Skip to content

Commit

Permalink
finetune : fix ggerganov#3404 (ggerganov#3437)
Browse files Browse the repository at this point in the history
the shapes for init model of gqa models was wrong
  • Loading branch information
xaedes authored Oct 2, 2023
1 parent a847676 commit a03ce38
Showing 1 changed file with 2 additions and 2 deletions.
4 changes: 2 additions & 2 deletions examples/finetune/finetune.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -332,8 +332,8 @@ static void init_model(struct llama_model * input, struct my_llama_model * model

assert_shape_1d(layer.attention_norm, hparams.n_embd);
assert_shape_2d(layer.wq, hparams.n_embd, hparams.n_embd);
assert_shape_2d(layer.wk, hparams.n_embd, hparams.n_embd);
assert_shape_2d(layer.wv, hparams.n_embd, hparams.n_embd);
assert_shape_2d(layer.wk, hparams.n_embd, hparams.n_embd_gqa());
assert_shape_2d(layer.wv, hparams.n_embd, hparams.n_embd_gqa());
assert_shape_2d(layer.wo, hparams.n_embd, hparams.n_embd);
assert_shape_1d(layer.ffn_norm, hparams.n_embd);
assert_shape_2d(layer.w1, hparams.n_embd, hparams.n_ff);
Expand Down

0 comments on commit a03ce38

Please sign in to comment.