vocab.get(token, self.vocab.get("[UNK]"))
+
+    def pad(self, encoded_inputs, padding=True, max_length=None, pad_to_multiple_of=None, return_tensors=None):
+        batch_ids = [x["input_ids"] for x in encoded_inputs]
+        max_len = max(len(ids) for ids in batch_ids)
+        padded_batch = []
+        for ids in batch_ids:
+            padded_ids = ids + [self.pad_token_id] * (max_len - len(ids))
+            padded_batch.append(padded_ids)
+        if return_tensors == "pt":
+            return {"input_ids": torch.tensor(padded_batch, dtype=torch.long)}
+        return {"input_ids": padded_batch}
+
+    def save_pretrained(self, save_directory):
+        os.makedirs(save_directory, exist_ok=True)
+        with open(os.path.join(save_directory, "vocab.json"), "w") as f:
+            json.dump(self.vocab, f, indent=4)
+
+    @classmethod
+    def from_pretrained(cls, load_directory):
+        with open(os.path.join(load_directory, "vocab.json"), "r") as f:
+            vocab = json.load(f)
+        return cls(vocab)

train.py

@@ -0,0 +1,73 @@
+
+import json
+import os
+from collections import Counter
+from datasets import load_dataset, Dataset
+from transformers import Trainer, TrainingArguments, DataCollatorForLanguageModeling
+from src.tokenizer import ChessTokenizer
+from src.model import ChessConfig, ChessForCausalLM, print_parameter_budget
+
+def build_vocab(dataset, max_vocab=1200):
+    counter = Counter()
+    for game in dataset:
+        moves = game["text"].split()
+        counter.update(moves)
+    special = ["[PAD]", "[BOS]", "[EOS]", "[UNK]"]
+    vocab_tokens = special + [t for t, _ in counter.most_common(max_vocab - len(special))]
+    vocab = {tok: i for i, tok in enumerate(vocab_tokens)}
+    return vocab
+
+def encode_game(game, tokenizer, max_len=256):
+    moves = game["text"].split()
+    tokens = ["[BOS]"] + moves + ["[EOS]"]
+    tokens = tokens[:max_len]
+    ids = [tokenizer._convert_token_to_id(t) for t in tokens]
+    ids += [tokenizer.pad_token_id] * (max_len - len(ids))
+    return ids
+
+def main():
+    print("Loading MICRO dataset...")
+    raw_ds = load_dataset("dlouapre/lichess_2025-01_1M", split="train[:10]")
+
+    print("Building tokenizer...")
+    vocab = build_vocab(raw_ds)
+    tokenizer = ChessTokenizer(vocab)
+
+    config = ChessConfig(vocab_size=tokenizer.vocab_size)
+    model = ChessForCausalLM(config)
+
+    print("Tokenizing dataset...")
+    input_ids = [encode_game(g, tokenizer) for g in raw_ds]
+    ds = Dataset.from_dict({"input_ids": input_ids})
+
+    train_output = "./my_model"
+
+    args = TrainingArguments(
+        output_dir=train_output,
+        per_device_train_batch_size=2,
+        num_train_epochs=1,
+        save_strategy="no",
+        report_to="none",
+        use_cpu=False
+    )
+
+    trainer = Trainer(
+        model=model,
+        args=args,
+        train_dataset=ds,
+        data_collator=DataCollatorForLanguageModeling(tokenizer, mlm=False),
+    )
+
+    print("🚀 Starting MICRO training...")
+    trainer.train()
+
+    print("💾 Saving model locally...")
+    final_path = os.path.join(train_output, "final_model")
+    
+    # SAUVEGARDE CORRIGÉE
+    model.save_pretrained(final_path, safe_serialization=False)
+    tokenizer.save_pretrained(final_path)
+    print("✅ Training complete.")
