ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update llama.py

Browse source
This commit is contained in:
Daniel Han-Chen 2024-01-21 19:10:48 +11:00
commit 7be801ff8d
1 changed files with 1 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

3
unsloth/models/llama.py
View file

@ -105,7 +105,6 @@ def LlamaAttention_fast_forward_inference(
Xn = hidden_states
bsz, _, _ = hidden_states.size()
K1, V1 = past_key_value
print(1)
n_heads = self.num_heads
n_groups = self.num_key_value_groups
@ -305,9 +304,9 @@ def LlamaDecoderLayer_fast_forward(
past_key_value (`Tuple(torch.FloatTensor)`, *optional*): cached past key and value projection states
"""
bsz, q_len, hd = hidden_states.size()
print(hidden_states.size())
if (past_key_value is not None and q_len == 1):
print(1)
# Self Attention
residual = hidden_states
hidden_states = fast_rms_layernorm_inference(self.input_layernorm, hidden_states)