Update dnaflash.py

dnaflash.py

@@ -7,7 +7,11 @@ from einops import rearrange
 from rotary_embedding_torch import RotaryEmbedding
 
 from transformers import PreTrainedModel, PretrainedConfig
-from transformers.modeling_outputs import MaskedLMOutput
+from transformers.modeling_outputs import MaskedLMOutput, SequenceClassifierOutput
+
+import torch.utils.checkpoint
+from torch import nn, Tensor
+from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
 
 # helper functions
 
@@ -413,3 +417,71 @@ class FLASHTransformerForPretrained(PreTrainedModel):
         logits, x = self.model(inputs["input_ids"], mask=inputs["attention_mask"])
         return MaskedLMOutput(logits=logits, hidden_states=x, loss=None, attentions=None)
 
+class FLASHTransformerForSequenceClassification(FLASHTransformerForPretrained):
+    def __init__(self, config):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+        self.config = config
+
+        self.score = nn.Linear(config.hidden_size, self.num_labels, bias=False)
+        if getattr(config, "use_mlp_classifier", False):
+            self.score = nn.Sequential(
+                nn.Linear(config.hidden_size, config.hidden_size),
+                nn.GELU(),
+                nn.Dropout(0.1),
+                nn.Linear(config.hidden_size, self.num_labels, bias=False),
+            )
+
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, SequenceClassifierOutput]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
+            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
+            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
+        """
+
+        # 获取基模型输出
+        outputs = super().forward(
+            input_ids=input_ids
+        )
+        hidden_states = outputs["hidden_states"]
+        input_mask_expanded = input_ids["attention_mask"].unsqueeze(-1).expand(hidden_states.size())  # 维度匹配
+        mean_pooled = torch.sum(token_embeddings * input_mask_expanded, dim=1) / input_mask_expanded.sum(dim=1)  # 计算加权平均
+        logits = self.score(mean_pooled)
+
+        loss = None
+        if labels is not None:
+            labels = labels.to(logits.device)
+
+            if self.config.problem_type is None:
+                if self.num_labels == 1:
+                    self.config.problem_type = "regression"
+                elif self.num_labels > 1 and (
+                    labels.dtype == torch.long or labels.dtype == torch.int
+                ):
+                    self.config.problem_type = "single_label_classification"
+                else:
+                    self.config.problem_type = "multi_label_classification"
+
+            if self.config.problem_type == "regression":
+                loss_fct = MSELoss()
+                if self.num_labels == 1:
+                    loss = loss_fct(logits.squeeze(), labels.squeeze())
+                else:
+                    loss = loss_fct(logits, labels)
+            elif self.config.problem_type == "single_label_classification":
+                loss_fct = CrossEntropyLoss()
+                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
+            elif self.config.problem_type == "multi_label_classification":
+                loss_fct = BCEWithLogitsLoss()
+                loss = loss_fct(logits, labels)
+        if not return_dict:
+            output = (logits,)
+            return ((loss,) + output) if loss is not None else output
+
+        return SequenceClassifierOutput(loss=loss, logits=logits)