books/LLMs-from-scratch
1
0
Fork 0
mirror of https://github.com/rasbt/LLMs-from-scratch.git synced 2026-04-10 12:33:42 +00:00
Code Issues Packages Projects Releases Wiki Activity

More efficient angles computation in RoPE (#830)

Browse Source
This commit is contained in:
Sebastian Raschka
2025-09-15 22:23:33 -05:00
committed by GitHub
parent 147dc49ab5
commit b6cd0a312f
11 changed files with 11 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ch05/12_gemma3/standalone-gemma3-plus-kvcache.ipynb
View File

@@ -200,7 +200,7 @@
" positions = torch.arange(context_length, dtype=dtype)\n",
"\n",
" # Compute the angles\n",
" angles = positions[:, None] * inv_freq[None, :] # Shape: (context_length, head_dim // 2)\n",
" angles = positions.unsqueeze(1) * inv_freq.unsqueeze(0) # Shape: (context_length, head_dim // 2)\n",
"\n",
" # Expand angles to match the head_dim\n",
" angles = torch.cat([angles, angles], dim=1) # Shape: (context_length, head_dim)\n",

2
ch05/12_gemma3/standalone-gemma3.ipynb
View File

@@ -200,7 +200,7 @@
" positions = torch.arange(context_length, dtype=dtype)\n",
"\n",
" # Compute the angles\n",
" angles = positions[:, None] * inv_freq[None, :] # Shape: (context_length, head_dim // 2)\n",
" angles = positions.unsqueeze(1) * inv_freq.unsqueeze(0) # Shape: (context_length, head_dim // 2)\n",
"\n",
" # Expand angles to match the head_dim\n",
" angles = torch.cat([angles, angles], dim=1) # Shape: (context_length, head_dim)\n",