+
+if __name__ == "__main__":
+    main()

vocab.json

@@ -0,0 +1,511 @@
+{
+    "[PAD]": 0,
+    "[BOS]": 1,
+    "[EOS]": 2,
+    "[UNK]": 3,
+    "WPe2e4": 4,
+    "BNg8f6": 5,
+    "BKe8g8(o)": 6,
+    "BPg7g6": 7,
+    "WPd2d4": 8,
+    "WNg1f3": 9,
+    "WPh2h3": 10,
+    "BPd7d5": 11,
+    "BNb8c6": 12,
+    "WNb1c3": 13,
+    "BPe7e5": 14,
+    "WBf1e2": 15,
+    "BPc7c6": 16,
+    "WKe1g1(o)": 17,
+    "WNb1d2": 18,
+    "BPd7d6": 19,
+    "BPe7e6": 20,
+    "WPb2b3": 21,
+    "WPa2a4": 22,
+    "WBc1e3": 23,
+    "WPc2c3": 24,
+    "WPg2g3": 25,
+    "BPb7b5": 26,
+    "WPd2d3": 27,
+    "BBc8g4": 28,
+    "WBe2f3(x)": 29,
+    "BPb7b6": 30,
+    "BNc6e5(x)": 31,
+    "BNe5f3(x+)": 32,
+    "WPh3h4": 33,
+    "BPf7f6": 34,
+    "BPa7a5": 35,
+    "WPh4h5": 36,
+    "WKg1h2": 37,
+    "BPf6f5": 38,
+    "WBc1f4": 39,
+    "BBf8d6": 40,
+    "WPc3c4": 41,
+    "BNf6e4": 42,
+    "BPf7f5": 43,
+    "BPa7a6": 44,
+    "WBf1d3": 45,
+    "WNd2b3": 46,
+    "BNb8d7": 47,
+    "WPf2f4": 48,
+    "BBf8g7": 49,
+    "BBf8c5": 50,
+    "BNf6h5": 51,
+    "BNh5f6": 52,
+    "WQf3g3": 53,
+    "BNf6g4": 54,
+    "BPh7h5": 55,
+    "WPf2f3": 56,
+    "WNc3d5": 57,
+    "WPe4e5": 58,
+    "BBg4f3(x)": 59,
+    "BPc6c5": 60,
+    "WQd1e2": 61,
+    "BNg8e7": 62,
+    "BPb6c5(x)": 63,
+    "WQe2b5": 64,
+    "WBc1d2": 65,
+    "WRa1d1": 66,
+    "WPf2g3(x)": 67,
+    "BNh4g6": 68,
+    "BPe6e5": 69,
+    "BPd5d4": 70,
+    "BRf8e8": 71,
+    "WPg3g4": 72,
+    "BPf5f4": 73,
+    "BPe4e3": 74,
+    "WPg4g5": 75,
+    "BKg8h8": 76,
+    "WPd4e5(x)": 77,
+    "BQd8d6": 78,
+    "WKg1h1": 79,
+    "BRa8e8": 80,
+    "WRa1e1": 81,
+    "BPd5e4(x)": 82,
+    "WRe1e4(x)": 83,
+    "BPh7h6": 84,
+    "WPa4a5": 85,
+    "BKh7h6": 86,
+    "BBc8e6": 87,
+    "BRa8d8": 88,
+    "WRf1e1": 89,
+    "WQd1d3": 90,
+    "BQa5b5": 91,
+    "BNd7f6": 92,
+    "WRh1e1": 93,
+    "BRf8d8": 94,
+    "WPh5g6(x)": 95,
+    "BPh7g6(x)": 96,
+    "WPc2c4": 97,
+    "BPc7c5": 98,
+    "WPd4d5": 99,
+    "BBc8d7": 100,
+    "WPa2a3": 101,
+    "WPb2b4": 102,
+    "BQd8e8": 103,
+    "WPe4f5(x)": 104,
+    "BPe6f5(x)": 105,
+    "WBf4e3": 106,
+    "BKg8g7": 107,
+    "BKg7h6": 108,
+    "BKh6h5": 109,
+    "WPg2g4": 110,
+    "WQd1h5": 111,
+    "WQh5e5(x+)": 112,
+    "BBc5e7": 113,
+    "WQe5f4": 114,
+    "WQf4f3": 115,
+    "WBe2g4(x)": 116,
+    "BBc8g4(x)": 117,
+    "BPh5h4": 118,
+    "WQg3g4(x)": 119,
+    "BBe7f8": 120,
+    "WBc1g5": 121,
+    "BQd8d7": 122,
+    "WNd5f6(+)": 123,
+    "BNf6g8": 124,
+    "WPd4c5(x)": 125,
+    "BBf8c5(x)": 126,
+    "WNc3a4": 127,
+    "WNa4c5(x)": 128,
+    "WQb5c5(x)": 129,
+    "WQc5c3": 130,
+    "WQc3f3(x)": 131,
+    "BNe7f5": 132,
+    "BRa8b8": 133,
+    "WBd2c3": 134,
+    "BQd8g5": 135,
+    "BNf5h4": 136,
+    "BQg5g3(x)": 137,
+    "WRd1d4": 138,
+    "BRb8b5": 139,
+    "WRd4g4": 140,
+    "WBc3d2": 141,
+    "BRb5c5": 142,
+    "WRf1c1": 143,
+    "BNg6f8": 144,
+    "WPh5h6": 145,
+    "WRg4h4": 146,
+    "BPe5e4": 147,
+    "WBd2f4": 148,
+    "BPd4d3": 149,
+    "WBf4d6": 150,
+    "BRc5c2(x)": 151,
+    "WRc1c2(x)": 152,
+    "BPd3c2(x)": 153,
+    "WBd6f4": 154,
+    "BNf8e6": 155,
+    "WBf4c1": 156,
+    "BNe6d4": 157,
+    "BNd4e2": 158,
+    "WBe3d2": 159,
+    "WBd2e3(x)": 160,
+    "BPf4e3(x)": 161,
+    "BPc2c1(Q)": 162,
+    "WRh4c4": 163,
+    "BQc1g1(+)": 164,
+    "WKh2h3": 165,
+    "BRe8f8": 166,
+    "WRc4c7": 167,
+    "BNe2f4(+)": 168,
+    "WKh3g3": 169,
+    "BQg1e1(+)": 170,
+    "WKg3h2": 171,
+    "BQe1h4(+)": 172,
+    "WKh2g1": 173,
+    "BQh4g5(x)": 174,
+    "WRc7g7(+)": 175,
+    "WRg7h7(x+)": 176,
+    "BKh8h7(x)": 177,
+    "WKg1f1": 178,
+    "BQg5h6(x)": 179,
+    "WPe2f3(x)": 180,
+    "WBf4e5(x)": 181,
+    "BBd6e5(x)": 182,
+    "BBe5h2": 183,
+    "BBh2g3(x)": 184,
+    "BQd6g3(x)": 185,
+    "WQd1e1": 186,
+    "BQg3h3(x+)": 187,
+    "WKh1g1": 188,
+    "WQe1f2": 189,
+    "BRe8e6": 190,
+    "WPf3e4(x)": 191,
+    "BRe6f6": 192,
+    "WQf2g2": 193,
+    "BRf6f1(x+)": 194,
+    "WBe2f1(x)": 195,
+    "BQh3g2(x+)": 196,
+    "WBf1g2(x)": 197,
+    "BRe8d8": 198,
+    "WNd2f3": 199,
+    "BRd8d3": 200,
+    "WPc4b5(x)": 201,
+    "BRd3d5": 202,
+    "BRd5h5(+)": 203,
+    "WKh2g3": 204,
+    "WRe4e5": 205,
+    "WKg3f4": 206,
+    "BPg6g5(+)": 207,
