fix(tokenizer): 移除异常处理，直接加载指定tokenizer

eval.py

@@ -78,16 +78,11 @@ class TextEvaluator:
 
     def load_tokenizer(self):
         """加载tokenizer"""
-        try:
+        tokenizer_path = (
-            tokenizer_path = (
+            Path(__file__).parent / "src" / "model" / "assets" / "tokenizer"
-                Path(__file__).parent / "src" / "model" / "assets" / "tokenizer"
+        )
-            )
+        self.tokenizer = AutoTokenizer.from_pretrained(str(tokenizer_path))
-            self.tokenizer = AutoTokenizer.from_pretrained(str(tokenizer_path))
+        print(f"✅ Tokenizer加载完成，词汇表大小: {self.tokenizer.vocab_size}")
-            print(f"✅ Tokenizer加载完成，词汇表大小: {self.tokenizer.vocab_size}")
-        except Exception as e:
-            print(f"⚠️ 无法加载tokenizer: {e}")
-            print("使用默认的bert-base-chinese tokenizer")
-            self.tokenizer = AutoTokenizer.from_pretrained("bert-base-chinese")
 
     def load_query_engine(self):
         """加载查询引擎用于字符-ID转换"""

src/model/components.py

@@ -180,14 +180,15 @@ class SlotMemory(nn.Module):
         Args:
             history_ids: [batch, total_steps]
                          Flattened sequence of history tokens.
-                         Empty positions should be filled with a special PAD or handled via mask.
+                         Zero positions use start_emb as learnable query.
         Returns:
             S: [batch, total_steps, 512] Slot sequence representation [1]
         """
-        # Embed history tokens
+        S = self.emb(history_ids)  # [B, num_slots, dim]
-        S = self.emb(history_ids)  # [B, 24, 512]
+
+        zero_mask = (history_ids == 0).unsqueeze(-1).float()  # [B, num_slots, 1]
+        S = S * (1 - zero_mask) + self.start_emb * zero_mask
 
-        # Add positional embeddings
         pos_ids = (
             torch.arange(S.size(1), device=S.device).unsqueeze(0).expand_as(history_ids)
         )

src/model/dataset.py

@@ -42,7 +42,8 @@ class PinyinInputDataset(IterableDataset):
         max_seq_length=128,
         text_field: str = "text",
         py_style_weight=(9, 2, 1),
-        shuffle_buffer_size: int = 5000,
+        shuffle_buffer_size: int = 100000,
+        retention_ratio: float = 0.5,
         length_weights={1: 10, 2: 50, 3: 50, 4: 40, 5: 15, 6: 10, 7: 5, 8: 2},
     ):
         # 频率调整参数 (可根据需要调整)
@@ -64,6 +65,16 @@ class PinyinInputDataset(IterableDataset):
         self.max_workers = max_workers
         self.py_style_weight = np.array(py_style_weight) / sum(py_style_weight)
         self.shuffle_buffer_size = shuffle_buffer_size
+        self.retention_ratio = retention_ratio
+        if not (0 < retention_ratio < 1):
+            raise ValueError(
+                f"retention_ratio必须在0和1之间，当前值: {retention_ratio}"
+            )
+        self.retention_size = int(shuffle_buffer_size * retention_ratio)
+        if self.retention_size <= 0:
+            raise ValueError(
+                f"计算出的retention_size必须大于0，当前值: {self.retention_size} (shuffle_buffer_size={shuffle_buffer_size}, retention_ratio={retention_ratio})"
+            )
         self.possible_lengths = list(length_weights.keys())
         self.weights = list(length_weights.values())
 
@@ -360,17 +371,31 @@ class PinyinInputDataset(IterableDataset):
                     # 添加到缓冲区
                     batch_samples.extend(samples)
 
-            # 处理shuffle buffer
+            # 处理shuffle buffer - 单缓冲区半保留方案
             if len(batch_samples) >= self.shuffle_buffer_size:
+                # 全量打乱缓冲区
                 indices = np.random.permutation(len(batch_samples))
-                for idx in indices:
+
+                # 计算实际保留大小（不超过缓冲区大小）
+                actual_retention = min(self.retention_size, len(batch_samples))
+
+                # 计算输出数量
+                output_count = len(batch_samples) - actual_retention
+
+                # 输出前output_count个样本
+                for i in range(output_count):
                     if current_iter_index >= worker_quota:
-                        # 清空batch_samples并返回
+                        # 配额用完，清空缓冲区并返回
                         batch_samples = []
-                        return  # 使用return而不是break，因为我们在生成器函数中
+                        return
-                    yield batch_samples[idx]
+                    yield batch_samples[indices[i]]
                     current_iter_index += 1
-                batch_samples = []
+
+                # 保留后actual_retention个样本（不清空缓冲区）
+                retained_samples = [
+                    batch_samples[idx] for idx in indices[output_count:]
+                ]
+                batch_samples = retained_samples
 
         # 处理剩余的样本
         if batch_samples:

src/model/model.py

@@ -124,20 +124,18 @@ class InputMethodEngine(nn.Module):
         moe_out = self.moe(fused)
 
