diff options
Diffstat (limited to 'candle-transformers')
| -rw-r--r-- | candle-transformers/src/models/quantized_phi3.rs | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/candle-transformers/src/models/quantized_phi3.rs b/candle-transformers/src/models/quantized_phi3.rs index 51a75f38..1ceb48d1 100644 --- a/candle-transformers/src/models/quantized_phi3.rs +++ b/candle-transformers/src/models/quantized_phi3.rs @@ -127,7 +127,7 @@ impl LayerWeights { .reshape((b_sz, seq_len, self.n_head, self.head_dim))? .transpose(1, 2)?; let k = k - .reshape((b_sz, seq_len, self.n_head, self.head_dim))? + .reshape((b_sz, seq_len, self.n_kv_head, self.head_dim))? .transpose(1, 2)?; let v = v .reshape((b_sz, seq_len, self.n_kv_head, self.head_dim))? |