v0.3.1
fixed using zero with tp cannot access weight correctly
[doc] Fix typo under colossalai and doc(#3618)
[shardformer] add Dropout layer support different dropout pattern (#3856)
[utils] refactor parallel layers checkpoint and bcast model on loading checkpoint (#1548)
improved allgather & reducescatter for 3d
[nfc] fix typo colossalai/nn (#3887)
[NFC] polish colossalai/nn/layer/utils/common.py code style (#983)
added skip_bias_add for non-tp linear
[NFC] polish colossalai/nn/layer/wrapper/pipeline_wrapper.py code style (#1303)
[MOE] changed parallelmode to dist process group (#460)
[shardformer] integrated linear 1D with dtensor (#3996)