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 optional arg to specify device for Transformer model. #165

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Add optional arg to specify device for Transformer model. #165

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a39b008
Add optional arg to specify device for Transformer model.
anordin95 Sep 30, 2024
13563f3
Add option to initialize multimodal model on devices other than cuda.…
anordin95 Oct 2, 2024
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
14 changes: 7 additions & 7 deletions models/llama3/reference_impl/model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -113,7 +113,7 @@ def repeat_kv(x: torch.Tensor, n_rep: int) -> torch.Tensor:


class Attention(nn.Module):
def __init__(self, args: ModelArgs):
def __init__(self, args: ModelArgs, device: torch.device = torch.device('cuda')):
super().__init__()
self.n_kv_heads = args.n_heads if args.n_kv_heads is None else args.n_kv_heads
model_parallel_size = fs_init.get_model_parallel_world_size()
Expand Down Expand Up @@ -158,15 +158,15 @@ def __init__(self, args: ModelArgs):
self.n_local_kv_heads,
self.head_dim,
)
).cuda()
).to(device)
self.cache_v = torch.zeros(
(
args.max_batch_size,
args.max_seq_len,
self.n_local_kv_heads,
self.head_dim,
)
).cuda()
).to(device)

def forward(
self,
Expand Down Expand Up @@ -245,12 +245,12 @@ def forward(self, x):


class TransformerBlock(nn.Module):
def __init__(self, layer_id: int, args: ModelArgs):
def __init__(self, layer_id: int, args: ModelArgs, device: torch.device = torch.device('cuda')):
super().__init__()
self.n_heads = args.n_heads
self.dim = args.dim
self.head_dim = args.dim // args.n_heads
self.attention = Attention(args)
self.attention = Attention(args, device)
self.feed_forward = FeedForward(
dim=args.dim,
hidden_dim=4 * args.dim,
Expand All @@ -274,7 +274,7 @@ def forward(


class Transformer(nn.Module):
def __init__(self, params: ModelArgs):
def __init__(self, params: ModelArgs, device: torch.device = torch.device('cuda')):
super().__init__()
self.params = params
self.vocab_size = params.vocab_size
Expand All @@ -286,7 +286,7 @@ def __init__(self, params: ModelArgs):

self.layers = torch.nn.ModuleList()
for layer_id in range(params.n_layers):
self.layers.append(TransformerBlock(layer_id, params))
self.layers.append(TransformerBlock(layer_id, params, device))

self.norm = RMSNorm(params.dim, eps=params.norm_eps)
self.output = ColumnParallelLinear(
Expand Down