+    "WKf4f5(x)": 208,
+    "BPg5g4(+)": 209,
+    "WKf5f6": 210,
+    "BPg4f3(x)": 211,
+    "WRe5h5(x)": 212,
+    "BKg8h7": 213,
+    "WBg2f3(x)": 214,
+    "WRh5h1": 215,
+    "BPc6b5(x)": 216,
+    "WRh1g1": 217,
+    "BPb5a4(x)": 218,
+    "WPb3a4(x)": 219,
+    "BPh6h5": 220,
+    "WBf3h5(x)": 221,
+    "WBh5e2": 222,
+    "BKh6h7": 223,
+    "WBe2d1": 224,
+    "WRg1h1(+*)": 225,
+    "WPe4d5(x)": 226,
+    "BPc6d5(x)": 227,
+    "WBd3e2": 228,
+    "BQd8c7": 229,
+    "WNb3d4": 230,
+    "WRe1e6(x)": 231,
+    "BPf7e6(x)": 232,
+    "WNd4e6(x)": 233,
+    "BQc7a5": 234,
+    "WNe6f8(x)": 235,
+    "BRd8f8(x)": 236,
+    "WBe3d4": 237,
+    "BBd6c5": 238,
+    "BBc5d4(x)": 239,
+    "WQd3d4(x)": 240,
+    "BQb5b2(x)": 241,
+    "WQd4d5(x+)": 242,
+    "WBf3e4(x)": 243,
+    "BQb2f2(x+)": 244,
+    "BQf2g3": 245,
+    "WQd5e6": 246,
+    "WQe6g6": 247,
+    "BQg3g6(x)": 248,
+    "WBe4g6(x)": 249,
+    "BRf8f2": 250,
+    "WPd3e4(x)": 251,
+    "BQd8d1(x+)": 252,
+    "WKe1d1(x)": 253,
+    "WKd1e2": 254,
+    "BNf6e4(x)": 255,
+    "BNe4d2(x)": 256,
+    "WNf3d2(x)": 257,
+    "WBf1g2": 258,
+    "BNf6d5": 259,
+    "WKe2d1": 260,
+    "BNd5e3(x+)": 261,
+    "WRe1e3(x)": 262,
+    "BBc8b7": 263,
+    "WNd2e4": 264,
+    "BBd6e7": 265,
+    "WKd1c1": 266,
+    "WNe4g5": 267,
+    "BBe7g5(x)": 268,
+    "WPf4g5(x)": 269,
+    "WPh2h4": 270,
+    "BKg8f8": 271,
+    "BKf8e7": 272,
+    "BRd8d4": 273,
+    "WBg2f3": 274,
+    "BRd8d7": 275,
+    "WBf3b7(x)": 276,
+    "BRd7b7(x)": 277,
+    "WRe3e4": 278,
+    "BRb7d7": 279,
+    "WKc1b2": 280,
+    "BRd4e4(x)": 281,
+    "BRd7d5": 282,
+    "BRd5g5(x)": 283,
+    "WPa5b6(x)": 284,
+    "BPa7b6(x)": 285,
+    "WRe4a4": 286,
+    "WPg4f5(x)": 287,
+    "BPg6f5(x)": 288,
+    "WKb2c1": 289,
+    "BRg5g4": 290,
+    "WRa4a6": 291,
+    "BRg4b4": 292,
+    "WKc1d2": 293,
+    "WKd2d3": 294,
+    "BPe5e4(+)": 295,
+    "WKd3e3": 296,
+    "BKe7e6": 297,
+    "WRa6a8": 298,
+    "BPc5c4": 299,
+    "WPb3c4(x)": 300,
+    "BRb4c4(x)": 301,
+    "WRa8e8(+)": 302,
+    "BKe6f7": 303,
+    "WRe8b8": 304,
+    "BRc4c3(+)": 305,
+    "WKe3f4": 306,
+    "BRc3f3(+)": 307,
