Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

revert partially Enable/Optimize flan t5 xxl on deepspeed z3 #257 #320

Merged
merged 4 commits into from
Aug 9, 2023
Merged

revert partially Enable/Optimize flan t5 xxl on deepspeed z3 #257 #320

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
a2761fd
revert partially Enable/Optimize flan t5 xxl on deepspeed z3 #257
hsubramony Aug 3, 2023
70bed73
remove unwanted comments
hsubramony Aug 3, 2023
8065dc8
Merge branch 'main' into flan_t5
hsubramony Aug 9, 2023
ea8ad1c
change contiguous_gradients to false for ds_flan_t5_z3_config_bf16
hsubramony Aug 9, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 0 additions & 16 deletions optimum/habana/transformers/modeling_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,12 +26,6 @@
GaudiLlamaForCausalLM,
GaudiOPTForCausalLM,
GaudiOPTLearnedPositionalEmbedding,
GaudiT5DenseActDense,
GaudiT5DenseGatedActDense,
GaudiT5LayerCrossAttention,
GaudiT5LayerFF,
GaudiT5LayerSelfAttention,
GaudiT5Stack,
_gaudi_esmfold_attention_wrap_up,
gaudi_albert_forward,
gaudi_bloom_attention_forward,
Expand Down Expand Up @@ -64,7 +58,6 @@
gaudi_opt_model_forward,
gaudi_rot_matmul,
gaudi_rot_vec_mul,
gaudi_T5Attention_forward,
gaudi_vit_self_attention_forward,
gaudi_wav2vec2_forward,
)
Expand Down Expand Up @@ -168,12 +161,3 @@ def adapt_transformers_to_gaudi():
transformers.models.llama.modeling_llama.LlamaModel.forward = gaudi_llama_model_forward
transformers.models.llama.modeling_llama.LlamaDecoderLayer.forward = gaudi_llama_decoder_layer_forward
transformers.models.llama.modeling_llama.LlamaAttention.forward = gaudi_llama_attention_forward

# Dropout kernel improvement for Flan-T5
transformers.models.t5.modeling_t5.T5Stack = GaudiT5Stack
transformers.models.t5.modeling_t5.T5DenseGatedActDense = GaudiT5DenseGatedActDense
transformers.models.t5.modeling_t5.T5LayerFF = GaudiT5LayerFF
transformers.models.t5.modeling_t5.T5LayerSelfAttention = GaudiT5LayerSelfAttention
transformers.models.t5.modeling_t5.T5LayerCrossAttention = GaudiT5LayerCrossAttention
transformers.models.t5.modeling_t5.T5DenseActDense = GaudiT5DenseActDense
transformers.models.t5.modeling_t5.T5Attention.forward = gaudi_T5Attention_forward
9 changes: 0 additions & 9 deletions optimum/habana/transformers/models/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,15 +46,6 @@
gaudi_opt_decoder_layer_forward,
gaudi_opt_model_forward,
)
from .t5 import (
GaudiT5DenseActDense,
GaudiT5DenseGatedActDense,
GaudiT5LayerCrossAttention,
GaudiT5LayerFF,
GaudiT5LayerSelfAttention,
GaudiT5Stack,
gaudi_T5Attention_forward,
)
from .vit import gaudi_vit_self_attention_forward
from .wav2vec2 import (
_gaudi_wav2vec2_compute_mask_indices,
Expand Down
9 changes: 0 additions & 9 deletions optimum/habana/transformers/models/t5/__init__.py
View file
Open in desktop

This file was deleted.

215 changes: 0 additions & 215 deletions optimum/habana/transformers/models/t5/modeling_t5.py
View file
Open in desktop

This file was deleted.

Loading