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

Add profiler annotation for the decoderonly example #8034

Draft
wants to merge 1 commit into
base: master
Choose a base branch
from
Draft

Add profiler annotation for the decoderonly example #8034

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
6 changes: 6 additions & 0 deletions examples/decoder_only_model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,7 @@
import torch.nn.functional as F
from torch import nn

import torch_xla.debug.profiler as xp

# the default config is intentionally kept low to make it runable on a sigle tpu v2-8 core.
@dataclass
Expand Down Expand Up @@ -44,6 +45,7 @@ def __init__(self, hidden_size, eps=1e-6):
self.weight = nn.Parameter(torch.ones(hidden_size))
self.variance_epsilon = eps

@xp.trace_me("RMSNorm")
def forward(self, hidden_states):
input_dtype = hidden_states.dtype
hidden_states = hidden_states.to(torch.float32)
Expand Down Expand Up @@ -79,6 +81,7 @@ def __init__(self, config: DecoderOnlyConfig):
self.num_heads * self.head_dim, self.hidden_size, bias=False)
self.flash_attention_impl = None

@xp.trace_me("attention")
def forward(
self,
hidden_states: torch.Tensor,
Expand Down Expand Up @@ -153,6 +156,7 @@ def __init__(self, config: DecoderOnlyConfig):
self.intermediate_size, self.hidden_size, bias=False)
self.act_fn = F.silu

@xp.trace_me("MLP")
def forward(self, x):
# [B, S, H] -> [B, S, I]
up_proj = self.up_proj(x)
Expand All @@ -173,6 +177,7 @@ def __init__(self, config: DecoderOnlyConfig):
self.input_layernorm = RMSNorm(config.hidden_size)
self.post_attention_layernorm = RMSNorm(config.hidden_size)

@xp.trace_me("DecoderLayer")
def forward(
self,
hidden_states: torch.Tensor,
Expand Down Expand Up @@ -209,6 +214,7 @@ def __init__(self, config: DecoderOnlyConfig):
self.norm = RMSNorm(config.hidden_size)
self.output = nn.Linear(config.hidden_size, self.vocab_size, bias=False)

@xp.trace_me("DecoderOnlyModel")
def forward(
self,
input_ids: torch.LongTensor = None,
Expand Down
Loading