我尝试改写你的代码在分类问题上，但是结果只有50%。这是为什么呢？我也尝试把bf16改为了float32，还有哪些可能存在的问题导致我的模型无法训练？（还有其他几个数据集也存在同样的问题）
x_enc = self.normalize_layers(x_enc, 'norm')
B, T, N = x_enc.size()
x_enc = x_enc.permute(0, 2, 1).contiguous().reshape(B * N, T, 1)
# 2. 计算统计特征 (保持不变，这些特征对分类也很有用)
min_values = torch.min(x_enc, dim=1)[0]
max_values = torch.max(x_enc, dim=1)[0]
medians = torch.median(x_enc, dim=1).values
lags = self.calcute_lags(x_enc)
trends = x_enc.diff(dim=1).sum(dim=1)
# 3. 修改 Prompt (重点修改部分)
prompt = []
for b in range(x_enc.shape[0]):
min_values_str = str(min_values[b].tolist()[0])
max_values_str = str(max_values[b].tolist()[0])
median_values_str = str(medians[b].tolist()[0])
lags_values_str = str(lags[b].tolist())
# --- 修改点：提示词改为分类任务 ---
prompt_ = (
f"<|start_prompt|>Dataset description: {self.description}"
# 这里的 class_names 需要你在初始化时定义，或者直接写死类别数量
f"Task description: Classify the input time series into one of the {self.num_classes} classes; "
"Input statistics: "
f"min value {min_values_str}, "
f"max value {max_values_str}, "
f"median value {median_values_str}, "
f"the trend of input is {'upward' if trends[b] > 0 else 'downward'}, "
f"top 5 lags are : {lags_values_str}<|<end_prompt>|>"
)
prompt.append(prompt_)
x_enc = x_enc.reshape(B, N, T).permute(0, 2, 1).contiguous()
# 4. LLM 编码过程 (基本保持不变)
prompt = self.tokenizer(prompt, return_tensors="pt", padding=True, truncation=True, max_length=2048).input_ids
prompt_embeddings = self.llm_model.get_input_embeddings()(prompt.to(x_enc.device))
source_embeddings = self.mapping_layer(self.word_embeddings.permute(1, 0)).permute(1, 0)
x_enc = x_enc.permute(0, 2, 1).contiguous()
enc_out, n_vars = self.patch_embedding(x_enc)
enc_out = self.reprogramming_layer(enc_out, source_embeddings, source_embeddings)
llama_enc_out = torch.cat([prompt_embeddings, enc_out], dim=1)
# 获取 LLM 输出
dec_out = self.llm_model(inputs_embeds=llama_enc_out).last_hidden_state
# ts_out = dec_out[:, -self.patch_nums:, :]
# cls_token_out = ts_out.reshape(B, self.patch_nums, -1).mean(dim=1)
# outputs = self.output_projection(cls_token_out)
dec_out = dec_out[:,0,:]
# cls_token_out = dec_out.mean(dim=1)
cls_token_out = dec_out.reshape(B, N, -1).contiguous()
# cls_token_out = cls_token_out.reshape(B, -1)
outputs = self.output_projection(cls_token_out)
`
我尝试改写你的代码在分类问题上，但是结果只有50%。这是为什么呢？我也尝试把bf16改为了float32，还有哪些可能存在的问题导致我的模型无法训练？（还有其他几个数据集也存在同样的问题）
这是我的分类forward：
` def classification(self, x_enc):