         # 5. 池化与分类：对槽位维度求平均（使用 mask 池化，完全兼容 torch.compile）
-        # 使用显式形状重塑，彻底杜绝广播歧义
         batch_size = input_ids.size(0)
         slot_mask = (history_slot_ids != 0).float().view(batch_size, self.num_slots, 1)
         numerator = (moe_out * slot_mask).sum(dim=1)  # [batch, dim]
         denominator = slot_mask.view(batch_size, -1).sum(dim=1)  # [batch]
-        # 若无有效槽位，使用上下文 H 的掩码均值
+
-        if torch.all(denominator == 0):
+        all_slot_mean = moe_out.mean(dim=1)  # [batch, dim]
-            # H: [batch, seq_len, dim], attention_mask: [batch, seq_len]
+        all_zero = denominator == 0  # [batch]
-            ctx_mask = attention_mask.float().unsqueeze(-1)  # [batch, seq_len, 1]
+        pooled = torch.where(
-            ctx_sum = (H * ctx_mask).sum(dim=1)  # [batch, dim]
+            all_zero.unsqueeze(-1),
-            ctx_cnt = ctx_mask.sum(dim=1) + 1e-8  # [batch, 1]
+            all_slot_mean,
-            pooled = ctx_sum / ctx_cnt
+            numerator / (denominator.unsqueeze(-1) + 1e-8),
-        else:
+        )
-            pooled = numerator / (denominator.unsqueeze(-1) + 1e-8)  # [batch, dim]
 
         logits = self.classifier(pooled)  # [batch, vocab_size]
         return logits

src/model/trainer.py

@@ -1070,7 +1070,7 @@ def train(
         max_seq_length=max_seq_len,
         text_field="text",
         py_style_weight=(9, 2, 1),
-        shuffle_buffer_size=50000,
+        shuffle_buffer_size=100000,
         length_weights={1: 10, 2: 50, 3: 50, 4: 40, 5: 15, 6: 10, 7: 5, 8: 2},
     )
 
@@ -1093,7 +1093,7 @@ def train(
         max_seq_length=max_seq_len,
         text_field="text",
         py_style_weight=(9, 2, 1),
-        shuffle_buffer_size=500000,
+        shuffle_buffer_size=100000,
         length_weights={1: 10, 2: 50, 3: 50, 4: 40, 5: 15, 6: 10, 7: 5, 8: 2},
     )
 
@@ -1388,7 +1388,7 @@ def expand_and_train(
         max_seq_length=max_seq_len,
         text_field="text",
         py_style_weight=(9, 2, 1),
-        shuffle_buffer_size=5000,
+        shuffle_buffer_size=100000,
         length_weights={1: 10, 2: 50, 3: 50, 4: 40, 5: 15, 6: 10, 7: 5, 8: 2},
     )
 
@@ -1409,7 +1409,7 @@ def expand_and_train(
         max_seq_length=max_seq_len,
         text_field="text",
         py_style_weight=(9, 2, 1),
-        shuffle_buffer_size=50000,
+        shuffle_buffer_size=100000,
         length_weights={1: 10, 2: 50, 3: 50, 4: 40, 5: 15, 6: 10, 7: 5, 8: 2},
     )
 
@@ -1711,7 +1711,7 @@ def expand_finetune(
         max_seq_length=final_max_seq_len,
         text_field="text",
         py_style_weight=(9, 2, 1),
-        shuffle_buffer_size=5000,
+        shuffle_buffer_size=100000,
         length_weights={1: 10, 2: 50, 3: 50, 4: 40, 5: 15, 6: 10, 7: 5, 8: 2},
     )
 
@@ -1730,7 +1730,7 @@ def expand_finetune(
         max_seq_length=final_max_seq_len,
         text_field="text",
         py_style_weight=(9, 2, 1),
-        shuffle_buffer_size=50000,
+        shuffle_buffer_size=100000,
         length_weights={1: 10, 2: 50, 3: 50, 4: 40, 5: 15, 6: 10, 7: 5, 8: 2},
     )
 
@@ -1802,4 +1802,3 @@ def expand_finetune(
 
 if __name__ == "__main__":
     app()
-

test.py

@@ -47,8 +47,8 @@ def text_to_pinyin_ids(pinyin_str: str) -> List[int]:
     return [CHAR_TO_ID.get(c, 0) for c in pinyin_str]
 
 
-part1 = "他是一名大学生，在上海读"
+part1 = "明明是国庆节，可是因为月底要结账，财务部所有人都"
-part2 = "dayi"
+part2 = "bxu"
 pinyin_ids = text_to_pinyin_ids(part2)
 len_py = len(pinyin_ids)
 if len_py < 24:
@@ -57,7 +57,7 @@ else:
     pinyin_ids = pinyin_ids[:24]
 pinyin_ids = torch.tensor(pinyin_ids, dtype=torch.long).unsqueeze(0)
 masked_labels = [0, 0, 0, 0, 0, 0, 0, 0]
-part3 = "。"
+part3 = ""
 part4 = "可行|特别|伤害"
 
 encoded = tokenizer(
@@ -83,7 +83,7 @@ sample = {
 
 model = InputMethodEngine(pinyin_vocab_size=30, compile=False)
 
-checkpoint = torch.load("/home/songsenand/下载/best_model.pt", map_location="cpu")
+checkpoint = torch.load("/home/songsenand/下载/20260411(acc34)final_model.ptrom", map_location="cpu")
 model.load_state_dict(checkpoint["model_state_dict"])
 
 input_ids = sample["input_ids"]