+    "WKf4e5": 308,
+    "WRb8b6(x)": 309,
+    "BPe3e2": 310,
+    "WRb6b7(+)": 311,
+    "BKf7e8": 312,
+    "WKe5f6": 313,
+    "BKe8d8": 314,
+    "WRb7e7": 315,
+    "BRf3f2": 316,
+    "WRe7e5": 317,
+    "WKf6f5": 318,
+    "BPf4f3": 319,
+    "WKf5f4": 320,
+    "BRf2f1": 321,
+    "BPe2e1(Q)": 322,
+    "WRe5e1(x)": 323,
+    "BRf1e1(x)": 324,
+    "WKf4f3(x)": 325,
+    "BKd8d7": 326,
+    "BBg7c3(x+)": 327,
+    "WPb2c3(x)": 328,
+    "BNd7e5": 329,
+    "WBe3h6": 330,
+    "BQd8a5": 331,
+    "BBc8a6": 332,
+    "WQd3d2": 333,
+    "BKe8c8(O)": 334,
+    "BPb5b4": 335,
+    "WRf1b1": 336,
+    "BPb4c3(x)": 337,
+    "WQd2f4": 338,
+    "BKc8d7": 339,
+    "WBf3g4(+)": 340,
+    "BKd7e8": 341,
+    "WBh6g7": 342,
+    "BRh8g8": 343,
+    "WBg7f6(x)": 344,
+    "BPe7f6(x)": 345,
+    "WQf4f6(x)": 346,
+    "WRb1b5(x)": 347,
+    "BBa6b5(x)": 348,
+    "WPe2e3": 349,
+    "WNg1e2": 350,
+    "BBd7c6": 351,
+    "WQd1c2": 352,
+    "WNb3a5": 353,
+    "BBc6d5": 354,
+    "BBd5c6": 355,
+    "WNa5c6(x)": 356,
+    "BNb8c6(x)": 357,
+    "BNc6b8": 358,
+    "WRa1c1": 359,
+    "WPe3e4": 360,
+    "BQe8f7": 361,
+    "WNe2g3": 362,
+    "WRe1e2": 363,
+    "WPd5c6(x)": 364,
+    "BNe7c6(x)": 365,
+    "WRc1e1": 366,
+    "BNc6d4": 367,
+    "WQc2d2": 368,
+    "BNd4e2(x+)": 369,
+    "WRe1e2(x)": 370,
+    "BRe8e2(x)": 371,
+    "WNg3e2(x)": 372,
+    "WNe2d4": 373,
+    "WBd3e4(x)": 374,
+    "BPf5e4(x)": 375,
+    "BPd6d5": 376,
+    "WPc4c5": 377,
+    "WPb4c5(x)": 378,
+    "BRd8c8": 379,
+    "WPc5c6": 380,
+    "BBg7d4(x)": 381,
+    "WBe3d4(x)": 382,
+    "BRc8c6(x)": 383,
+    "WPa3a4": 384,
+    "BRc6c4": 385,
+    "WBd4b6": 386,
+    "BRc4a4(x)": 387,
+    "WQd2c2": 388,
+    "BRa4a1(+)": 389,
+    "BQf7f4(+)": 390,
+    "BQf4f3": 391,
+    "WQc2c8(+)": 392,
+    "WBb6d4(+)": 393,
+    "WBd4e3(+)": 394,
+    "WPg3g4(+)": 395,
+    "BKh5h4": 396,
+    "WQc8d8(+)": 397,
+    "WQd1d2": 398,
+    "WKe1c1(O)": 399,
+    "WBe3g5": 400,
+    "WBg5h4": 401,
+    "BNg4h6": 402,
+    "BNh6f7": 403,
+    "WBd3c4": 404,
+    "BQe8d8": 405,
+    "WBh4g3": 406,
+    "BNc6a5": 407,
+    "WBc4d3": 408,
+    "BNa5c6": 409,
