mirror of
https://github.com/ggerganov/llama.cpp.git
synced 2024-12-25 02:44:36 +00:00
parent
1ddb52ec38
commit
22da05536f
@ -1433,7 +1433,8 @@ void ggml_metal_graph_compute(
|
|||||||
const int n_past = ((int32_t *) dst->op_params)[0];
|
const int n_past = ((int32_t *) dst->op_params)[0];
|
||||||
const int n_dims = ((int32_t *) dst->op_params)[1];
|
const int n_dims = ((int32_t *) dst->op_params)[1];
|
||||||
const int mode = ((int32_t *) dst->op_params)[2];
|
const int mode = ((int32_t *) dst->op_params)[2];
|
||||||
const int n_orig_ctx = ((int32_t *) dst->op_params)[3];
|
// skip 3, n_ctx, used in GLM RoPE, unimplemented in metal
|
||||||
|
const int n_orig_ctx = ((int32_t *) dst->op_params)[4];
|
||||||
|
|
||||||
float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow;
|
float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow;
|
||||||
memcpy(&freq_base, (int32_t *) dst->op_params + 5, sizeof(float));
|
memcpy(&freq_base, (int32_t *) dst->op_params + 5, sizeof(float));
|
||||||
|
Loading…
Reference in New Issue
Block a user