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

remove dtype from llama precompute_freqs_cis #546

Closed
wants to merge 1 commit into from
Closed

remove dtype from llama precompute_freqs_cis #546

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
9 changes: 5 additions & 4 deletions extra/models/llama.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,11 +3,10 @@
from tinygrad.helpers import getenv

# https://github.com/facebookresearch/llama/blob/1076b9c51c77ad06e9d7ba8a4c6df775741732bd/llama/model.py#L47
def precompute_freqs_cis(dim: int, end: int, theta: float = 10000.0, dtype=dtypes.half) -> Tensor:
def precompute_freqs_cis(dim: int, end: int, theta: float = 10000.0) -> Tensor:
freqs = 1.0 / (theta ** (Tensor.arange(0, dim, 2)[:(dim // 2)] / dim))
freqs = Tensor.arange(end).unsqueeze(dim=1) * freqs.unsqueeze(dim=0)
# TODO: move dtype outside this
return Tensor.stack(freqs.cos().cast(dtype), freqs.sin().cast(dtype), dim=-1).reshape(1, end, 1, dim//2, 2)
return Tensor.stack(freqs.cos(), freqs.sin(), dim=-1).reshape(1, end, 1, dim//2, 2)

# (a+i*b) * (c+i*d) = (ac-bd) + i*(ad+bc)
def complex_mult(A, c, d):
Expand Down Expand Up @@ -163,9 +162,11 @@ def __init__(self, dim:int, hidden_dim:int, n_heads:int, n_layers:int, norm_eps:

def forward(self, tokens:Tensor, start_pos:Union[Variable,int], temperature:float, top_k:int, top_p:float, alpha_f:float, alpha_p:float):
_bsz, seqlen = tokens.shape
h = self.tok_embeddings(tokens)

self.freqs_cis = self.freqs_cis.cast(h.dtype).realize()
freqs_cis = self.freqs_cis.shrink((None, (start_pos, start_pos+seqlen),None,None,None))

h = self.tok_embeddings(tokens)
mask = Tensor.full((1, 1, seqlen, start_pos+seqlen), float("-inf"), dtype=h.dtype, device=h.device).triu(start_pos+1).realize() if seqlen > 1 else None
for layer in self.layers: h = layer(h, start_pos, freqs_cis, mask)
logits = self.output(self.norm(h)).float()[:, -1, :]
Expand Down