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.

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

Fixing Dtype error: expected scalar type BFloat16 but found Float for Torch parts #59

Closed
wants to merge 1 commit into from

Conversation

nanowell
Copy link

@nanowell nanowell commented Oct 9, 2024

This PR will fix dtype errors: RuntimeError: expected scalar type BFloat16 but found Float

This PR will fix dtype errors: RuntimeError: expected scalar type BFloat16 but found Float
@stillmatic
Copy link

stillmatic commented Oct 10, 2024

I fixed it with this

diff --git a/entropix/torch_model.py b/entropix/torch_model.py
index 0ebb3e9..bda7629 100644
--- a/entropix/torch_model.py
+++ b/entropix/torch_model.py
@@ -25,7 +25,7 @@ from typing import Tuple, Optional
 def rms_norm(x: torch.Tensor, w: torch.Tensor, eps: float = 1e-6) -> torch.Tensor:
   return w * (x * torch.rsqrt(torch.pow(x, 2).mean(-1, keepdim=True) + eps))

-def apply_rotary_emb(xq: torch.Tensor, xk: torch.Tensor, freqs_cis: torch.Tensor, dtype: torch.dtype = torch.float32) -> Tuple[torch.Tensor, torch.Tensor]:
+def apply_rotary_emb(xq: torch.Tensor, xk: torch.Tensor, freqs_cis: torch.Tensor, dtype: torch.dtype = torch.bfloat16) -> Tuple[torch.Tensor, torch.Tensor]:
     reshape_xq = xq.float().reshape(*xq.shape[:-1], -1, 2)
     reshape_xk = xk.float().reshape(*xk.shape[:-1], -1, 2)
     xq_ = torch.complex(reshape_xq[..., 0], reshape_xq[..., 1])
@@ -54,7 +54,7 @@ def attention(x: torch.Tensor, layer_weights: LayerWeights, model_params, cur_po
         scores = scores + attn_mask
     mask = torch.where(scores != 0.0, scores, DEFAULT_MASK_VALUE)
     padded_logits = torch.where((mask >= DEFAULT_MASK_VALUE * 0.5), scores, DEFAULT_MASK_VALUE)
-    scores = F.softmax(padded_logits, dim=-1).to(torch.float32)
+    scores = F.softmax(padded_logits, dim=-1).to(x.dtype)
     output = torch.matmul(scores, values)
     output = output.transpose(1, 2).reshape(xq.shape[0], xq.shape[2], -1)
     out = F.linear(output, layer_weights.wo)

alternatively you could pass dtype=x.dtype to the apply_rotary_emb call

edit: looks like same problem as #39

@nanowell nanowell closed this Oct 16, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants