From 26329163291e2f8142494bece2db5548aef469c9 Mon Sep 17 00:00:00 2001
From: wangbluo <2538539015@qq.com>
Date: Wed, 1 May 2024 09:23:43 +0000
Subject: [PATCH] remove useless code
---
colossalai/shardformer/modeling/mistral.py | 1 -
colossalai/shardformer/policies/mistral.py | 1 -
2 files changed, 2 deletions(-)
diff --git a/colossalai/shardformer/modeling/mistral.py b/colossalai/shardformer/modeling/mistral.py
index 642fa3b40..796aeca51 100644
--- a/colossalai/shardformer/modeling/mistral.py
+++ b/colossalai/shardformer/modeling/mistral.py
@@ -270,7 +270,6 @@ class MistralForwards:
shift_labels = labels[..., 1:].contiguous()
# Flatten the tokens
loss_fct = CrossEntropyLoss()
- #shift_logits = shift_logits.view(-1, self.config.vocab_size)
shift_labels = shift_labels.view(-1)
# Enable model parallelism
shift_labels = shift_labels.to(shift_logits.device)
diff --git a/colossalai/shardformer/policies/mistral.py b/colossalai/shardformer/policies/mistral.py
index 362fd11e5..936fd2d24 100644
--- a/colossalai/shardformer/policies/mistral.py
+++ b/colossalai/shardformer/policies/mistral.py
@@ -277,7 +277,6 @@ class MistralForCausalLMPolicy(MistralPolicy):
suffix="lm_head",
target_module=VocabParallelLMHead1D,
kwargs={
- #gather_output=True,
"gather_output": not self.shard_config.parallel_output,
"make_vocab_size_divisible_by": self.shard_config.make_vocab_size_divisible_by,
},