fix ChatGLM RLHF

src/llmtuner/tuner/ppo/trainer.py

@@ -182,6 +182,8 @@ class PPOPeftTrainer(PPOTrainer, PeftTrainer):
         replace_model(unwrapped_model, target="reward")
         batch = self.prepare_model_inputs(queries, responses)
         _, _, values = self.model(**batch, output_hidden_states=True, return_dict=True)
+        if values.size(0) != batch["input_ids"].size(0):
+            values = torch.transpose(values, 0, 1)
         rewards = [reward for reward in values[:, -1].float().detach().cpu()] # use fp32 type
         replace_model(unwrapped_model, target="default")
         return rewards

src/llmtuner/tuner/rm/trainer.py

@@ -42,6 +42,8 @@ class PairwisePeftTrainer(PeftTrainer):
         """
         batch_size = inputs["input_ids"].size(0) // 2
         _, _, values = model(**inputs, output_hidden_states=True, return_dict=True)
+        if values.size(0) != inputs["input_ids"].size(0):
+            values = torch.transpose(values, 0, 1)
         r_accept, r_reject = values[:, -1].split(batch_size, dim=0)
         loss = -torch.log(torch.sigmoid(r_accept - r_reject)).mean()
         return (loss, [loss, r_accept, r_reject]) if return_outputs else loss