We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent fda6cd1 commit e23bfedCopy full SHA for e23bfed
penzai/models/transformer/variants/llamalike_common.py
@@ -596,7 +596,7 @@ def llamalike_from_huggingface_model(
596
num_decoder_blocks=hf_config.num_hidden_layers,
597
vocab_size=hf_config.vocab_size,
598
mlp_variant="swiglu",
599
- rope_wavelength=10_000,
+ rope_wavelength=hf_config.rope_theta,
600
tie_embedder_and_logits=False,
601
attention_type=attention_type,
602
rms_norm_eps=hf_config.rms_norm_eps,
0 commit comments