[fix] loss-issues-430

trainer/train_distill_reason.py

@@ -79,7 +79,7 @@ def train_epoch(epoch, wandb):
         if step % args.log_interval == 0 or step == iter_per_epoch - 1:
             spend_time = time.time() - start_time
             Logger(
-                'Epoch:[{}/{}]({}/{}) loss:{:.3f} lr:{:.12f} epoch_Time:{}min:'.format(
+                'Epoch:[{}/{}]({}/{}) loss:{:.6f} lr:{:.12f} epoch_Time:{}min:'.format(
                     epoch + 1,
                     args.epochs,
                     step,

trainer/train_distillation.py

@@ -113,7 +113,7 @@ def train_epoch(epoch, wandb, alpha=0.0, temperature=1.0):
         if step % args.log_interval == 0 or step == iter_per_epoch - 1:
             spend_time = time.time() - start_time
             Logger(
-                'Epoch:[{}/{}]({}/{}) loss:{:.4f} lr:{:.12f} epoch_Time:{}min:'.format(
+                'Epoch:[{}/{}]({}/{}) loss:{:.6f} lr:{:.12f} epoch_Time:{}min:'.format(
                     epoch,
                     args.epochs - 1,
                     step,

trainer/train_dpo.py

@@ -103,7 +103,7 @@ def train_epoch(epoch, wandb):
         if step % args.log_interval == 0 or step == iter_per_epoch - 1:
             spend_time = time.time() - start_time
             Logger(
-                'Epoch:[{}/{}]({}/{}) loss:{:.3f} lr:{:.12f} epoch_Time:{}min:'.format(
+                'Epoch:[{}/{}]({}/{}) loss:{:.6f} lr:{:.12f} epoch_Time:{}min:'.format(
                     epoch + 1,
                     args.epochs,
                     step,

trainer/train_full_sft.py

@@ -66,7 +66,7 @@ def train_epoch(epoch, wandb):
         if step % args.log_interval == 0 or step == iter_per_epoch - 1:
             spend_time = time.time() - start_time
             Logger(
-                'Epoch:[{}/{}]({}/{}) loss:{:.3f} lr:{:.12f} epoch_Time:{}min:'.format(
+                'Epoch:[{}/{}]({}/{}) loss:{:.6f} lr:{:.12f} epoch_Time:{}min:'.format(
                     epoch + 1,
                     args.epochs,
                     step,

trainer/train_grpo.py

@@ -169,7 +169,7 @@ def grpo_train_epoch(epoch, wandb):
 
             Logger(
                 f'Epoch: {epoch}, Step: {step + 1}/{iter_per_epoch}, '
-                f'Actor Loss: {policy_loss_val:.4f}, Reward: {avg_reward_val:.4f}, '
+                f'Actor Loss: {policy_loss_val:.6f}, Reward: {avg_reward_val:.6f}, '
                 f'Avg Response Len: {avg_len_val:.2f}, LR: {current_lr:.2e}')
 
             if wandb and (not ddp or dist.get_rank() == 0):

trainer/train_lora.py

@@ -67,7 +67,7 @@ def train_epoch(epoch, wandb):
         if step % args.log_interval == 0 or step == iter_per_epoch - 1:
             spend_time = time.time() - start_time
             Logger(
-                'Epoch:[{}/{}]({}/{}) loss:{:.3f} lr:{:.12f} epoch_Time:{}min:'.format(
+                'Epoch:[{}/{}]({}/{}) loss:{:.6f} lr:{:.12f} epoch_Time:{}min:'.format(
                     epoch + 1,
                     args.epochs,
                     step,

trainer/train_ppo.py

@@ -193,8 +193,8 @@ def ppo_train_epoch(epoch: int, wandb_run, old_actor_model, ref_model, actor_sch
                 })
 
             Logger(f"Epoch: {epoch}, Step: {step + 1}/{len(train_loader)}, "
-                   f"Actor Loss: {actor_loss_val:.4f}, Critic Loss: {critic_loss_val:.4f}, "
+                   f"Actor Loss: {actor_loss_val:.6f}, Critic Loss: {critic_loss_val:.6f}, "
-                   f"Reward: {reward_val:.4f}, KL: {kl_val:.4f}, KL_ref: {kl_ref_val:.4f}, "
+                   f"Reward: {reward_val:.6f}, KL: {kl_val:.6f}, KL_ref: {kl_ref_val:.6f}, "
                    f"Avg Response Len: {avg_len_val:.2f}, Actor LR: {actor_lr:.2e}, Critic LR: {critic_lr:.2e}")
 
         if (step + 1) % args.update_old_actor_freq == 0:

trainer/train_pretrain.py

@@ -66,7 +66,7 @@ def train_epoch(epoch, wandb):
         if step % args.log_interval == 0 or step == iter_per_epoch - 1:
             spend_time = time.time() - start_time
             Logger(
-                'Epoch:[{}/{}]({}/{}) loss:{:.3f} lr:{:.12f} epoch_Time:{}min:'.format(
+                'Epoch:[{}/{}]({}/{}) loss:{:.6f} lr:{:.12f} epoch_Time:{}min:'.format(
                     epoch + 1,
                     args.epochs,
                     step,

trainer/train_spo.py

@@ -216,7 +216,7 @@ def spo_train_epoch(epoch, wandb, value_tracker):
 
             Logger(
                 f'Epoch: {epoch}, Step: {step + 1}/{iter_per_epoch}, '
-                f'Actor Loss: {policy_loss_val:.4f}, Reward: {avg_reward_val:.4f}, '
+                f'Actor Loss: {policy_loss_val:.6f}, Reward: {avg_reward_val:.6f}, '
                 f'Baseline: {avg_baseline_val:.4f}, KL: {kl_val:.4f}, Rho: {rho:.4f}, Avg Response Len: {avg_len_val:.2f}, LR: {current_lr:.2e}')
 
             if wandb and (not ddp or dist.get_rank() == 0):