+    "BNc6b4": 410,
+    "WRe1h1": 411,
+    "BNb4d3(x+)": 412,
+    "WQd2d3(x)": 413,
+    "BNf7g5(x)": 414,
+    "WBg3h4": 415,
+    "BBg7f6": 416,
+    "WBh4g5(x)": 417,
+    "BBf6g5(x+)": 418,
+    "WKc1b1": 419,
+    "BQd8f6": 420,
+    "WQd3c4(+)": 421,
+    "WQc4c7(x)": 422,
+    "BQf6f7": 423,
+    "WNf3g5(x)": 424,
+    "BQf7e7": 425,
+    "WRh1h7(+)": 426,
+    "BKg7f6": 427,
+    "WRh7e7(x)": 428,
+    "BKf6e7(x)": 429,
+    "WRd1e1(+)": 430,
+    "BKe7f6": 431,
+    "WNg5h7(+)": 432,
+    "BKf6g7": 433,
+    "WNh7f8(x)": 434,
+    "BRa8f8(x)": 435,
+    "WQc7d7(x+)": 436,
+    "BRf8f7": 437,
+    "WQd7d6(x)": 438,
+    "BRf7f6": 439,
+    "BRf6d6(x)": 440,
+    "WKb1c2": 441,
+    "BPa5a4": 442,
+    "WKc2c3": 443,
+    "WNd5e3": 444,
+    "BPa4b3(x)": 445,
+    "WRe1h1(+*)": 446,
+    "BPe5f4(x)": 447,
+    "BBf8e7": 448,
+    "BBe7h4(+)": 449,
+    "WKe1e2": 450,
+    "WBc1f4(x)": 451,
+    "BBg4f3(x+)": 452,
+    "WKe2f3(x)": 453,
+    "BBh4g5": 454,
+    "BBg5f4(x)": 455,
+    "WPg3f4(x)": 456,
+    "BNg8h6": 457,
+    "WBf1h3": 458,
+    "BNc6e7": 459,
+    "BNe7g6": 460,
+    "BNg6h4(x+)": 461,
+    "WKf3g3": 462,
+    "WNc3d5(x)": 463,
+    "BQd8h4(+)": 464,
+    "WKg3f3": 465,
+    "BQh4h5(+)": 466,
+    "WKf3e3": 467,
+    "BQh5h4": 468,
+    "WNd5c7(x)": 469,
+    "BQh4g3(+)": 470,
+    "WKe3d2": 471,
+    "BQg3f4(x+)": 472,
+    "WKd2c3": 473,
+    "BRa8c8": 474,
+    "WQb5b7(x)": 475,
+    "BQf4e3(+)": 476,
+    "WKc3c4": 477,
+    "BQe3e2(+)": 478,
+    "WKc4c5": 479,
+    "BQe2c2(x+)": 480,
+    "WKc5d6": 481,
+    "BRf8d8(+)": 482,
+    "WKd6e6": 483,
+    "BRc8c7(x)": 484,
+    "WQb7b3": 485,
+    "BQc2c6(+*)": 486,
+    "BNc6b4(x)": 487,
+    "WPa3b4(x)": 488,
+    "BBc5b4(x+)": 489,
+    "BBb4c5": 490,
+    "BBe6d7": 491,
+    "WKe1f2": 492,
+    "WKf2g3": 493,
+    "BPh5h4(+)": 494,
+    "WKg3f2": 495,
+    "BNf6h7": 496,
+    "WBd2c1": 497,
+    "BPg7g5": 498,
+    "WBc1a3": 499,
+    "WBa3c5(x)": 500,
+    "BQd6c5(x)": 501,
+    "BQc5b6": 502,
+    "WRa1b1": 503,
+    "WNb3d4(x)": 504,
+    "BQb6d6": 505,
+    "WNd4b5": 506,
+    "BQd6d4(+)": 507,
+    "WNb5d4(x)": 508
+}